LEFT BEHIND LEFT BEHIND Chronic Poverty in Latin America and the Caribbean Renos Vakis, Jamele Rigolini, and Leonardo Lucchetti © 2016 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW, Washington, DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved 1 2 3 4 19 18 17 16 This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, 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 concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. RIGHTS AND PERMISSIONS This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO) http://creativecommons.org/ licenses/by/3.0/igo. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: • Attribution—Please cite the work as follows: Vakis, Renos, Jamele Rigolini, and Leonardo Lucchetti. 2016. Left Behind: Chronic Poverty in Latin America and the Caribbean. Latin American Development Forum. Washington, DC: World Bank. doi:10.1596/978-1-4648-0660-5. License: Creative Commons Attribution CC BY 3.0 IGO • Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by The World Bank and should not be considered an official World Bank translation. The World Bank shall not be liable for any content or error in this translation. • Adaptations—If you create an adaptation of this work, please add the following disclaimer along with the attribution: This is an adaptation of an original work by The World Bank. Views and opinions expressed in the adaptation are the sole responsibility of the author or authors of the adaptation and are not endorsed by The World Bank. • Third-party content—The World Bank does not necessarily own each component of the content contained within the work. The World Bank therefore does not warrant that the use of any third-party-owned individual component or part contained in the work will not infringe on the rights of those third parties. The risk of claims resulting from such infringement rests solely with you. If you wish to re-use a component of the work, it is your responsibility to determine whether permission is needed for that re-use and to obtain permission from the copyright owner. Examples of components can include, but are not limited to, tables, figures, or images. All queries on rights and licenses should be addressed to the Publishing and Knowledge Division, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. ISBN (paper): 978-1-4648-0660-5 ISBN (electronic): 978-1-4648-0661-2 DOI: 10.1596/978-1-4648-0660-5 Cover image: © iStock.com/DNY59. Used with the permission of iStock.com/DNY59. Further permission required for reuse. Cover design: Bill Pragluski/Critical Stages Library of Congress Cataloging-in-Publication Data Names: Vakis, Renos Nicos, author. | Rigolini, Jamele, author. | Lucchetti, Leonardo, author. Title: Left behind : chronic poverty in Latin America and the Caribbean / Renos Vakis, Jamele Rigolini and Leonardo Lucchetti. Description: Washington, D.C. : World Bank, 2016. | Series: Latin American Development Forum | Includes bibliographical references. | Description based on print version record and CIP data provided by publisher; resource not viewed. Identifiers: LCCN 2016021906 (print) | LCCN 2016014211 (ebook) | ISBN 9781464806612 () | ISBN 9781464806612 | ISBN 9781464806605 (pdf) Subjects: LCSH: Poverty—Caribbean Area. | Poverty—Latin America. | Caribbean Area—Economic conditions—21st century. | Latin America—Economic conditions—21st century. | Caribbean Area—Economic policy—21st century. | Latin America— Economic policy—21st century. Classification: LCC HC151.Z9 (print) | LCC HC151.Z9 V35 2016 (ebook) | DDC 339.4/6098—dc23 LC record available at https://lccn.loc.gov/2016021906 Latin American Development Forum Series This series was created in 2003 to promote debate, disseminate information and analysis, and convey the excitement and complexity of the most topical issues in economic and social development in Latin America and the Caribbean. It is sponsored by the Inter-American Development Bank, the United Nations Economic Commission for Latin America and the Caribbean, and the World Bank, and represents the highest quality in each institution’s research and activity output. Titles in the series have been selected for their relevance to the academic community, policy makers, researchers, and interested readers, and have been subjected to rigorous anonymous peer review prior to publication. Advisory Committee Members Alicia Bárcena Ibarra, Executive Secretary, Economic Commission for Latin America and the Caribbean, United Nations Inés Bustillo, Director, Washington Office, Economic Commission for Latin America and the Caribbean, United Nations Augusto de la Torre, Chief Economist, Latin America and the Caribbean Region, World Bank Daniel Lederman, Deputy Chief Economist, Latin America and the Caribbean Region, World Bank Santiago Levy, Vice President for Sectors and Knowledge, Inter-American Development Bank Roberto Rigobon, Professor of Applied Economics, MIT Sloan School of Management José Juan Ruiz, Chief Economist and Manager of the Research Department, Inter- American Development Bank Ernesto Talvi, Director, Brookings Global-CERES Economic and Social Policy in Latin America Initiative Andrés Velasco, Cieplan, Chile v Titles in the Latin American Development Forum Series Left Behind: Chronic Poverty in Latin America and the Caribbean (2016) by Renos Vakis, Jamele Rigolini, and Leonardo Lucchetti Cashing in on Education: Women, Childcare, and Prosperity in Latin America and the Caribbean (2016) by Mercedes Mateo Díaz and Lourdes Rodriguez-Chamussy Work and Family: Latin American and Caribbean Women in Search of a New Balance (2016) by Laura Chioda Great Teachers: How to Raise Student Learning in Latin America and the Caribbean (2014) by Barbara Bruns and Javier Luque Entrepreneurship in Latin America: A Step Up the Social Ladder? (2013) by Eduardo Lora and Francesca Castellani, editors Emerging Issues in Financial Development: Lessons from Latin America (2013) by Tatiana Didier and Sergio L. Schmukler, editors New Century, Old Disparities: Gaps in Ethnic and Gender Earnings in Latin America and the Caribbean (2012) by Hugo Ñopo Does What You Export Matter? In Search of Empirical Guidance for Industrial Policies (2012) by Daniel Lederman and William F. Maloney From Right to Reality: Incentives, Labor Markets, and the Challenge of Achieving Universal Social Protection in Latin America and the Caribbean (2012) by Helena Ribe, David Robalino, and Ian Walker Breeding Latin American Tigers: Operational Principles for Rehabilitating Industrial Policies (2011) by Robert Devlin and Graciela Moguillansky New Policies for Mandatory Defined Contribution Pensions: Industrial Organization Models and Investment Products (2010) by Gregorio Impavido, Esperanza Lasagabaster, and Manuel García-Huitrón The Quality of Life in Latin American Cities: Markets and Perception (2010) by Eduardo Lora, Andrew Powell, Bernard M. S. van Praag, and Pablo Sanguinetti, editors vii Discrimination in Latin America: An Economic Perspective (2010) by Hugo Ñopo, Alberto Chong, and Andrea Moro, editors The Promise of Early Childhood Development in Latin America and the Caribbean (2010) by Emiliana Vegas and Lucrecia Santibáñez Job Creation in Latin America and the Caribbean: Trends and Policy Challenges (2009) by Carmen Pagés, Gaëlle Pierre, and Stefano Scarpetta China’s and India’s Challenge to Latin America: Opportunity or Threat? (2009) by Daniel Lederman, Marcelo Olarreaga, and Guillermo E. Perry, editors Does the Investment Climate Matter? Microeconomic Foundations of Growth in Latin America (2009) by Pablo Fajnzylber, Jose Luis Guasch, and J. Humberto López, editors Measuring Inequality of Opportunities in Latin America and the Caribbean (2009) by Ricardo de Paes Barros, Francisco H. G. Ferreira, José R. Molinas Vega, and Jaime Saavedra Chanduvi The Impact of Private Sector Participation in Infrastructure: Lights, Shadows, and the Road Ahead (2008) by Luis Andres, Jose Luis Guasch, Thomas Haven, and Vivien Foster Remittances and Development: Lessons from Latin America (2008) by Pablo Fajnzylber and J. Humberto López, editors Fiscal Policy, Stabilization, and Growth: Prudence or Abstinence? (2007) by Guillermo Perry, Luis Servén, and Rodrigo Suescún, editors Raising Student Learning in Latin America: Challenges for the 21st Century (2007) by Emiliana Vegas and Jenny Petrow Investor Protection and Corporate Governance: Firm-level Evidence Across Latin America (2007) by Alberto Chong and Florencio López-de-Silanes, editors Natural Resources: Neither Curse nor Destiny (2007) by Daniel Lederman and William F. Maloney, editors The State of State Reform in Latin America (2006) by Eduardo Lora, editor Emerging Capital Markets and Globalization: The Latin American Experience (2006) by Augusto de la Torre and Sergio L. Schmukler Beyond Survival: Protecting Households from Health Shocks in Latin America (2006) by Cristian C. Baeza and Truman G. Packard Beyond Reforms: Structural Dynamics and Macroeconomic Vulnerability (2005) by José Antonio Ocampo, editor Privatization in Latin America: Myths and Reality (2005) by Alberto Chong and Florencio López-de-Silanes, editors Keeping the Promise of Social Security in Latin America (2004) by Indermit S. Gill, Truman G. Packard, and Juan Yermo viii TITLES IN THE LATIN AMERICAN DEVELOPMENT FORUM SERIES Lessons from NAFTA: For Latin America and the Caribbean (2004) by Daniel Lederman, William F. Maloney, and Luis Servén The Limits of Stabilization: Infrastructure, Public Deficits, and Growth in Latin America (2003) by William Easterly and Luis Servén, editors Globalization and Development: A Latin American and Caribbean Perspective (2003) by José Antonio Ocampo and Juan Martin, editors Is Geography Destiny? Lessons from Latin America (2003) by John Luke Gallup, Alejandro Gaviria, and Eduardo Lora TITLES IN THE LATIN AMERICAN DEVELOPMENT FORUM SERIES ix Contents Foreword xvii Acknowledgments xix About the Authors xxi Abbreviations xxiii Overview 1 A Successful Decade with Challenges Ahead 1 A Framework for Studying Chronic Poverty 3 Measuring Chronic Poverty in the Absence of Longitudinal Data 5 Five Stylized Facts about Chronic Poverty in Latin America and the Caribbean 5 From Diagnosis to Policies: Design Elements Supporting the Chronically Poor 18 Notes 24 References 24 Chapter 1: The Roaring 2000s and the Left Behind 27 Notes 31 References 31 Chapter 2: What Is Chronic Poverty and How Is It Measured? 33 Conceptual Underpinnings 33 The Empirical Challenges of Measuring Chronic Poverty 40 Notes 48 References 49 Chapter 3: Five Facts about Chronic Poverty in Latin America and the Caribbean 53 Stylized Fact 1: One of Five People in Latin America and the Caribbean Lives in Chronic Poverty 53 xi Stylized Fact 2: Chronic Poverty Tends to Be Geographically Concentrated 57 Stylized Fact 3: Chronic Poverty Is As Big a Problem in Urban Areas as in Rural Areas 60 Stylized Fact 4: Economic Growth Was Not Sufficient to Lift the Chronically Poor out of Poverty 62 Stylized Fact 5: The Chronically Poor Have Limited Income Opportunities 66 Notes 70 References 70 Chapter 4: Unraveling the Complexities of Chronic Poverty 71 Endowments, the Context, and Chronic Poverty 72 State of Mind and the Process of Emerging from Poverty 87 Notes 94 References 94 Chapter 5: From Diagnosis to Policies: Crafting Coordinated Policies That Reduce Chronic Poverty 99 Improving Endowments and the Enabling Context 99 Coordinating Poverty-Reduction Efforts 101 Recognizing the State of Mind of the Poor in Crafting Policy 111 Developing Coordinated Solutions that Address Behavioral Constraints: Social Intermediation Services 117 Note 119 References 119 Appendix 121 Index 125 Boxes 2.1 The framework in practice 39 3.1 Monetary versus nonmonetary measures of chronic poverty 55 4.1 Ethnicity and chronic poverty in rural Guatemala 74 4.2 Connectivity pays off: Reducing poverty in rural Peru 78 4.3 Improving institutions, reducing poverty in Colombia 81 4.4 Does increased public expenditures reduce crime and chronic poverty? Evidence from Mexico 85 4.5 Does poverty cause stress? 88 xii CONTENTS 5.1 Does increased public spending help reduce poverty? Evidence from Peru 103 5.2 Coordinating poverty-reduction efforts in Brazil, Mexico, and Peru 106 Figures O.1 Poverty, vulnerability, and the middle class in Latin America and the Caribbean, 2000–12 2 O.2 From chronic poverty to upward mobility: Inputs and the process of emerging from poverty 3 O.3 Chronic poverty and downward mobility in selected countries in Latin America and the Caribbean, 2004–12 6 O.4 Subnational chronic poverty rates in Latin America and the Caribbean, 2012 7 O.5 Concentration of chronic poverty in Latin America and the Caribbean, 2012 8 O.6 Chronic poverty in rural and urban areas in selected countries in Latin America and the Caribbean, 2012 9 O.7 Median income in selected countries in Latin America and the Caribbean, 2012 10 O.8 Number of labor income earners in household in selected countries in Latin America and the Caribbean, by poverty group, 2012 11 O.9 Correlation between chronic poverty and access to services at subnational level in Latin America and the Caribbean, 2012 14 O.10 Role of endowments in explaining income differences between the chronically poor and people who escaped poverty in Latin America and the Caribbean, 2012 15 O.11 Correlation between chronic poverty and expectations in Latin America and the Caribbean, 2010 16 O.12 Level of optimism about the next 12 months by the chronically poor, people who escaped poverty, and people who were never poor in Latin America and the Caribbean, 2010 17 O.13 Correlation between educational aspirations of children ages 12–15 and their peers in Latin America and the Caribbean, 2011 18 O.14 Intergenerational transmission of chronic poverty in selected countries in Latin America and the Caribbean, 2012 19 O.15 Number of new social assistance programs in selected countries in Latin America and the Caribbean, 1990–99 and 2000–11 22 1.1 GDP per capita and income inequality in Latin America and the Caribbean, 2000–12 28 1.2 Poverty, vulnerability, and the middle class in Latin America and the Caribbean, 2000–12 29 1.3 Decomposition of changes in poverty in Latin America and the Caribbean by sources of income, 2003–12 30 CONTENTS xiii 2.1 From chronic poverty to upward mobility: Inputs and the process of emerging from poverty 39 3.1 Chronic poverty and downward mobility in selected countries in Latin America and the Caribbean, 2004–12 54 B3.1.1 Correlation between estimates of chronic poverty and multidimensional poverty in selected countries in Latin America and the Caribbean 56 3.2 Subnational chronic poverty rates in Latin America and the Caribbean, 2012 57 3.3 Concentration of chronic poverty in Latin America and the Caribbean, 2012 59 3.4 Chronic poverty in rural and urban areas in selected countries in Latin America and the Caribbean, 2012 61 3.5 Correlation between per capita GDP growth and chronic poverty in selected countries in Latin America and the Caribbean, 2012 63 3.6 Initial (2004) income distribution of chronically poor and people who escaped poverty by 2012 in Peru 64 3.7 Median income in selected countries in Latin America and the Caribbean, 2012 64 3.8 Number of labor income earners in household in selected countries in Latin America and the Caribbean, by poverty group, 2012 66 3.9 Female labor force participation and chronicity of poverty in selected countries in Latin America and the Caribbean, 2012 67 3.10 Correlation between chronic poverty and sector of employment in selected countries in Latin America and the Caribbean, 2012 68 3.11 Correlation between chronic poverty and rate of growth of formal sector in Latin America and the Caribbean, 2012 69 3.12 Correlation between chronic poverty and dependence on nonlabor income in Latin America and the Caribbean, 2012 69 B4.1.1 Chronic poverty and ethnicity in rural Guatemala, 2011 74 4.1 Mean standardized test scores of richest and poorest children in urban and rural areas of five countries in Latin America 75 4.2 Intergenerational transmission of chronic poverty in selected countries in Latin America and the Caribbean, 2012 76 4.3 Correlation between chronic poverty and access to services at subnational level in Latin America and the Caribbean, 2012 77 B4.2.1 Effect of mobile phone coverage on poverty in rural Peru 78 4.4 Access to services in rural Guatemala, 2011 79 4.5 Correlation between chronic poverty and incidence of weather shocks in Latin America and the Caribbean, 2012 83 4.6 Location of and food shortages from weather shocks in rural Guatemala, 2011 84 xiv CONTENTS 4.7 Role of endowments in explaining income differences between the chronically poor and people who escaped poverty in Latin America and the Caribbean, 2012 86 4.8 Correlation between chronic poverty and expectations in Latin America and the Caribbean, 2010 90 4.9 Level of optimism about next 12 months by the chronically poor, people who escaped poverty, and people who were never poor in Latin America and the Caribbean, 2010 91 4.10 Expectations about next 12 months by the chronically poor in selected countries in Latin America and the Caribbean, 2010 92 4.11 Educational aspirations of Peruvian children and their parents 93 4.12 Correlation between educational aspirations of children ages 12–15 and their peers in Latin America and the Caribbean, 2011 93 5.1 Number of new social assistance programs in selected countries in Latin America and the Caribbean, 1990–99 and 2000–11 102 5.2 Spending on social assistance in selected countries in Latin America and the Caribbean, 2000–10 102 B5.1.1 Median public spending per capita in “improved” and chronically poor districts of Peru 103 5.3 Adherence to treatment for tuberculosis among depressed and nondepressed patients in Peru 113 5.4 Effect of socioeconomic intervention on vaccination against tuberculosis of children in Peru 114 5.5 Effect on income of social interaction with local leaders in Nicaragua 116 Map 3.1 Chronic poverty in Peru, by district, 2012 58 Tables O.1 Movement in and out of poverty in Latin America and the Caribbean between 2004 and 2012 5 O.2 Characteristics of people in Latin America and the Caribbean who were chronically poor, who escaped poverty, and who were never poor, 2004 13 2.1 Poverty estimates for Peru based on actual and synthetic panel data, 2004–05 46 3.1 Movement in and out of poverty in Latin America and the Caribbean between 2004 and 2012 54 3.2 Characteristics of the chronically poor in urban and rural areas of Latin America and the Caribbean, 2012 62 4.1 Characteristics of people in Latin America and the Caribbean who were chronically poor, who escaped poverty, and who were never poor, 2004 73 CONTENTS xv 4.2 Population size and remoteness between districts that were never poor, escaped poverty, and were chronically poor in Peru 80 5.1 Social development and institutional arrangements in Latin America and the Caribbean 105 A.1 Household surveys used in the analysis 121 A.2 Movement into and out of poverty in Latin America and the Caribbean, by country, 2004–12 123 xvi CONTENTS Foreword Latin America and the Caribbean’s story in the 2000s was one of rapid progress for many. More than 70 million people moved out of poverty and started on the path to a better life for themselves and their families. The middle class grew at an impres- sive rate, and despite recent slowdowns in growth, remains a large segment of Latin American society. Income inequalities shrank between those at the top and those at the bottom, albeit not nearly enough. Despite these improvements, still too many Latin Americans are trapped, unable to see the progress that their compatriots have experienced. One in four people is still living in poverty. And of the millions who have moved out of poverty, most are stuck between poverty and the middle class, making up a growing vulnerable class that remains at risk of losing their hard-won gains. One very concerning group is the focus of this book: the nearly 130 million chronic poor in Latin America and the Caribbean. These are the one in five people in the region who have never known anything but poverty. They have not benefitted from the rising tide driven by the growth over the past fifteen years, and thus have thus been left behind by policies and programs that have otherwise been effective in improving the lives of millions. These are the poorest of the poor, concentrated in communities that share the same issues, and for whom an escape from poverty seems unattainable. Chronic poverty stems from a number of things—and for each new generation born to chronically poor parents, the way out becomes more difficult. Communities across Latin America and the Caribbean suffer from a lack of access to basic needs and services like water, electricity, healthcare, and education. Local institutions do not have the capacity to serve the needs of the poorest. The chronically poor, who often suffer earliest and worst following shocks and disasters are also less likely to have risk insurance or safety nets that can support them through hard times. When added together, all of these factors can lead to lowered aspirations and a depressed state of mind, completing a vicious circle that makes it nearly impossible for the chronically poor to even dream of escaping the conditions in which they live. xvii To tackle the issue of chronic poverty in Latin America and the Caribbean and around the world, governments and institutions will need to think and act differently, not relying on the usual approaches that have benefitted many but passed others by. This book points to a few areas where policies and programs can have more and better impact. First, improving the enabling environment for the chronically poor to succeed, balancing direct support to poor families with critical improvements in their communities as a whole. It is not enough for people to have skills or education; they need to be able to find good jobs, have access to universal quality health care, and build their businesses in safe neighborhoods. Second, coordinating poverty reduction efforts to maximize their impact. The proliferation of social programs has contributed to the advances that many of Latin America and the Caribbean’s poorest have seen, but in order to reach those who have been left behind, it will be crucial to ensure that these programs are part of a bigger picture that is aligned across areas, agencies, and partners. And third, considering the state of mind and low aspirations of the chronically poor when planning programs and policies. A number of promising interventions have been successful in reaching those most entrenched in poverty, and the World Bank Group has a growing body of work in this area following on the 2015 World Development Report on Mind, Society, and Behavior as well as on the emerging find- ings from this book. Even more encouraging, many of these behavioral interventions do not require a complete reinvention of processes. Often, a small addition or tweak to an existing program can mean the difference between impact and the status quo. In Colombia, for example, changing the timing on a conditional cash transfer helped get more children into school. In Peru, removing the stigma around banking for women has increased their savings and helped get them into commercial activities. We are heartened by the success of these and other efforts, but recognize that for progress to take hold, small interventions must become impactful policies, and they must be underpinned by growth that is inclusive, improved basic services and more equal opportunities, and social insurance systems that protect the poor from risk and shocks. With this book, the World Bank Group hopes to set the stage for governments in Latin America, and indeed in other regions, to recognize the unique situation of those living in chronic poverty, and to take the steps needed to help them move them- selves out. This will be crucial not only to the region’s prosperity in the face of slower growth, but also to the world’s efforts to end extreme poverty by 2030 and improve the lives of the least advantaged in every country. It is our sincere hope that this evidence and these interventions can be used for progress, to help pave the way for today’s chronically poor children to break the cycle into which they were born, giving hope for a more prosperous future for Latin America and the Caribbean. Jorge Familiar Calderon Ana Revenga Vice President, Latin America and the Caribbean Senior Director, Poverty and Equity xviii FOREWORD Acknowledgments This book was prepared by a team led by Renos Vakis and co-managed with Leonardo Lucchetti and Jamele Rigolini. The team was ably assisted by Oscar Barriga Cabanillas, who also contributed substantially to some of the background material. Important additional contributions were made by Javier E. Baez, Kiyomi E. Cadena, Adriana Camacho, Wendy Cunningham, Hai-Anh H. Dang, Tania Valeria Diaz Bazan, Maria Marta Favara, Eugenia Genoni, James Robinson, Carlos Rodriguez Castelan, Veronica Silva Villalobos, and Carolina Trivelli. The work was conducted under the general guidance of Augusto de la Torre, chief economist of the Latin America and Caribbean Region. The team was fortunate to receive advice and guidance from a number of peer reviewers and advisors, includ- ing Omar Arias, Michael Carter, Louise Cord, German Freire, Michele Gragnolati, Jesko Hentchel, Ede Jorge Ijjasz-Vasquez, Peter Lanjouw, Luis Felipe López-Calva, Andy Mason, Jaime Saavedra, and Carolina Trivelli. Additional insights and sugges- tions from Tito Cordella, Margaret Grosh, Peter Lanjouw, Daniel Lederman, and Kyla Wethli are gratefully acknowledged. The team also benefited from comments received in two authors’ workshops. Raka Banerjee provided editorial assistance in English. Susy Carrera translated the manuscript from Spanish. Sara Burga, Cristina Cifuentes, Carmen Contreras Buse, Ruth Delgado, Karem Edwards, Bernarda Erazo, Erika Bazan Lavanda, Jacqueline Larrabure Rivero, Amparo Lezama, Tatianna Guerrante Schlottfeldt, and Jessica Terry provided unfailing administrative support. xix About the Authors Leonardo Lucchetti is an economist in the Poverty and Equity Global Practice of the World Bank. His research focuses on poverty and inequality reduction, income mobil- ity, chronic poverty, and the impact of shocks on households’ welfare. Before joining the Poverty and Equity Global Practice, he worked for the Social Protection Unit of the Latin America and the Caribbean Region of the World Bank; the Centro de Estudios Distributivos Laborales y Sociales (CEDLAS), in Argentina; and the Economic and Financial Affairs Unit of the European Commission, in Brussels. He has published in leading academic journals, including the Journal of Development Economics, the Journal of Agricultural Economics, and the Journal of Economic Inequality. He holds a PhD in economics from the University of Illinois at Urbana-Champaign. Jamele Rigolini is a lead economist in the World Bank’s Latin America and the Caribbean Region (LAC), where he coordinates and supervises the human develop- ment portfolio for the Andean countries. His areas of expertise include social protec- tion, human development, labor markets, innovation, political economy, and economic growth. Before joining LAC, he worked for the East Asia and Pacific Region’s social protection unit, where he managed cross-sectoral lending projects and advisory activities; taught economics at the University of Warwick; and worked for the Inter- American Development Bank, the International Union for Conservation of Nature, and McKinsey. He has published in leading academic journals, including the Journal of Public Economics, the Journal of Development Economics, Economics & Politics, Economic Letters, and World Development. He holds a diploma in physics from the Swiss Federal Institute of Technology (ETH) in Zurich and a PhD in economics from New York University. Renos Vakis is a lead economist in the Poverty and Equity Global Practice as well as a member of the Living Standards Measurement Study (LSMS) team in the Development Research Group of the World Bank. He leads the Behavioral Initiative for Poverty Reduction and Equity, which aims to integrate behavioral science in the xxi design of social policies, and is involved in the design of impact evaluations of anti- poverty interventions. His research focuses on poverty reduction, equity, and gender, with an emphasis on understanding how behavioral insights can influence policies and outcomes. He has written extensively on poverty dynamics and mobility, risk management, social protection, market failures, and rural development, especially in Latin America and South Asia. He has taught at the Paul H. Nitze School of Advanced International Studies (SAIS) of Johns Hopkins University. He holds a PhD from the University of California, Berkeley. xxii ABOUT THE AUTHORS Abbreviations CEDLAS Center for Distributional Labor and Social Studies GDP gross domestic product LAC Latin America and the Caribbean MPI Multidimensional Poverty Index SEDLAC Socio-Economic Database for Latin America and the Caribbean All amounts are presented in U.S. dollars unless otherwise indicated. xxiii Overview A Successful Decade with Challenges Ahead By most measures, the 2000s were one of the most impressive decades of eco- nomic development in Latin America and the Caribbean (LAC) in recent his- tory. The region’s gross domestic product (GDP) per capita grew consistently at an average annual rate of 2.5 percent between 2000 and 2012 (except in 2009). Inequality narrowed substantially, with the regional Gini coefficient for per capita income falling by an unprecedented 5 basis points, from 0.57 in 2000 to 0.52 in 2012 (Alvaredo and Gasparini 2015). Sustained economic growth and substantial reductions in income inequality led to large increases in the incomes of people at the bottom of the income distri- bution (Cord and others 2015; Cord, Genoni, and Rodríguez-Castelán 2015). As a result, total poverty decreased by more than 16 percentage points within a single decade (from 41.6 percent of the population in 2003 to 25.3 percent in 2012), and extreme poverty was cut in half (from 24.5 percent to 12.3 percent) (figure O.1).1 Overall, some 70 million people moved out of poverty, the largest poverty reduc- tion in the region in decades. Another facet of the dramatic reduction in poverty was the emergence of a large middle class, which increased from about 23 percent of the population in 2003 to 34 percent in 2012. In 2010, for the first time the number of middle class people in the region exceeded the number of people living in poverty (see figure O.1).2 These improvements notwithstanding, LAC is not yet a middle-class region. One in four people is poor. Some of them emerged from poverty but experienced shocks that pushed them back into it (the “transitory” poor), others never escaped (the “chronically” poor). The chronically poor did not benefit much from the impressive growth rates of the 2000s and may have fallen through the cracks of the social assistance system. 1 FIGURE O.1 Poverty, vulnerability, and the middle class in Latin America and the Caribbean, 2000–12 50 40 Percent of population 30 20 10 0 01 00 02 03 04 06 08 09 10 11 12 05 07 20 20 20 20 20 20 20 20 20 20 20 20 20 Moderately poor (per capita income of less than $4 a day) Vulnerable (per capita income of $4–$10 a day) Extremely poor (per capita income of less than $2.50 a day) Middle class (per capita income of $10–$50 a day) Source: World Bank 2014b, using SEDLAC data (CEDLAS and World Bank). Note: Estimates are population-weighted averages of country estimates. Poverty lines and incomes are expressed in 2005 purchasing power parity dollars. The prospects of the chronically poor escaping poverty in the near future are weak (Arim and others 2013). GDP growth has slowed significantly, from about 6 percent in 2010 to an estimated 0.8 percent in 2014 (World Bank 2014a). Improved labor market prospects may therefore not be sufficient to pull the chron- ically poor out of poverty. Investment in social assistance has been significant—but it remains modest compared with other parts of the world. What can be done to reduce chronic poverty in the region? To provide policy makers with guidance, this book identifies who the chronically poor are, explores their lives, and evaluates factors that may prevent them from escaping poverty. It then identifies design elements that could be integrated into development and social assistance policies in order to support them more effectively. 2 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN A Framework for Studying Chronic Poverty Studying chronic poverty involves both conceptual and empirical challenges. Conceptually, chronic poverty involves both a welfare and a time dimension. It is therefore complex in nature. This book uses a simple conceptual framework to understand what characterizes and drives chronic poverty. The framework draws from the literature on poverty traps (Carter and Barrett 2006, Cord and López- Calva 2012) and Sen’s (1999) capabilities approach. Figure O.2 summarizes this framework. It divides inputs into endowments, the enabling context, and the state of mind. Poverty exists and persists because of constraints that prohibit the optimal accumulation and use of existing endow- ments, such as skills and physical assets. An enabling context is also a necessary input: Chronic poverty may exist not because endowments are low per se, but because contextual factors affect the returns to those endowments differentially. Households with similar endowments that live in different contexts may therefore face different trajectories in life. The third input is the state of mind. An emerging field of research studies the ways in which it affects people’s life trajectories and the resulting implications for social policies (see, for example, World Development Report 2015: Mind, Society, and Behavior). The pressure of poverty affects people’s state of mind and decision- making process in a wide range of ways, causing them to place greater emphasis on FIGURE O.2 From chronic poverty to upward mobility: Inputs and the process of emerging from poverty Endowments Enabling context State of mind • Skills • Markets • Aspirations • Physical assets • Services • Psychological • Risk welfare • Institutions Converting action into welfare Transforming outcomes Upward choice into action Mobility Considering opportunities Action Choice Poverty Inputs Process Outcomes OVERVIEW 3 short-term rather than long-term problems, for example, even if focusing on long- term issues would lead to better outcomes over time (Banerjee and Mullainathan 2010). The fact that poor people must devote much of their mental resources to tackling survival problems can cause them to underinvest in their children’s upbringing, perpetuating poverty across generations. Aspirations—the presence of forward-looking goals and a desire to attain them (Locke and Latham 2002)—are also critical. Weak aspirations can reduce the capability to act in order to achieve a desired outcome, as Sen (1999) shows. They are associated with a limited temporal horizon, which could translate into failing to save, opting out of good investments, and engaging in harmful behaviors. “Hopelessness destroys both the will and the ability to invest in one’s future and oneself,” notes Duflo (2012). Poverty may generate poverty. Inputs are relevant insofar as they affect the process of upward mobility (see figure O.2). Understanding this process is therefore also important. The first step in emerging from poverty is to engage in decision making. Consider, for example, the decision-making process involved in choosing to participate in an income- generation opportunity. Poor people may choose not to do so for several reasons, including not knowing the option exists; concluding that their abilities would pre- vent them from being able to take advantage of the opportunity; or having low aspirations, despite their best efforts. The context could amplify these effects through social norms. Peer effects, for example, may accentuate low informational flows and the formation of aspirations. The second step in moving out of poverty is transforming choices into actions. The differences between the two stages are subtle but important: Considering participation begins a lengthy process in which an individual evaluates costs and benefits before deciding whether to take action. Endowments, context, and the state of mind all affect the decision to do so. The last part of the process of upward mobility is “conversion” (the more tra- ditional part of the overall process). In general, the productivity of an investment will depend on individual ability and effort; endowments such as skills and assets are therefore central. An enabling context will also affect the returns. For example, an uninsured weather shock may reduce yields by destroying part of the harvest. The state of mind may also affect outcomes at this stage of the process: Positive peer effects and social norms (themselves partly driven by context) may improve motivation and effort, which in turn enhance the likelihood of success (Macours and Vakis 2014). The framework remains silent on the ways in which these elements interact. Low levels of endowments, a disabling context, or a defeatist state of mind can affect all stages of the process, giving rise to a state of chronic poverty induced by different channels. These interactions highlight the policy challenges involved in both identifying the chronically poor and designing effective policies for pulling them out of poverty. 4 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Measuring Chronic Poverty in the Absence of Longitudinal Data Studying chronic poverty in LAC is difficult because of the almost complete lack of longitudinal data. An important contribution of this book is therefore its meth- odological approach to the measurement of chronic poverty. To capture chronic poverty between 2004 and 2012, we rely on an innovative technique originally proposed by Dang and others (2014) and improved by Dang and Lanjouw (2014) that uses information contained in repeated cross-sectional data sets to construct “synthetic” panels. The approach allows us to use cross- sectional data to define a household as chronically poor if it was poor in both 2004 and 2012. Validation exercises on the technique suggest that although synthetic panels are not a substitute for actual panels, they are among the most accurate remedies in the absence of longitudinal data. Five Stylized Facts about Chronic Poverty in Latin America and the Caribbean Five stylized facts characterize chronic poverty in the region. Together with our conceptual framework, they structure the policy discussion. Stylized Fact 1: One of Five People in Latin America and the Caribbean Lives in Chronic Poverty In 2012, 21.6 percent, or one in five poor people in the region, had also been poor in 2004 (table O.1). This figure implies that 130 million people—about half of all poor people in 2012—were chronically poor. The rest of the poor—about 8 per- cent of the population that was nonpoor in 2004—fell into poverty over this period. These figures reveal that despite extraordinary success in reducing poverty in the region, many people were left behind, either staying or becoming poor. TABLE O.1 Movement in and out of poverty in Latin America and the Caribbean between 2004 and 2012 (percent of population) 2012 Poor Nonpoor Total 2004 Poor 21.6 23.4 44.9 Nonpoor 4.2 50.9 55.1 Total 25.7 74.2 100.0 Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are population-weighted averages of country-specific estimates, which are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. OVERVIEW 5 FIGURE O.3 Chronic poverty and downward mobility in selected countries in Latin America and the Caribbean, 2004–12 60 Percent of population 40 20 0 Ca ric co Ho ua la Ni bia Co hile na Pa ica ma Gu uras Ec lic eru Ar uay Pa y Re r Co dor ia il be nd an do ua ma az liv b i rag nti the Am Mex R lom an na rib a a P C pu nic lva ua ug rag r Bo nd B ate ge sta ca Ur mi Sa e Do El tin La Downwardly mobile Chronically poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates of poverty at the regional level are population-weighted averages of country-specific estimates, which are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Total poverty does not necessarily match country-specific poverty rates, which are estimated based on the synthetic panel approach, which uses a subpopulation of households with adult household heads. Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. Chronically poor people are people whose income remained below $4 between 2004 and 2012. Downwardly mobile people are people who were nonpoor in 2004 but poor in 2012. Chronic poverty and downward mobility both vary considerably across countries. Uruguay, Argentina, and Chile have the lowest rates of chronic poverty, with rates of about 10 percent (figure O.3). Nicaragua, Honduras, and Guatemala have very high rates of chronic poverty, ranging from 37 percent in Nicaragua to 50 percent in Guatemala. Stylized Fact 2: Chronic Poverty Tends to Be Geographically Concentrated Geography is an extremely important factor for understanding chronic poverty. Within Brazil, for example, Santa Catarina has a chronic poverty rate of about 5 percent, which is lower than the national average of 20 percent (figure O.4). This rate is close to the average for Uruguay, the LAC country with the lowest rate of chronic poverty. By contrast, about 40 percent of the population in Ceará, Brazil, is chronically poor—twice the Brazilian average. This rate is close to the average for Honduras, one of the countries with the highest chronic poverty rates in the region. Looking exclusively at poverty rates provides only part of the picture, how- ever, because a large number of the chronically poor reside in densely popu- lated areas with relatively low rates of chronic poverty (for example, regional 6 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE O.4 Subnational chronic poverty rates in Latin America and the Caribbean, 2012 100 90 Percent of population that is chronically poor 80 70 60 50 Ceará (Brazil) 40 Latin America and the Caribbean 30 Santa Catarina 20 (Brazil) Tocantins (Brazil) 10 0 0 20 40 60 80 100 Ranking Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Horizontal axis presents a normalized ranking of regions (out of a total of 168) based on their chronic poverty rates. Selection of subnational regions varies depending on level of representativeness of surveys. Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. and national capitals). Figure O.5 presents the cumulative distribution (Lorenz curve) of chronically poor households across subnational regions (the vertical axis), sorted by the contribution of each region in terms of the absolute number of chronically poor (the horizontal axis). It shows that 20 of the 187 regions considered contain half of the region’s chronically poor population. The regions where the majority of the chronically poor reside are not nec- essarily the regions with the highest rates of chronic poverty. In Mexico, for instance, the incidence of chronic poverty in the Distrito Federal is about equal to the LAC regional average, but its large population means that it is home to almost 3 percent of the region’s chronically poor, making it the sixth-largest con- tributor in the region. The rate of chronic poverty is higher in Baja California (Mexico), but it is home to only 0.2 percent of the region’s chronically poor, because its population is much smaller. Stylized Fact 3: Chronic Poverty Is As Big a Problem in Urban Areas as in Rural Areas In every country in LAC, the rate of chronic poverty is higher in rural areas than urban areas (figure O.6, panel a). Although rural areas have higher rates of chronic poverty, in at least five countries (Chile, Brazil, Mexico, Colombia, and OVERVIEW 7 FIGURE O.5 Concentration of chronic poverty in Latin America and the Caribbean, 2012 100 Distrito Federal 90 (Mexico) Cumulative distribution of percent of population 80 70 Chiapas (Mexico) that is chronically poor 60 50 40 30 20 Baja California (Mexico) 10 0 0 20 40 60 80 100 Ranking Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Selection of subregions varies depending on level of representativeness of surveys. the Dominican Republic) the number of chronically poor people is greater in urban areas (figure O.6, panel b). In Brazil, for example, for every one chronically poor household in rural areas, two live in urban areas. The number of chronically poor is about the same in urban and rural areas in many other countries The rural poor tend to face worse conditions than the urban poor. Chronically poor households in urban areas tend to be smaller in size, have household heads with more years of education, and face better economic opportunities than their counterparts in rural areas. Stylized Fact 4: Economic Growth Was Not Sufficient to Lift the Chronically Poor out of Poverty The impressive economic growth of the 2000s had little positive impact on the lives of the chronically poor, for two main reasons. First, countries with the high- est rates of chronic poverty grew the least. Guatemala, for example, where about half the population was chronically poor in 2012, grew less than 1 percent a year. In contrast, Panama, where the rate of chronic poverty was 20 percent, grew 6 percent a year. Second, chronically poor households tend to be poorer than poor households that escape poverty. The median initial per capita income of people who were 8 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE O.6 Chronic poverty in rural and urban areas in selected countries in Latin America and the Caribbean, 2012 a. Share of rural and urban population that is chronically poor 80 70 60 Percent of urban population that is chronically poor 50 40 Guatemala Nicaragua Honduras 30 Dominican Republic Colombia 20 El Salvador Ecuador Paraguay Peru Brazil 10 Mexico Bolivia Chile Costa Rica Panama 0 0 10 20 30 40 50 60 70 Percent of rural population that is chronically poor b. Urbanization rate and ratio of number of people who are chronically poor in urban versus rural areas 7 6 Chile Ratio of number of chronically poor 5 people in urban to rural areas 4 3 2 Brazil Costa Rica Dominican Republic Mexico Paraguay Bolivia Colombia 1 Honduras Nicaragua Ecuador Guatemala El Salvador Peru 0 Panama –1 –2 45 55 65 75 85 95 Urbanization rate (percent) Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Vertical line indicates average urbanization rate in the region. OVERVIEW 9 poor in 2004 but not poor in 2012 was $3 a day (in 2005 purchasing power par- ity terms); it rose to $6 a day in 2012, an annualized growth rate of 9.0 percent (figure O.7, panel a). The incomes of the chronically poor were lower in 2004 ($1.50 a day) and rose much less (to just $2.70 a day, an annualized growth rate of 7.9 percent). These trends are evident in almost all of the 17 countries analyzed. FIGURE O.7 Median income in selected countries in Latin America and the Caribbean, 2012 a. People who exited poverty 8 Daily per capita income power parity dollars) (2005 purchasing 6 4 2 0 Ho ala Ni ico mi Co le Ec a na ia Re ia Pa a ma Co lic Me s Pa eru Sa y y ate r Ur or il a Gu do Ric u El ua ua i liv b az Ch b ur d nti rag x m nic lom na P pu ua lva Bo rag ug Br nd ge sta ca Ar an Do b. People who remained poor 8.0 Daily per capita income 6.0 power parity dollars) (2005 purchasing 4.0 2.0 0 –2.0 Re a Ho ala Ni ico Co ile na mi Cos bia ia Pa a ma Ec lic as Pa ru Sa y y Gu dor Ur or il an Ric u El gua ua liv az Ch b Pe ur d nti rag x m lom na pu ua lva Bo ug Me Br nd ate ge nic ta ra ca Ar Do Initial income (median) Change in median income Median income in Latin America Median income in Latin America and the Caribbean, 2012 and the Caribbean, 2004 Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Figure is updated version of figure 4.5 in Ferreira and others (2012). Note: Figure shows lower-bound mobility estimates using the Dang and others (2014) technique. Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. 10 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Stylized Fact 5: The Chronically Poor Have Limited Income Opportunities Labor income was by far the most powerful driver behind the strong reduction in poverty between 2004 and 2012. The chronically poor face greater barriers to entering the labor force and rely relatively more on nonlabor incomes. They are also more active than the nonpoor in low-productivity and subsistence sectors. Weak labor earnings partly explain why people remain chronically poor. In every LAC country there were fewer labor income earners among chroni- cally poor households than there were among households that escaped poverty or were not poor in either 2004 or 2012 (figure O.8). On average, just 1.3 adults generate income in chronically poor households—30 percent fewer than the 1.7 adults in other households. In contrast, there does not appear to be a large differ- ence in the number of labor income earners in households that escaped poverty and households that were never poor. A similar story emerges for female labor market participation. Gender equal- ity is crucial for poverty reduction, because greater economic opportunities for women can enhance productivity gains and increase households’ welfare (World Bank 2013). LAC increased female labor force participation more than any other FIGURE O.8 Number of labor income earners in household in selected countries in Latin America and the Caribbean, by poverty group, 2012 2.5 Average number of labor income 2.0 earners in household 1.5 1.0 0.5 0 a ua Ca ica via nic Ecu a Pa xico Ho ala bia ile Ar blic na ca s El Peru Me r Co ay Pa uay Re or Co zil be nd do Ni ura Ric m Ch an ad a rag nti u i m lom na rib a l pu Ur n lva Br the Ame Bo rag ug nd a ate sta ge Sa Gu r tin mi La Do Never poor Exited poverty Chronically poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are based on 2004 and 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. OVERVIEW 11 region since the 1980s, with more than 70 million women entering the labor force between 1984 and 2006 (World Bank 2012). This performance does not appear to have benefited the chronically poor, however: Female labor force participation lagged significantly among chronically poor households in almost every country in LAC in 2012, with participation rates 16 percentage points lower than among the nonpoor. In addition to more limited labor force participation, the chronically poor tend to work in lower-productivity sectors. Subnational regions with high concentra- tions of people employed in agriculture also tend to have higher rates of chronic poverty. In contrast, regions with lower rates of chronic poverty are more likely to employ larger numbers of people in high-tech industry, services, construction, and retail. What limits the income-earning opportunities of the poor? Particularly impor- tant are the interactions between endowments and the enabling context, as well as the role of the state of mind in limiting aspirations and the process of upward mobility. The role of endowments and context The initial endowments of the chronically poor are markedly different from those of the nonpoor but similar to the initial endowments of people who escaped pov- erty (table O.2). The primary difference between people who escape poverty and people who do not is access to services. In 2004, only 79 percent of the chronically poor had access to water, compared with 89 percent of people who escaped pov- erty and 95 percent of those who were never poor. Only 58 percent of the chroni- cally poor had a minimum level of assets, compared with 78 percent of people who escaped poverty and 90 percent of those who were never poor.3 Subnational regions with lower rates of access to clean water, sewerage systems, or sanitation facilities tend to have higher rates of chronic poverty. The chronically poor are also more likely to reside in regions with lower coverage of electricity and mobile communications (although these correlations are weaker, partly because of improvements in the accessibility of both services during the last decade). It also appears that services complement one another as drivers of chronic poverty: Households with no access to three or more basic services are more likely to reside in regions with higher levels of chronic poverty (panel f in figure O.9). The context in which people live affects the returns they reap from a given endowment. Returns to endowments matter at least as much, if not more, than endowments themselves (figure O.10). Supporting individuals may therefore not be sufficient if the context does not provide them with the chance to embrace opportunities. Without an enabling context, chronic poverty may prevail because people will be unable to use their endowments. Regional development efforts must therefore be implemented in conjunction with social policies that focus on improving context. 12 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN TABLE O.2 Characteristics of people in Latin America and the Caribbean who were chronically poor, who escaped poverty, and who were never poor, 2004 Characteristic Chronically poor Escaped poverty Never poor Demographics Household size 4.8 4.9 3.6 Age 0–12 2.0 1.9 0.9 0–15 2.4 2.3 1.1 13–18 0.7 0.8 0.5 19–70 2.1 2.2 2.2 70+ 0.1 0.1 0.1 Age of household head 37.8 40.0 40.8 Education Years of education of household head 6.0 5.1 9.2 Access to services Drinking water 0.8 0.9 1.0 Shelter 0.7 0.8 0.9 School attendance 0.9 1.0 1.0 At least five years of schooling 0.9 0.9 1.0 Sanitation 0.5 0.7 0.9 Electricity 0.9 1.0 1.0 Assets 0.6 0.8 0.9 Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are population-weighted averages of country-specific estimates, which are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. Chronically poor = poor in both 2004 and 2012. Escaped from poverty = poor in 2004 but not 2012. Never poor = not poor in either year. Access to services is not the only important feature of this context. All external factors that influence returns to endowments and households’ decisions should be considered. Two key factors are institutions and uninsured risk. Institutions The type and quality of national and local institutions have a strong impact on welfare in general and chronic poverty in particular. It is easier for the poor to escape poverty if their voices are heard and taken into consideration. The poor must therefore be represented by leaders who understand their needs and the OVERVIEW 13 FIGURE O.9 Correlation between chronic poverty and access to services at subnational level in Latin America and the Caribbean, 2012 a. Access to clean water b. Access to sanitantion 1.0 1.0 Regional access to Regional access to 0.8 0.8 clean water sanitantion 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Chronically poor (%) Chronically poor (%) c. Access to sewer d. Access to mobile phone 1.0 Regional access to 1.0 Regional access to 0.8 mobile phone 0.8 0.6 0.6 sewer 0.4 0.4 0.2 0.2 0 0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Chronically poor (%) Chronically poor (%) e. Access to electricity f. Overall access to services Regional lacks of at least 1.0 1.0 Regional access to 0.8 0.8 3 services (%) electricity 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Chronically poor (%) Chronically poor (%) 95% Confidence interval Fitted values Cronically poor by region (%) Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Figures come from 2012 surveys (or nearest year in cases in which 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. Selection of regions varies depending on level of representativeness of surveys. challenges they face. Local governments, services, and social programs must be staffed by civil servants who discharge their duties in ways that address the needs of the poor rather than pose additional obstacles. The foundation for such an institution is an inclusive social contract that recognizes every citizen as equal and supports efforts to provide equal oppor- tunities for all. Without such a foundation, it is nearly impossible to eradicate poverty. 14 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE O.10 Role of endowments in explaining income differences between the chronically poor and people who escaped poverty in Latin America and the Caribbean, 2012 30 25 Percent of income differential explained 20 15 10 5 0 –5 –10 –15 Ni ivia ua Sa l r o Ho or as y Pa a ma na Re le Co blic Pa ica y ru El razi do ua ua xic bi i d Pe Ch ur rag nti R lom na lva ua l pu ug rag Bo Me B nd ge sta Ec ca Ur Co Ar an nic mi Do Endowments Returns Unexplained component Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Endowments include household composition, assets, and education. Figures are based on 2012 surveys (or nearest year in cases in which 2012 data were not available). A strong and inclusive social contract is insufficient to guarantee strong institu- tions, however. In order to operate effectively and support citizens, institutions must function well at three levels: There must be sufficient bureaucratic capacity to deliver good services, the state must have the credibility to enforce the legal framework and justice system, and institutional processes must be transparent and follow clear accountability rules (Fukuyama 2011). The importance of bureaucratic capacity is often greatly underestimated. Social programs often have low impact on the chronically poor because the poor reside in regions with less capacity to implement social programs. Loayza, Rigolini, and Calvo-González (2014) study the ability of Peruvian municipalities to spend addi- tional budget streams generated by mining revenues. They find that although poorer municipalities spend a larger share of the additional budget allocated to them (signifying greater need), all else being equal, municipalities with lower average educational attainments have less ability to spend the additional budget. Galasso (2011) studies the impact of Chile Solidario, a program that tries to tailor assistance to the needs of the extremely poor. She finds that the impacts of the pro- gram are much greater when extremely poor families are assisted by social workers who receive positive performance assessments. OVERVIEW 15 Uninsured risk The chronically poor face disproportionately high levels of uninsured risk. The prevalence of natural disasters, for example, is higher in countries with higher rates of chronic poverty. In the absence of some form of insurance and savings, shocks—particularly repeated ones—can exacerbate chronic poverty. Shocks can reduce welfare by directly depleting assets, triggering coping mechanisms that involve the decapitalization of productive assets and the depletion of human capital. They can lead to slow recovery or even poverty traps, reducing welfare over the long term. The importance of state of mind The depressed state of mind and lower aspirations of the chronically poor can be major obstacles to upward mobility. This relationship likely works both ways: Poverty leads to a depressed state of mind, which in turn lowers aspirations and the chances for upward mobility, potentially perpetuating a vicious cycle of poverty. Expectations about the future are lower in countries in which the incidence of chronic poverty is higher (figure O.11). Guatemala, Honduras, and Nicaragua— the countries in the region with the highest chronic poverty rates—are also among FIGURE O.11 Correlation between chronic poverty and expectations in Latin America and the Caribbean, 2010 Brazil Uruguay Paraguay 2.6 Average level of expected well-being Colombia Dominican Republic 2.4 Costa Rica Panama Peru Chile Bolivia 2.2 Mexico El Salvador Ecuador Nicaragua Honduras Argentina Guatemala 2.0 0 10 20 30 40 50 60 70 Percent of population living in chronic poverty Source: Chronic poverty data come from SEDLAC (CEDLAS and World Bank). Note: Expected well-being data are based on responses (on a scale of 1–5) to the question, “Do you think that in the next 12 months your economic situation will be better/same/worse?” in 2010 Latin American Public Opinion Project survey. 16 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN the countries with the lowest levels of positive expectations for the future. Similar trends can be observed at the subnational level. Differences in perceptions about the future across socioeconomic groups are also striking. The chronically poor are the most pessimistic about their outlooks, with one out of five expecting his or her economic situation to decline in the next year. This rate is twice as high as the rate among people who exited poverty or who were never poor (figure O.12). The chronically poor are also half as likely as peo- ple who exited poverty to expect their situation to improve. Perhaps unsurpris- ingly, people who exited poverty are the most optimistic, consistent with causal evidence about how positive events can affect people’s outlook in life. Social norms and peer effects also influence state of mind, behaviors, and aspi- rations. Paraphrasing Appadurai (2004) and Ray (2006), aspirations are socially determined, and the capacity to aspire is inherently unequal between rich and poor. The poor often lack the aspirational resources (or capacity) to act and change the conditions of their own poverty. Breaking such a transmission chan- nel is essential. Learning about the positive experiences of others can be beneficial. Increasing the aspirations of poor children is particularly important. Children in Peru living in communities in which the educational attainment aspirations of their peers are high also have higher aspirations (and vice versa), suggest- ing that social interactions can trigger positive changes in the decision-making process (figure O.13). FIGURE O.12 Level of optimism about the next 12 months by the chronically poor, people who escaped poverty, and people who were never poor in Latin America and the Caribbean, 2010 100 80 Percent of respondents 60 40 20 0 Chronically poor Never poor Escaped poverty Better The same Worse Source: Latin American Public Opinion Project 2010. OVERVIEW 17 FIGURE O.13 Correlation between educational aspirations of children ages 12–15 and their peers in Latin America and the Caribbean, 2011 20 Years of education aspired to by peers 15 10 5 0 5 10 15 20 Years of education aspired to by children Source: Young Lives Surveys Peru. The intergenerational transmission of chronic poverty Children born into poverty are less likely to have access to the opportunities avail- able to nonpoor children, creating a channel for low mobility and poverty per- sistence across generations. Children from chronically poor families are severely disadvantaged in schooling outcomes (figure O.14). In Peru, for example, 60 per- cent of children from households that were never poor complete middle school, compared with just 34 percent of children from chronically poor households. State of mind also appears to affect the intergenerational transmission of chronic poverty. Children’s educational aspirations are affected not only by the aspirations of their peers but also by the aspirations of their parents, which tend to be lower among the chronically poor. From Diagnosis to Policies: Design Elements Supporting the Chronically Poor No single approach can eliminate chronic poverty. Policies must be consis- tent with a variety of country-specific factors, including the social contract and political vision in society, budgets, bureaucratic capacity, and the over- all institutional setting (including existing social programs). In recognition of 18 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE O.14 Intergenerational transmission of chronic poverty in selected countries in Latin America and the Caribbean, 2012 a. Finished 6th grade on time 100 School completion rate (percent) 80 60 40 20 0 Ec o a Co ma Ho bia ile Pa ia Ur a Sa s Gu agua la tin Pa ublic ru Ca eric ay y Bo r Ni dor Re il be nd do El ura xic Ric tin ua an raz liv ma Pe Ch the m gu lom na Co an rib a a ua lva en ug Me nd nic B p ate sta r a ca g r Ar A mi La Do Chronically poor Never poor b. Finished 10th grade on time 100 School completion rate (percent) 80 60 40 20 0 Ho bia Ur a ile Ec co Co ama a Gu gua la Pa ia c Sa s ru y Ca eric ay Ni dor Bo r Re il be nd El ura tin Pa ubli n Ric do ua ma an raz liv Pe Ch xi nti the Am ragu lom Co an rib a a lva ua ra ug Me n nd nic B p ate ge sta ca Ar mi La Do Chronically poor Never poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Figures come from 2012 surveys (or nearest year in cases in which 2012 data were not available). the complexity and specificity of policy design, we refrain from presenting a comprehensive review of factors affecting chronic poverty or providing spe- cific recommendations for the design of policies. Instead, we identify elements of policy design that should be taken into consideration in designing social protection programs. OVERVIEW 19 This section begins by synthesizing findings that are relevant for policy design. It then investigates an area of increasing relevance in the policy arena: the coordination of poverty-reduction efforts to build social protection systems that effectively assist the chronically poor. It closes by making the case for bun- dling behavioral and coordination solutions through “social intermediation services.” Refocusing Policy to Reduce Chronic Poverty Five key lessons emerge from this analysis for policy design. Enhance the environment in which poor people live The environment in which people live matters at least as much as their skills and characteristics. A chronically poor family living in a remote district of the Andean sierra or Amazon selva might not be poor if it lived in São Paulo, Bogota, or Lima. Paved roads increase the opportunities to sell one’s prod- ucts. Industrialization brings manufacturing jobs, which generally offer higher wages than agricultural. People covered by universal health coverage are bet- ter able to cope with expenses related to medical emergencies, which are one of the most damaging types of shocks. Better policing reduces crime rates and improves the profitability of local businesses. An optimal social policy should thus balance direct support to the chronically poor with broader investments that improve the environment in which they reside. Improve poor people’s state of mind and raise their aspirations Strategies that help improve poor people’s state of mind and raise their aspi- rations can help prevent them from falling through the cracks of the social safety net system by failing to register for social programs or fulfilling their co-responsibilities. In Peru the national tuberculosis prevention and treatment program nearly doubled uptake by adding a component that addressed depres- sion, which was widespread among potential beneficiaries (Rocha and others 2011). Considering the state of mind can also help the poor deal with stress and engage in long-term planning, as a tweak to a conditional cash transfer program in Colombia did (Barrera-Osorio and others 2011). Promoting interactions with role models has also shown results. In Nicaragua, for example, beneficiaries of a business grant program who interacted more with local leaders invested more in their own children and earned more from nonagricultural activities than did other beneficiaries (Macours and Vakis 2014). It is thus not always necessary to develop new programs: small modifications to existing programs that explore pathways of behavioral change are often a cost-effective way to improve the abil- ity of programs to address chronic poverty. 20 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Address the intergenerational transmission of chronic poverty Malnutrition, inadequate stimulation, fragile health, absent parents, and a risky or violent environment keep people mired in poverty and limit their ability to develop to their full potential. These differences grow over time; by the time social programs finally reach them, many chronically poor adults may lack the skills or mind-set necessary to escape poverty. Policy makers need to break the chronic poverty cycle early in childhood. The emerging policy dialogue and integration of early childhood development into the social development agenda in the region is welcome and should be expanded. Boost labor income Boosting labor income can sustainably lift people out of poverty. Comprehensive poverty-reduction programs should therefore include training and labor-insertion programs. Large inequalities early in life can result in fewer skills and aspirations in adulthood, however, severely weakening the potential impacts of income- generation programs. Income-promotion strategies alone therefore cannot be expected to eliminate chronic poverty. Focus on urban as well as rural areas Many social programs base geographic coverage decisions on the incidence of poverty. As a result, many programs operate mostly in rural areas, where the inci- dence of poverty is highest. Although the incidence of chronic poverty is higher in rural areas, more chronically poor families live in urban areas than in rural ones. Supporting the urban poor can be more challenging than supporting the rural poor, because they are more mobile, making them more difficult to identify and reach. The rural and urban poor also have different sources of income and face different types of shocks and sources of vulnerability. Coordinating Poverty-Reduction Efforts After decades in which the social contract disregarded the needs of the poor, LAC is now creating more inclusive societies. Living conditions of the poor have improved, in part as a result of the surge in social programs. During the past 20 years, Columbia increased the number of new programs by a factor of seven. El Salvador, which had a single program at the beginning of the period, had 30 at the end (figure O.15). The surge in social assistance programs is welcome. Also needed, however, is better alignment of policies and initiatives under a systematic and coordinated approach. Social assistance programs must not only have clear and measurable goals, they also need to be part of a comprehensive poverty-reduction framework in which programs and initiatives communicate and collaborate with one another. OVERVIEW 21 FIGURE O.15 Number of new social assistance programs in selected countries in Latin America and the Caribbean, 1990–99 and 2000–11 40 Number of new programs 30 20 10 0 a il ile bia r or as o ru ay do az tin xic ad Pe Ch gu ur lom Br ua en Me nd lv u Ec Sa Ur g Co Ho Ar El 1990–99 2000–11 Source: World Bank LAC Social Protection database. Coordination is needed at all levels. At the level of the state, there must be a broad agreement that social programs are not populist initiatives but effective social inclusion tools. To achieve such a consensus, the generosity of benefits and type of programs may need to be adjusted to societal perceptions of the scope of mutual help and solidarity. At the level of the executive branch, ministries and pro- grams must be given adequate resources, and technically competent program staff who are shielded from political pressures should be appointed. Lack of coordination of programs, agencies, and ministries substantially lim- its the effectiveness of poverty-reduction efforts. To address the problem, many countries in LAC have established coordinating agencies. These agencies take many forms, including supra-ministries (Ecuador), ministries (Peru), and agencies within a ministry or the presidency (the National Agency for Overcoming Extreme Poverty [ANSPE] in Colombia). The extent to which they are effective in coordi- nating social policies depends greatly on the institutional and financial power they wield, as well as the extent to which they develop clear strategies and institutional arrangements. Recognizing that coordination is costly, countries should attempt to establish the degree of coordination that maximizes benefits while minimizing costs. It is crucial to design policies with clear, specific, and measurable objectives, and to define clear competencies and accountabilities in the implementation of each intervention. It is also important to provide incentives that go beyond good- will, building an accountability system that rewards performance. There is a need to show strong evidence of the benefits of coordination for social policies so that 22 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN social services cease being perceived as inefficient expenses and start being per- ceived as central pillars of countries’ development efforts. Finally, coordination must happen not only at the top but also at the local level. Civil servants, teachers, doctors, and suppliers must be rewarded and benchmarked against measurable achievements (as opposed to inputs). Bundling Behavioral and Coordination Solutions: The Emergence of Social Intermediation Services In the classical “passive” social assistance approach, budgetary considerations and poverty status determine eligibility, and the poor are assumed to seek out and enroll in social programs intended to benefit them. In an “active” approach, pro- grams seek out and enroll beneficiaries and help them identify their own devel- opment goals. Because behavioral barriers can induce the chronically poor to exclude themselves from social programs designed to assist them (Galasso 2011), the active approach is likely to be more effective. Efforts to offer a more systematic response to both behavioral and coordination constraints have been emerging in various LAC countries in the form of “social intermediation services,” intended to help the poor overcome access, information, and other barriers through a holistic, systemic, and household-based approach. This approach, which aims to provide people with the building blocks they need to overcome their specific challenges, deserves attention. Camacho and others (2014) review two of these programs (Chile Solidario, the first such program in the region, and Red Unidos in Colombia). They con- clude that social intermediation services can facilitate the poor’s access to social programs; improve beneficiaries’ social and emotional well-being; and, if the right conditions are created, increase their employment perspectives. The programs’ success shows that psychosocial constraints are not insurmountable barriers and that it is possible to ensure that the chronically poor benefit from state assistance. Social intermediation services do not directly bring material benefits to fami- lies; they facilitate access to existing social programs. They must therefore be well integrated within the social assistance system, have interoperable informa- tion systems to track the supply of and demand for social services, and employ a staff of well-trained social workers to actively work with the target population to match them to social programs that address family-specific needs. Strong social programs must already be in place; granting access to services of poor quality or services that are poorly tailored to the needs of the extreme and chronically poor may lead to little or no impact. Social workers are the backbone of family support. They need to be well- trained and must meet minimum qualifications. They need to be knowledgeable of the eligibility rules and procedures of all social assistance programs and famil- iar with the informational and psychosocial barriers facing the chronically poor, OVERVIEW 23 so that they can help families overcome them. Visits to families must be conducted on a regular basis and tailored to individual needs, as behavioral barriers can be overcome only if families feel that their constraints are understood and that social workers are willing and have the means to help them overcome them. Social intermediation services vary in sophistication. The ambition of an inter- vention such as Chile Solidario goes beyond mere access to social programs to include psychosocial support and programs to fill supply gaps. Where resources and capacity are more limited, simpler programs that focus only on increasing access can also lead to positive impacts. Notes 1. Given LAC’s level of economic development, a $4 a day poverty line was used to measure total poverty in the region; it is about equal to the average of moderate poverty lines in the region. A $2.50 a day extreme poverty line (about equal to the average of national extreme poverty lines) is considered more appropriate for LAC than the World Bank’s $1.25 a day global extreme poverty line. 2. We follow Ferreira and others (2012), who define four economic classes based on the concept of economic security: (a) the poor (per capita income of less than $4 a day); (b) the vulnerable (per capita income of $4–$10 a day); (c) the middle class (per capita income of $10–$50 a day); and (d) the rich (per capita income of more than $50 a day), all in 2005 purchasing power parity. 3. Households are considered to have a minimum level of assets if they have at least three of the following four assets: a landline or mobile phone; a bicycle, car, or motorcycle; a refrigerator; and a television. References Alvaredo, F. and L. Gasparini. 2015. “Recent Trends in Inequality and Poverty in Developing Countries.” In Handbook of Income Distribution, vol. 2, edited by Anthony B. Atkinson and Francois Bourguignon. Science Direct. Appadurai, Arjun. 2004. “The Capacity to Aspire: Culture and the Terms of Recognition.” In Culture and Public Action, ed. Vijayendra Rao and Michael Walton, 59–84. Stanford, CA: Stanford University Press. Arim, R., M. Brum, A. Dean, M. Leites, and G. Salas. 2013. “Movilidad de ingreso y trampas de pobreza: nueva evidencia para los países del Cono sur.” Estudios Económicos 28 (1): 3–38. Banerjee, Abhijit, and Sendhil Mullainathan. 2010. “The Shape of Temptation: Implications for the Economic Lives of the Poor.” NBER Working Paper 15973, National Bureau of Economic Research, Cambridge, MA. Barrera-Osorio, Felipe, Marianne Bertrand, Leigh L. Linden, and Francisco Perez-Calle. 2011. “Improving the Design of Conditional Transfer Programs: Evidence from a Randomized Education Experiment in Colombia.” American Economic Journal: Applied Economics 3 (2): 167–95. 24 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Camacho, Adriana, Wendy Cunningham, Jamele Rigolini, and Veronica Silva. 2014. “Addressing Access and Behavioral Constraints through Social Intermediation Services: A Review of Chile Solidario and Red Unidos.” Policy Research Working Paper 7136, World Bank, Washington, DC. Carter, Michael R., and Christopher B. Barrett. 2006. “The Economics of Poverty Traps and Persistent Poverty: An Asset-Based Approach.” Journal of Development Studies 42 (2): 178–99. Cord, Louise, Oscar Barriga Cabanillas, Leonardo Lucchetti, Carlos Rodríguez-Castelán, Liliana D. Sousa, and Daniel Valderrama. 2015. “Inequality Stagnation in Latin America in the Aftermath of the Global Financial Crisis.” Policy Research Working Paper 7146, World Bank, Washington, DC. Cord, Louise, Maria Eugenia Genoni, and Carlos Rodríguez-Castelán, eds. 2015. Prosperidad compartida y fin de la pobreza en América Latina y el Caribe. Washington, DC: World Bank. Cord, L., and L. F. López-Calva. 2012. “The Equity Agenda in Latin America and the Caribbean: Opportunities.” World Bank, Poverty and Equity Unit, Washington, DC. Dang, Hai-Anh, and Peter Lanjouw. 2014. “Measuring Poverty Dynamics with Synthetic Panels Based on Cross-Sections.” World Bank, Washington, DC. Dang, Hai-Anh, Peter Lanjouw, Jill Luoto, and David McKenzie. 2014. “Using Repeated Cross-Sections to Explore Movements into and out of Poverty.” Journal of Development Economics 107: 112–28. Duflo, Esther. 2012. “Human Values and the Design of the Fight against Poverty.” Tanner Lecture, Massachusetts Institute of Technology, Cambridge, MA. Ferreira, Francisco H. G., Julian Messina, Jamele Rigolini, Luis-Felipe López-Calva, Maria Ana Lugo, and Renos Vakis. 2012. Economic Mobility and the Rise of the Latin American Middle Class. Washington, DC: World Bank. Fukuyama, Francis. 2011. The Origins of Political Order: From Prehuman Times to the French Revolution. London: Profile Books. Galasso, Emanuela. 2011. “Alleviating Extreme Poverty in Chile: The Short Term Effects of Chile Solidario.” Estudios de Economía 38 (1): 101–27. Latin American Public Opinion Project. 2010. Vanderbilt University, Nashville, TN. Loayza, Norman V., Jamele Rigolini, and Oscar Calvo-González. 2014. “More Than You Can Handle: Decentralization and Spending Ability of Peruvian Municipalities.” Economics & Politics 26 (1): 56–78. Locke, Edwin A., and Gary P. Latham. 2002. “Building a Practically Useful Theory of Goal Setting and Task Motivation: A 35-Year Odyssey.” American Psychologist 57 (9): 705–17. Macours, Karen, and Renos Vakis. 2014. “Changing Households’ Investments and Aspirations through Social Interactions: Evidence from a Randomized Transfer Program.” Economic Journal 124: 607–33. Ray, Debraj. 2006. “Aspirations, Poverty, and Economic Change.” In Understanding Poverty. Edited by Abhijit Vinayak Banerjee, Roland Bénabou, and Dilip Mookherjee, 409–21. Oxford, U.K.: Oxford University Press. Rocha, C., R. Montoya, K. Zevallos, A. Curatola, W. Ynga, J. Franco, F. Fernandez, and others. 2011. “The Innovative Socio-economic Interventions against Tuberculosis (ISIAT) Project: An operational assessment.” International Journal of Tuberculosis and Lung Disease: The Official Journal of the International Union against Tuberculosis and Lung Disease 15 (Suppl. 2): S50. OVERVIEW 25 Sen, Amartya. 1999. Development as Freedom. Oxford, U.K.: Oxford University Press. World Bank. 2012. World Development Report 2013: Jobs. Washington, DC: World Bank. ———. 2013. Inclusion Matters: The Foundation for Shared Prosperity. Washington, DC: World Bank. ———. 2014a. “Social Gains in the Balance: A Fiscal Challenge for Latin America and the Caribbean.” LCSPP Poverty and Labor Brief No. 6, Latin America and the Caribbean Poverty and Gender Unit, Washington, DC. ———. 2014b. World Development Report 2015: Mind, Society, and Behavior. Washington, DC: World Bank. 26 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN CHAPTER 1 The Roaring 2000s and the Left Behind B y most measures, the 2000s were one of the most impressive decades of economic development in Latin America and the Caribbean (LAC) in recent history. With the exception of 2009, LAC’s gross domestic product (GDP) per capita grew consistently at an average rate of 2.5 percent between 2000 and 2012 (figure 1.1). Inequality narrowed substantially, with the regional Gini coefficient for per capita income falling by an unprecedented 5 percentage points, from 0.57 in 2000 to 0.52 in 2012 (Alvaredo and Gasparini 2015). Sustained economic growth and substantial reductions in income inequality led to remarkable increases in the incomes of people at the bottom of the income distribution (Cord and others 2015; Cord, Genoni, and Rodríguez-Castelán 2015). As a result, poverty decreased by more than 16 percentage points within a single decade, from 41.6 percent in 2003 to 25.3 percent in 2012.1 Extreme poverty was cut in half, from 24.5 percent to 12.3 percent (figure 1.2). Overall, some 70 million people moved out of poverty, the largest poverty reduction in the region in decades. Another facet of the dramatic reduction in poverty is the emergence of a large middle class, which increased from about 23 percent of the population in 2003 to 34 percent in 2012; in 2010, for the first time, the number of middle class people exceeded the number of poor people.2 Nevertheless, LAC is not yet a middle class society: Most people exiting poverty did not join the ranks of the middle class but the ranks of the vulnerable class, which remains at risk of falling back into poverty. The size of the vulnerable class in LAC is changing the social policy dialogue in the region, where much recent concern has focused on how to establish more integrated safety nets for people who managed to exit poverty but still remain vul- nerable (Ferreira and others 2012). 27 FIGURE 1.1 GDP per capita and income inequality in Latin America and the Caribbean, 2000–12 11,000 0.58 0.57 10,000 0.56 GDP per capita (dollars) 0.55 Gini coefficient 9,000 0.54 8,000 0.53 0.52 7,000 0.51 6,000 0.50 00 01 02 03 11 08 09 10 12 04 06 05 07 20 20 20 20 20 20 20 20 20 20 20 20 20 Gini coefficient GDP per capita Source: Gini coefficient: World Bank 2014, based on SEDLAC data (CEDLAS and World Bank). GDP per capita: World Development Indicators. Note: GDP is in 2005 purchasing power parity dollars. The income-generating capacity of households remains by far the most impor- tant driver of poverty reduction: The increase in labor market earnings accounted for 68 percent of the reduction in poverty in the region between 2003 and 2012 (World Bank 2014). Increases in female labor market participation and earnings made an important contribution, with almost 40 percent of the contribution of labor incomes to poverty reduction resulting from increases in earnings by women (figure 1.3). More generous social transfers and pensions also played a role. Transfers con- tributed roughly 9 percent of the reduction in poverty between 2003 and 2012, and pensions (both contributory and noncontributory) contributed 14 percent (World Bank 2014). Despite these dramatic advances, one in four people in the region remains poor. Some of them have experienced shocks that pushed them back into poverty (the “transitory” poor), while others never escaped (the “chronically” poor). These people did not benefit much from the impressive growth rates of the 2000s and may have fallen through the cracks of the social assistance system. The prospects for their escaping poverty in the near future are weak (Arim and others 2013). GDP growth has slowed significantly, from about 6 percent 28 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 1.2 Poverty, vulnerability, and the middle class in Latin America and the Caribbean, 2000–12 50 40 Percent of population 30 20 10 0 01 00 02 03 04 06 08 09 10 11 12 05 07 20 20 20 20 20 20 20 20 20 20 20 20 20 Moderately poor (per capita income of less than $4 a day) Vulnerable (per capita income of $4–$10 a day) Extremely poor (per capita income of less than $2.50 a day) Middle class (per capita income of $10–$50 a day) Source: World Bank 2014, using SEDLAC data (CEDLAS and World Bank). Note: Estimates are population-weighted averages of country estimates. Poverty lines and incomes are expressed in 2005 purchasing power parity dollars. in 2010 to an estimated 0.8 percent in 2014 (World Bank 2014). Improved labor market prospects may therefore not be sufficient to pull the chronically poor out of poverty. And despite significant investment in social assistance, levels remain modest compared with other parts of the world. The mechanisms by which poverty reduction occurs and the ways in which social spending and economic growth can affect it are complex. Economic growth may not necessarily trickle down to the poorest, and heavy social spending does not always translate into significant declines in poverty. In order to understand why some people cannot escape poverty—and to design social assistance programs that will help them leave poverty for good—it is necessary first to identify the chronically poor and the nature of the challenges THE ROARING 2000S AND THE LEFT BEHIND 29 FIGURE 1.3 Decomposition of changes in poverty in Latin America and the Caribbean by sources of income, 2003–12 0 –5 –10 Share of change (percent) –15 –20 –25 –30 –35 –40 –45 –50 Share of Labor Share of Labor Pensions Transfers Other occupied income occupied income (public and nonlabor private) income Male labor Female labor Other sources of income Source: World Bank 2014, using SEDLAC data (CEDLAS and World Bank). Note: Estimates of poverty at the regional level are population-weighted averages of country-specific poverty rates. The figure shows the Shapley decomposition of poverty changes. See Azevedo, Cong Nguyen, and Sanfelice (2012) for details about this technique. Totals for female labor and other income differ from sum of components because of rounding. they face. This book does so, analyzing as well the extent to which social assis- tance systems are designed to respond to their needs. Studies of chronic poverty in the region have long suffered from the absence of longitudinal data, which prevented them from identifying and pro- filing the chronically poor. This book benefits from recent methodological advances in statistical methods that can partly substitute for the absence of longitudinal data. Its institutional analysis benefits from the establishment of a new wave of social programs and institutions over the last decade designed to tackle chronic poverty, which are reviewed in order to assess factors that hinder or enhance their effectiveness. The book examines three questions: 1. How many people in LAC live in chronic poverty? 2. Who are they, and what precludes them from escaping poverty? 3. How effective has the policy toolkit in the region been in addressing chronic poverty? 30 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN The first question is central to the book, as it establishes the magnitude of the issue. Understanding chronic poverty is a dynamic concept, which requires data that are not typically available. We solve this problem by using new empirical techniques that allow us to create synthetic panels, which make dynamic analysis of individuals feasible. Chapter 2 explains these conceptual and technical issues. The second question, addressed in chapters 3 and 4, examines what it means to live in chronic poverty in LAC and identifies both the microlevel aspects (gen- der, ethnicity, voice, aspirations) and macrolevel issues (geography, risks, connec- tivity to opportunities, markets, institutions, and policies) associated with it. In addition to the chronic poverty estimates using synthetic panel techniques, we also use spatial chronic poverty panels, when available, to explore how geography may play a key role in persistence. We also explore multidimensional measures of chronic poverty to understand the linkages among the various dimensions of chronic poverty. Chapter 5 reviews how recent policies in the region have evolved to deal with chronic poverty. Its goal is to connect the findings and trends identified in the first four chapters with the policies underway in order to assess the extent to which the existing models in the region can address chronic poverty. The aim is not to provide policy prescriptions but to highlight efforts that have worked and could be considered as part of the redesign of social policy that is underway in many countries in the region. Notes 1. Given LAC’s level of economic development, a $4 a day poverty line was used to measure total poverty in the region; it is about equal to the average of moderate poverty lines in the region. A $2.50 a day extreme poverty line (about equal to the average of national extreme poverty lines) is considered more appropriate for LAC than the World Bank’s $1.25 a day global extreme poverty line. 2. We follow Ferreira and others (2012), who define four economic classes based on the concept of economic security: (a) the poor (per capita income of less than $4 a day); (b) the vulnerable (per capita income of $4–$10 a day); (c) the middle class (per capita income of $10–$50 a day); and (d) the rich (per capita income of more than $50 a day), all in 2005 purchasing power parity. References Alvaredo, F. and L. Gasparini. 2015. “Recent Trends in Inequality and Poverty in Developing Countries.” In Handbook of Income Distribution, vol. 2, edited by Anthony B. Atkinson and Francois Bourguignon. Science Direct. Arim, R., M. Brum, A. Dean, M. Leites, and G. Salas. 2013. “Movilidad de ingreso y trampas de pobreza: nueva evidencia para los países del Cono sur.” Estudios Económicos 28 (1): 3–38. THE ROARING 2000S AND THE LEFT BEHIND 31 Azevedo, Joao Pedro, Minh Cong Nguyen, and Viviane Sanfelice. 2012. “ADECOMP: Stata Module to Estimate Shapley Decomposition by Components of a Welfare Measure.” Boston College, Department of Economics, Chestnut Hill, MA. Cord, Louise, Oscar Barriga Cabanillas, Leonardo Lucchetti, Carlos Rodríguez-Castelán, Liliana D. Sousa, and Daniel Valderrama. 2015. “Inequality Stagnation in Latin America in the Aftermath of the Global Financial Crisis.” Policy Research Working Paper 7146, World Bank, Washington, DC. Cord, Louise, Maria Eugenia Genoni, and Carlos Rodríguez-Castelán, eds. 2015. Prosperidad compartida y fin de la pobreza en América Latina y el Caribe. Washington, DC: World Bank. Ferreira, Francisco H. G., Julian Messina, Jamele Rigolini, Luis-Felipe López-Calva, Maria Ana Lugo, and Renos Vakis. 2012. Economic Mobility and the Rise of the Latin American Middle Class. Washington, DC: World Bank. World Bank. 2014. “Social Gains in the Balance: A Fiscal Challenge for Latin America and the Caribbean.” LCSPP Poverty and Labor Brief 6, World Bank, Latin America and the Caribbean Poverty Gender and Equity Group, Washington, DC. 32 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN CHAPTER 2 What Is Chronic Poverty and How Is It Measured? Conceptual Underpinnings The notion of chronic poverty is intrinsically linked to the dimension of time and the persistence of low welfare equilibria. Within generations, chronic poverty is related to the weak links between economic growth and the income-generation capabilities that hinder individuals’ ability to integrate productively into society. The situation of an individual who is vulnerable to falling into poverty as a con- sequence of an adverse shock is very different from the situation of an individual who is poor for long spells of time. Distinguishing transitional from structural or chronic poverty and identifying what drives each can therefore shed light on the best policies to address them. Jalan and Ravallion (1998, 2000) define the chronically poor as people whose welfare is at a low level over long periods. Clark and Hulme (2005) identify the time and duration of poverty as one of the three meta-dimensions of poverty that are key to understanding the phenomenon and the processes that generate and reduce it. Chronic poverty adds an important dimension to the measurement of poverty beyond depth, severity, breadth, and multidimensionality: the time dimension. Chronic poverty also captures dimensions of social progress that transcend its monetary manifestation. The chronicity or persistence of poverty is linked to the fairness of the social system, the lack of voice of some population segments, and the lack of responsiveness of the political system to their needs (Hickey and Bracking 2005). 33 The idea that chronic poverty goes beyond income is reflected in the work of Sen (1983, 1985, 1999), which builds on Rawlsian concepts to argue that access to or ownership of material goods should not be the goal of develop- ment. Rather, development and progress should be seen in terms of capabili- ties (what a person is able to do) and functioning (what a person manages to do). The direct implication of this notion is that to understand and deal with chronic poverty, simply understanding the external constraints and factors that prohibit a person from leading a productive life is not sufficient. It is also crucial to understand the internal process by which individuals’ capabilities can shape aspirations and thus directly influence their ability to define and actively seek to achieve goals. From an intergenerational perspective, chronic poverty can be thought of in terms of the long-term structural constraints that persist from one genera- tion to another, which can be framed as a problem of fairness or “equality of opportunity.” In an equal and fair society, the advantages and achieve- ments of individuals should be independent of their initial circumstances and social background (Roemer 1998; Barros and others 2009). Hulme and Shepherd (2003) define chronic poverty as “the poverty that persists for many years or a life course and that may be transmitted across generations.” One of the premises underlying the analysis of chronic poverty is the fact that “poverty repeated over time has a greater impact than poverty that does not recur” on a number of different negative outcomes, particularly for chil- dren (Foster 2009). Looking at the interaction between the various facets of chronic poverty is also essential for understanding its nature and figuring out how to address it. For instance, the two-way link between chronic poverty and voice constitutes a cru- cial aspect that has often been overlooked, in part because of the lack of data with which to carry out rigorous empirical work. “Faustian bargains,” in which the poor give up their agency and remain loyal to clientelistic institutions in exchange for economic security, may arise (Wood 2003). Lipton (1983, 1986) coined the term ultra-poverty, which he defines to comprise very poor people whose poverty persists over long periods of time and includes more than one dimension of depri- vation. Poverty and decisions about investments in human capital accumulation are linked by the constraints parents face today, which directly affect their chil- dren’s opportunities tomorrow. An extensive body of literature on chronic poverty explores the concept of poverty traps—indefinite persistent poverty, not just poverty in a few time periods. It focuses on “identifying and explaining the existence of low well-being ‘basins of attraction’ within an economy and naturally emphasizes how stocks of assets (and resulting flows of income) evolve—or fail to evolve—over time” (Barrett and Carter 2013).1 Poverty creates poverty. This work explains how deep pockets of poverty can be created by multiple channels and mechanisms (assets, human capital, and 34 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN behavior, to name a few) and can operate at different levels (individual, household, community, region, country). The empirical definition of chronic poverty used in this book differs from the definition above because of data constraints. Nevertheless, this literature provides a natural entry point for discussing and studying poverty persistence in Latin America and the Caribbean (LAC) over the last decade. We use a simple conceptual framework that attempts to connect many of the aspects discussed above. The framework is by no means new; we borrow heav- ily and adapt freely from the theoretical and empirical literature, particularly from Sen’s work on capabilities but also from work on poverty traps by Carter and Barrett (2006), Cord and López-Calva (2012), and the World Bank (2013), among others. The framework involves the close interaction between two fea- tures: inputs (building on the poverty trap literature) and the process of upward mobility (adapting Sen’s capabilities approach). We organize inputs into three broad categories: endowments, the enabling context, and the state of mind. The process of upward mobility involves three critical steps that can lead to a pathway out of poverty: engaging in decision making (for instance, considering investing in a nonfarm business); transforming choices into action (that is, “acting” and deciding to invest in a nonfarm business); and converting action into an out- come (that is, realizing the investment in the nonfarm business). The interaction of these elements gives rise to various scenarios in which poverty can persist, as discussed below. Inputs We start from the asset-based approach proposed by Carter and Barret (2006). The advantage of this approach is that it allows chronic poverty to be examined from a forward-looking perspective. In this framework, poverty exists and persists because of constraints that prohibit the optimal accumulation and use of existing asset endowments (skills, physical assets). For example, lack of access to credit or insurance markets may induce people to make suboptimal investment decisions (think of being stuck in subsistence crop production instead of investing in cash crops because of the need to ensure a minimum level of consumption). These con- straints can then be used to predict future well-being for a given household by looking at its set of assets and how these constraints affect the income-generation process. An enabling context is also a necessary input. Chronic poverty may persist not because endowments are low but because of the uneven distribution of geographic and local factors (for example, climate, infrastructure, supply of public services) that differentially affect the returns to those endowments. A great product cannot be sold in a market if there is no road to transport it there. The absence of financial services makes it impossible for people to pur- sue productive opportunities if they are liquidity constrained. The context WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 35 therefore influences the probability of escaping poverty. Individuals who live in an enabling context have better chances of escaping poverty than people who do not. Inclusive and efficient institutions are one aspect of the enabling context. They are potentially of great importance for the chronically poor, who often lack the voice to push for reforms and defend their rights. Institutions there- fore play a role in facilitating (or hindering) upward mobility. Social pro- grams will be of little effect if they are not complemented by high-quality services and infrastructure. For example, income-generation opportunities will be limited if a woman lives in a setting in which social norms (an infor- mal institution) or lack of support from the justice system (a formal institu- tion) results in a situation in which she can work only in subsistence farming. The same woman living in a setting in which women’s rights and institutions actively facilitate women’s involvement in the labor market faces very differ- ent possibilities. Individuals endowed with the same assets may not be equally successful in converting them into a livelihood even within the same context, however, because of differences in innate abilities or behavioral constraints. Carter and Ikegami (2009) suggest that the critical minimum stock of assets needed to get ahead in the future is also a function of intrinsic abilities. The crucial question is which abilities are needed (given existing assets) to get ahead (Appadurai 2004; Ray 2006). We refer to these behavioral constraints as the “state of mind” and focus on three interconnected dimensions. First, poverty adversely affects individu- als’ state of mind and decision-making process. For example, “present biases” may focus poor people’s attention on today’s problems rather than tomorrow’s, even if doing so may be detrimental in the long run, thereby perpetuating pov- erty (Banerjee and Mullainathan 2010). Reduced mental resources are also rel- evant for the persistence of intergenerational poverty. Because poor people must devote much of their mental resources to tackling survival problems, they have less time available for other tasks, such as raising their children and investing in their future. Second, poverty affects aspirations—the presence of forward-looking goals and a willingness to attain them (Locke and Latham 2002). Sen’s capability approach makes it clear that a weak capacity to aspire can weaken the capa- bility to act in order to achieve desired outcomes. Low aspirations can affect perceived returns and influence whether or not an individual takes advan- tage of an opportunity. People with low aspirations may also be more likely to engage in behaviors that are detrimental to their own socioeconomic mobility and that of their children (for example, substance abuse, unprotected sex by teenagers). As Sen notes, behavioral inputs and capabilities shape aspirations 36 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN and thus directly influence individuals’ ability to define and actively seek to achieve goals. A weak capacity to aspire is associated with a limited temporal horizon that could translate, for instance, into low savings, failure to take advan- tage of good investments, and unproductive behaviors. “Hopelessness destroys both the will and the ability to invest in one’s future and oneself,” notes Duflo (2012). Poverty generates poverty. Third, the brain is malleable throughout life; changes in attitudes are there- fore feasible. As early as the 1930s (Conel 1939), researchers understood that brain development continues throughout the life cycle. Cognitive, motor, lan- guage, social, and emotional skills start developing in the first years of life, but they continue to evolve through adolescence and adulthood. Cognitive, emotional, and social capacities are especially relevant for poverty. Emotional well-being and social competence provide a strong foundation for cognitive abilities and positive aspirations and form the foundations of brain architec- ture. The emotional and physical health, social skills, and cognitive-linguistic capacities that emerge in the early years are important for success in school, the workplace, and the larger community (Radner and Shonkoff 2012). For this reason, early childhood development programs are critical to break the intergenerational transmission of chronic poverty. It is never too late to improve, reshape, or even develop noncognitive skills and aspirations, however. Dweck (2006) explores how one’s mind-set influ- ences one’s actions. She describes two groups of people: people who believe or behave under the assumption that their success is based on innate abil- ity (fixed mind-set) and people who believe that success is based on hard work, learning, training, and persistence (growth mind-set). She shows how people’s reaction to failure depends critically on their mind-set: People with a fixed mind-set see failure as a negative affirmation of their basic (in)abili- ties, whereas people with a growth mind-set flourish in failure, seizing it as an opportunity for improvement. Dweck argues that the growth mind-set allows a person to live a less stressful and more successful life—and that mind-sets are malleable. Duckworth and others (2007) find that believing that change through self- mastery is possible can lead to sustained efforts to achieve one’s goals. Helping individuals shift beliefs toward the growth mind-set is therefore critical, especially in the context of chronic poverty. The Process of Upward Mobility The first element in the process of upward mobility is “engaging in decision making,” which is where the state of mind may matter the most. Think about the following example: In considering whether to participate in an income-generation opportunity (notice the explicit emphasis and differentiation on “considering” as WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 37 opposed to “participating,” which is the next part of the process), an individual may decide to pass it up because she (a) does not know the option exists (that is, she faces an informational constraint [Jensen 2010; Baird, McIntosh, and Özler 2011]); (b) believes that, given her skills, she would not be able to advan- tage of it; or (c) lacks aspirations. Someone with low aspirations may perceive that the returns to an investment are lower than they actually are or that she will fail despite whatever efforts she makes. Context could amplify both effects. For example, local peer effects may accentuate low informational flows and the forma- tion of aspirations (Wilson 1987). To summarize: As a result of endowments, the context, and the state of mind, some people may never consider opportunities, impeding investment and increasing the risk of poverty persistence. After considering an opportunity, there is a need to transform choices into actions. The differences between the two stages are subtle, but impor- tant: Considering participation kicks off a lengthy process in which a person evaluates all the pros and cons and ultimately decides whether to take action. Endowments, the context, and the state of mind also affect the decision to take action. If credit markets function poorly, for instance, poor individuals may consider participating in an income-generation opportunity but ultimately realize that they would not be able to act on it because of credit constraints. The context can also affect the decision to engage: Poor local infrastructure (such as bad roads) may make the costs of marketing a product prohibitive. The state of mind matters as well: If an individual lacks aspirations, he may (incorrectly) perceive that an opportunity will generate only low returns or is too risky. Finally, an individual who must focus on urgent daily tasks (like fetching water) may make faster, simpler calculations about an investment decision, potentially coming to the wrong conclusion about its profitability (Mullainathan and Shafir 2013). The last part of the process is conversion, in which all three inputs matter. In general, the productivity of an investment depends on individual ability and effort, hence endowments like skills and assets are central. The enabling context also affects returns. An uninsured weather shock, for example, has a different effect on income than an insured shock. The state of mind may also affect outcomes at this stage of the process. Positive peer effects and social norms (themselves partly driven by context) may improve motivation and effort, which in turn enhance the likelihood of success (Macours and Vakis 2014). Figure 2.1 presents all the elements in this framework. It does not show how they interact with one another, however. There are many ways in which low levels of endowments, the lack of an enabling context, or a poor state of mind can affect the stages of the process, giving rise to chronic poverty induced through different channels (box 2.1), all of which highlight the policy challenge of identifying and designing policy solutions for the chronically poor. 38 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 2.1 From chronic poverty to upward mobility: Inputs and the process of emerging from poverty Endowments Enabling context State of mind • Skills • Markets • Aspirations • Physical assets • Services • Psychological • Risk welfare • Institutions Converting action into welfare Transforming outcomes Upward choice into action Mobility Considering opportunities Action Choice Poverty Inputs Process Outcomes BOX 2.1 The framework in practice Heidi is a poor indigenous woman in her early 20s living in Pomacanchi, a village about an hour south of Cusco, Peru. She was hoping to improve life for her family by starting her own business—a corner grocery store. In the hope that this dream would someday be realized, she poured her meager savings into raising cuys, a local delicacy. Rural families in the region often raise these small guinea pigs as a source of supplementary income. Investing in cuys was a safe investment, but it was unlikely to generate sufficient resources to change Heidi’s living conditions. In 2008 the Cusco branch of Credinka Bank launched a product aimed at poor rural women: a savings account that did not require a large initial deposit. Opening a savings account could grant Heidi access to the formal financial system, increasing the likelihood of receiving credit in the future. Opening an account seemed like an opportunity that could finally allow her to realize her dream. Nevertheless, and despite her high aspirations, she did not pursue the opportunity to join the formal banking system. Like many residents of Pomacanchi, Heidi found the very idea of entering a bank surreal, something that “people like her” simply did not do. Social norms prevented her from considering formal banking as a real option—they created a mental constraint. Everything changed when a financial literacy program, Corredor Puno Cusco, came to Pomacanchi, hosting a series of self-esteem talks and informational events at which local women (continued on next page) WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 39 BOX 2.1 The framework in practice (continued) could learn more about their rights, including the right to access the financial system. The highlight of these events for Heidi was seeing rural women from neighboring communities talk about their experience using savings accounts. These women were very similar to Heidi—they were indigenous and rural, some were not literate—but they had been regularly making deposits and receiving interest on their savings. They were not intimidated by the formal banks. Their saving strategies had yielded a multitude of positive results, allowing them to pay for their children’s education and invest in productive activities. These women had pushed beyond the mental frame that Heidi was using in her decision-making process, which assumed that opening a savings account was not a possibility for her. Heidi chose to open a savings account. Three years after doing so, she owned three corner groceries. She saved enough to buy a motorcycle, which she rented out as a moto-taxi. And she had new aspirations: expanding her three grocery stores into a local chain. Heidi’s story illustrates the linkages among the various aspects of the framework. Heidi had skills, abilities, and high aspirations. She was financially constrained, but a bigger constraint was her state of mind, which was influenced by the context and social norms that convinced her that opening a savings account—her entry into accessing credit—was not a possibility. Even when Credinka introduced the savings account scheme, she did not engage. The catalyst for change was the talks organized by Corredor, which triggered a change in Heidi’s perception of her compatibility with the banking system and shifted her view of her rights as a Peruvian citizen. Seeing women like herself achieving dreams like hers changed her mind-set, motivating her to take action. The lifting of external constraints alone did not have much effect until her mental state changed. Once it did, Heidi was ready to act and work toward achieving her dreams. The “conversion” process and eventual outcome were the easy parts of the process. Source: Perova and Vakis 2013. The Empirical Challenges of Measuring Chronic Poverty Measuring chronic poverty faces two challenges. The first is reducing a complex notion into one or a few dimensions that are simple to understand and can be cap- tured by household surveys. The second is measuring chronic poverty, which has an inherent time dimension, in the absence of longitudinal surveys. Review of the Literature One way to study chronic poverty is to adopt the approach taken by Yaqub (2000), who splits the methodology for measuring chronic poverty into two broad classes, the “component approach” and the “spells approach.” Jalan and Ravallion (2000) propose what has become the most common form of the components approach. It is grounded in the theoretical notion that consumption and utility are based on long-term expected earnings. The poor can at least partly insure against temporary 40 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN income shocks. A measure of chronic poverty should therefore be based on the long-term expected component of income. Income is separated into two compo- nents. The chronic component is defined as expected income (or consumption) over time, represented by the arithmetic mean of income over time. The transient component consists of the difference between total poverty and the chronic com- ponent of poverty. An individual is identified as chronically poor if the average income over time lies below the poverty line. Chronicity in this case refers to a component of an individual’s income, not the state of poverty of an individual. Therefore, when using Jalan and Ravallion’s approach strictly, one cannot identify people as chronically poor, one can identify only components of income that con- tribute to chronic poverty. The implicit assumption of this model is that income is perfectly transferrable across periods. Under this assumption, identification of the chronically poor is not very sensitive to the amount of time an individual actually spends in poverty. In an extreme case, individuals can be nonpoor in all but one period and still be consid- ered chronically poor if their mean income is below the poverty line. Additionally, if one assumes that income is perfectly transferable across periods, it is safe to pre- sume that an individual who is identified as poor in a given period is chronically poor. As a result, the concepts of chronic poverty and poverty become conflated. Foster and Santos (2013) introduce a measure based on the components approach that relaxes the assumption of perfectly substitutable incomes across time. They calculate poverty in each period and aggregate poverty over time by taking the mean of poverty from each period, allowing the researcher to choose the level of substitutability of incomes across periods. The second-most common way to empirically measure chronic poverty is to use a spells approach. In this approach, individuals are typically identified as chronically poor if they are poor in a certain number of periods. The spells approach is particu- larly useful for identifying transitions in and out of poverty (Hulme and Shepherd 2003). In their classic paper, Bane and Ellwood (1986) define a spell as a set of contiguous periods in which an individual’s income falls below the poverty line. A basic flaw of using the spells approach with the typical data collection methodology is that it is impossible to determine if an individual is poor before, between, or after waves of a panel. In order to calculate chronic poverty, it is essential to make assumptions or imputations about poverty for individuals dur- ing the unobserved periods. In order to mitigate bias, researchers have used techniques such as exit probabilities, hazard models (Bane and Ellwood 1986), and survival analysis (Ruggles and Williams 1989) to estimate the duration of poverty for truncated datasets. An additional drawback of the spells approach is that time is incorporated only into the identification, not the aggregation, step. Therefore, the length someone spends in poverty does not affect the magnitude of the poverty measurement. This approach also makes the implicit assumption that income cannot be smoothed over periods. WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 41 Foster (2009) introduces a class of chronic poverty measures based on the Foster-Greer-Thorbecke (FGT) class that improves the properties of the aggregation step in the spells approach. This measure establishes a dual cutoff for chronic poverty. The first, the traditional poverty line, identifies whether individuals are poor in a given period. The second, the duration cutoff, estab- lishes the proportion of periods in which an individual must be identified as poor to be considered chronically poor. The extent to which the consecutive- ness of time spells should be taken into consideration in a measure of chronic poverty has been subject to debate (Foster 2007; Gradín, del Río and Cantó 2012; Porter and Quinn 2008; Calvo and Dercon 2009; Mendola, Busetta, and Milito 2011). Hoy and Zheng (2011) unite the spells approach and the components approach. To calculate lifetime poverty, the authors aggregate two components. First, spells of poverty in each period of an individual’s life are identified and aggregated. Second, a “lifetime” poverty line is identified that represents permanent consumption over time in a way that is similar to the components approach proposed by Jalan and Ravallion (2000). The most obvious drawback of this approach is that there is no dataset that tracks a representative sample over its lifetime. We are interested in long-term chronicity, defined as an individual who has been poor for a long period of time. Even such a “simple” definition presents an empirical challenge. Poverty estimates using cross-sectional data like the ones in figure 1.2 do not distinguish between people who have been persistently poor and people who became poor and may escape poverty. They represent the net flow of movements across the income distribution. Studying the gross movement within generations by focusing on people who remain poor requires panel datasets that follow individuals or households over time. Because panel data are costly and complex to administer, they are rarely avail- able. It is also difficult to get a general understanding of mobility within and across countries, because the analysis depends on different methods, time periods, and welfare measures. The problem is compounded by the fact that most panels are usually not representative of a country’s entire population, and even if they are, nonrandom attrition is common in panel datasets and can significantly bias results (Antman and McKenzie 2007). Finally, panel datasets usually cover short periods of time, reducing their usefulness for understanding the long-term trends associ- ated with economic immobility. To overcome these concerns, researchers have focused on the “pseudo-panel” approach, which allows cohorts of individuals to be followed over several peri- ods of time. Pseudo-panels can help overcome the main limitation of panel data- sets, but they usually require significant structural assumptions (Dang and others 2014). Moreover, by aggregating average trends for a given cohort, this technique ignores intragroup mobility, a key element that is often even more relevant than aggregate mobility. 42 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN A few alternative approaches exist. The Multidimensional Poverty Index (MPI), based on Alkire and Foster (2011), is a summary measure that evaluates the number of deprivations (access gaps) across various welfare dimensions con- sidered relevant in a given country context (such as school attendance or access to sanitation, safe water, or electricity). Within this context, the chronically poor are individuals who experience gaps in access in more than a given number of those dimensions. Although this approach is highly appealing, because it is simple to estimate, it involves a number of subjective decisions (such as which dimen- sions to include or the number of gaps) that make this measure harder to interpret. Moreover, the approach misses the time dimension and presumes that there is an association between deprivation and chronicity. Although the two are clearly related, the MPI approach cannot quantify this association. López-Calva, Mitra, and Ortiz-Juarez (2013) and Castaneda and others (2013) propose a method for identifying chronic poverty based on the depth, complexity, and link to persistence. They show that individuals who are both multidimen- sionally and income poor are more likely to remain poor between two periods than people who are only income poor. Using this insight, they propose defining chronic poverty by combining income (monetary) poverty with the MPI.2 People are classified into four groups: people who are multidimensionally and income poor (the chronically poor); (b) people who are monetarily poor but not poor in the other dimensions; (c) people who are not poor in monetary terms but poor in nonincome dimensions; and (d) people who are not deprived in any indicator (the “better off ”). Castaneda and others (2013) validate the methodology using panel datasets for Chile, Mexico, and Peru. The main advantage of this method is that it allows researchers to identify the chronically poor without using information that follows them over time. A draw- back is that it does not capture other measures of mobility, such as the income change of people who remain in or exit poverty. Synthetic Panels and Chronic Monetary Poverty We use an innovative technique originally developed by Dang and others (2014) and improved by Dang and Lanjouw (2014) to understand chronic poverty between 2004 and 2012 in LAC. The basic steps are as a follows. Assume that two rounds of cross-sectional data are available on a country. A model of consumption (or income) can then be estimated based on the first round of cross-sectional data, using a specification that includes only time-invariant covariates (such as house- hold head gender, education) or (if available) retrospective covariates that can be recalled from one period to the other. Assuming that the underlying population in the first round is the same as the one in the second round, parameter estimates of consumption (or income) in the second round can be produced simply by using the same time-invariant and retrospective covariates to estimate the (unobserved) second period’s consumption (or income) for the same individuals/households WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 43 surveyed in the first round, using the coefficients from the first period. Depending on the choice of additional assumptions about the distribution of the residuals, one can apply a nonparametric approach (to derive an upper and lower bound for individual mobility) or a parametric approach (to derive point estimates of condi- tional probability of poverty transitions). Formally, let yit be round t household per capita consumption or income (where t = 1, 2) of household i, and let z represent the poverty line. The main goal is to estimate conditional probabilities of poverty transitions—the proportion of poor households that escaped poverty (Pr( yi2> z and yi1< z)) between the first and sec- ond period that remained poor (Pr( yi2 < z and yi1 < z)), became poor (Pr(yi2 < z and yi1 > z)), and remained nonpoor (Pr( yi2 > z and yi1 > z)). The linear projection of household consumption or income on household time-invariant characteristics xit of household i in round t is given by yit = β t′ x it + e it t = 1, 2 (2.1) Assuming that the error terms (ei1 and ei2) have a bivariate normal distribution with a nonnegative correlation coefficient r and standard deviations se1 and se 2, respectively, the probability of escaping poverty can be expressed as follow ⎛ z − β ′x z − β2′ x i 2 ⎞ P( yi1 < z1 and y i 2 > z2 ) = Φ2 ⎜ 1 1 i2 ,− 2 ,− ρ⎟ (2.2) ⎜ ⎝ σ e1 σ e2 ⎟ ⎠ where Φ2is a standard bivariate normal distribution function. If r is known, the probability of escaping poverty can be estimated by obtaining the predicted coefficients β ˆ ′ as well as the predicted standard error σ ˆ ′, β ˆ ε1 and σˆ ε 2 from 1 2 round 1 and 2, respectively. However, in practice, r is unknown. Dang and others (2014) offer two options. The first, a nonparametric approach, bounds the correlation between zero and one, which allows one to derive upper and lower bounds for individual mobility that are expected to sandwich true mobility.3 The second, a parametric approach, assumes normal- ity of the error terms and estimates the residual correlation between the two rounds using actual panel datasets, which can then be used to calculate a point estimate. Dang and Lanjouw (2014) improved the parametric technique to obtain a point estimate of mobility using the age-cohort correlation of residuals obtained from the same cross-sectional data at hand (instead of panel data). The additional assumption that cohort-based analysis is sufficient to provide a reasonable estimate of r means that a simple procedure can be used to obtain a point estimate. It con- sists of first calculating the simple correlation between the welfare measure of the age-cohorts ryl1,yl2. Once this correlation is computed, and using the fact that the 44 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN ryl1 , yl2 var ( yi1 ) var ( yi 2 ) − B1′ var ( x i ) B2 residual partial correlation equals , s e 1s e 2 researchers can estimate r as well as a household’s probabilities of income mobil- ity. These probabilities can then be aggregated into a country- or group-specific mobility estimate. How well does the approach work in practice? A number of validations using actual panel data have been conducted using both approaches. The results are compelling. Dang and others (2014) validate the nonparametric technique with data from Indonesia and Vietnam. Drawing from both cross-sectional and true panel data, they compare mobility estimates based on synthetic panels to estimates from actual panel data. They find that the “true” estimate of mobility (as revealed by the actual panel data) is generally sandwiched between the upper-bound and lower-bound estimates produced by the technique. Cruces and others (2015) validate the nonparametric technique in Chile, Nicaragua, and Peru, benefitting from the availability of actual panel data of vary- ing lengths. Their results, which include a wide range of sensitivity analyses and robustness checks, confirm the validity of the technique even for longer spans (as long as 10 years in the case of Chile). Other research that confirms the valid- ity of these techniques includes Bierbaum and Gassmann (2012) in the Kyrgyz Republic, Martinez and others (2013) in the Philippines, and Davalos and Meyer (2015) in Moldova. Dang and Lanjouw (2014) conduct extensive validation of the parametric method using actual panel data in a geographically diverse set of five countries with very different income levels (Bosnia-Herzegovina 2001–04, Lao People’s Democratic Republic 2002/03–2007/08, Peru 2005–06, the United States 2007–09, and Vietnam 2006–08).4 Their synthetic panel estimates fall within the 95 percent confidence interval—one standard error in many cases—of those based on actual panel data, further validating the appropriateness of the technique. Table 2.1 reports their results for Peru. Although this approach is not a substitute for actual panel datasets, in the absence of panel data it presents several advantages (Ferreira and others 2012). First, it allows conditional poverty probabilities to be constructed for every household in the sample. Aggregate poverty dynamics measures can thus be constructed based on household estimates instead of cohort averages, as in the pseudo-panel approach. Second, the approach can be applied in all countries for which two or more cross-sectional data points exist. Given the validation results, it allows us to consistently compare chronic poverty trends for 17 coun- tries in LAC. A few caveats are in order with respect to what the approach does and does not do. First, it can be used only to explore associations between the probabilities of transitional poverty and additional variables of interest; in its current form, it WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 45 TABLE 2.1 Poverty estimates for Peru based on actual and synthetic panel data, 2004–05 (percent of population) Poverty status in 2004 and 2005 Estimate from actual panel Estimate from synthetic panel Poor, poor 29.9 30.9 Poor, nonpoor 11.6 12.3 Nonpoor, poor 8.9 10.0 Nonpoor, nonpoor 49.7 46.8 Number of observations 2,250 9,084 Source: Dang and Lanjouw 2014. does not allow deeper causal relationships of chronic poverty or upward mobility to be explored. However, merely understanding the extent of chronic poverty in the region and its association with relevant socioeconomic aspects goes a long way toward filling the knowledge gap and exploring potential mechanisms and pol- icy channels, which can then be used to push the policy agenda for dealing with chronic poverty. We complement these associations with insights from poverty mapping techniques (discussed below) and experimental evidence to understand the channels through which chronic poverty operates. Second, the parametric approach—the approach we adopt in this book— requires a strong assumption. In order to move from bounds to point estimates of mobility, it requires the estimate of an additional parameter, r (the age-cohort partial residual correlations). Like the pseudo-panel approach, this approach assumes that cross-cohort differences alone can explain this correlation (that is, it rules out any role for within-cohort differences). The validation results of Dang and Lanjouw (2014) show that in the few countries for which panel data can be analyzed, the estimated r is very close to the directly measured r from the panels, suggesting that this assumption is not unreasonable. Although there is no guaran- tee that r will be correctly estimated in all cases, their results provide a measure of support. Third, the approach used in the book is based on just two rounds of data for each country, limiting the ability to claim that it captures the “true” poverty dynamics over the decade studied. Dynamics that occur between the two rounds are not captured.5 We adopt the parametric approach of the synthetic panel technique to obtain point estimates of conditional poverty probabilities for each household between 2004 and 2012. Following the literature, we apply the technique to households with an adult head and estimate conditional poverty transition probabilities based on income. This technique allows us to classify households into one of 46 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN four states: chronically poor (poor in both years), exited poverty (poor in 2004 but not poor in 2012), downwardly mobile (not poor in 2004 but poor in 2012), and never poor. We estimate aggregate poverty status in a consistent manner for 17 countries for the period 2004–12, focusing on the first and last years (or the nearest year in cases in which data for those years were unavailable). Focusing on these years allows us to capture the extensive welfare improvements in the region and maxi- mizes the number of countries and length of the period, allowing us to identify long-term trends. We use the Socio-Economic Database for Latin America and the Caribbean (SEDLAC) database of household surveys, compiled by the Center for Distributional Labor and Social Studies (CEDLAS) at the University de La Plata, Argentina and the World Bank (http://sedlac.econo.unlp.edu.ar /eng/). It includes harmonized information on a range of socioeconomic char- acteristics, including comparable incomes from 17 countries in the region: Argentina, Bolivia, Brazil, Chile, Colombia, Costa Rica, the Dominican Republic, Ecuador, El Salvador, Guatemala, Honduras, Mexico, Nicaragua, Panama, Paraguay, Peru, and Uruguay (see table A.1 in the appendix for the coverage of each survey). We define a household as chronically poor if its income remained below $4 a day (in 2005 purchasing power parity dollars) in both 2004 and 2012 using the synthetic panel approach. This threshold allows us to identify people whose income grew enough to lift them out of poverty. We also use complementary sources of data, as discussed in the relevant chapters. Capturing Chronic Poverty at High Levels of Geographical Disaggregation We use poverty mapping techniques to explore the role geographic factors play in poverty persistence in LAC. This methodology allows estimation of monetary poverty indicators at a higher level of geographical disaggregation than household surveys permit. Household surveys include information on consumption or income that allows poverty to be estimated, but they include too few household observations to do so at high levels of disaggregation. Censuses include the universe of households but generally do not collect consumption or income. The methodology developed by Elbers, Lanjouw, and Lanjouw (2003) combines information from surveys and censuses to estimate consumption or income poverty in small geographic areas, such as municipalities. Several steps are required to apply the technique. First, income or consumption econometric models are estimated using microdata from house- hold surveys. Second, the parameters estimated are exported to the census data to obtain estimates of household-level welfare, information not available in census data. Third, poverty measures are estimated for specific geographic WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 47 areas based on the welfare measure estimated in the second step. An important element of this methodology is that it takes into account the error associated with the imputation of consumption and produces standard errors for the esti- mates of poverty, which helps determine the degree of statistical precision of these estimates. In a first stage, household surveys are used to estimate the following model ln( ych) = x′ch b  + µch where ln( ych) is the logarithm of per capita consumption of household h in area c; x′ch is a vector of characteristics; b is a vector of regression coefficients; and µch is an error term composed of two parts, hc, which is common to all house- holds located in the same area c, and ech, which is specific to each household ( µch = hc + ech). Parameters and error terms are estimated using generalized least squares. In the second stage of the methodology, estimates of poverty along with their standard errors are computed using census data. Parameter estimates obtained in the first stage are used to simulate consumption for every household in the census data. This simulation is repeated R times, so that it can generate an estimate of the standard error of every poverty indicator. (For more details on this methodology, see Elbers, Lanjouw, and Lanjouw 2003.) When relevant, we draw evidence from Guatemala, Mexico, and Peru, where poverty map panels were recently estimated. These countries are geo- graphically diverse and have significant regional inequalities, making them excellent for illustrating the relationship between geographic conditions and chronic poverty. Notes 1. The debate over whether poverty traps exist in practice is ongoing (see, for example, Kraay and McKenzie 2014). 2. The authors consider the following dimensions: children’s school attendance; years of schooling; access to sanitation, safe water, electricity, and shelter; and an asset index. 3. Dang and others (2014) argue that the correlation between error terms will most likely be positive if the terms include an individual fixed effect and shocks to consumption persist over time. They present empirical support for this assumption. Cruces and others (2015) find positive correlation terms in Chile, Nicaragua, and Peru. 4. To avoid life-cycle effects, which could violate the time invariance assumption, they restrict estimates to households in which the head is between ages 25 and 65. 5. Research based on panel data faces the same limitations, as at best household-level panel data in a given setting exist for a finite number of rounds, seldom extending beyond two or three rounds covering a few years. 48 Left Behind: Chronic Poverty in Latin America and the Caribbean References Alkire, Sabina, and James Foster. 2011. “Counting and Multidimensional Poverty Measurement.” Journal of Public Economics 95 (7): 476–87. Antman, Francisca, and David J. McKenzie. 2007. “Earnings Mobility and Measurement Error: A Pseudo-Panel Approach.” Economic Development and Cultural Change 56 (1): 125–61. Appadurai, Arjun. 2004. “The Capacity to Aspire: Culture and the Terms of Recognition.” In Culture and Public Action, edited by Vijayendra Rao and Michael Walton, 59–84. Stanford, CA: Stanford University Press. Baird, Sarah, Craig McIntosh, and Berk Özler. 2011. “The Regressive Demands of Demand-Driven Development.” Policy Research Working Paper 5883, World Bank, Washington, DC. Bane, Mary Jo, and David T. Ellwood. 1986. “Slipping into and out of Poverty: The Dynamics of Spells.” Journal of Human Resources 21 (1): 1–23. Banerjee, Abhijit, and Sendhil Mullainathan. 2010. “The Shape of Temptation: Implications for the Economic Lives of the Poor.” NBER Working Paper 15973, National Bureau of Economic Research, Cambridge, MA. Barrett, Christopher B., and Michael R. Carter. 2013. “The Economics of Poverty Traps and Persistent Poverty: Empirical and Policy Implications.” Journal of Development Studies 49 (7): 976–90. Barros, Ricardo, Francisco H. G. Ferreira, José Molinas, and Jaime Saavedra. 2009. Measuring Inequality of Opportunities in Latin America and the Caribbean. Washington, DC: World Bank. Bierbaum, Mira, and Franziska Gassmann. 2012. “Chronic and Transitory Poverty in the Kyrgyz Republic: What Can Synthetic Panels Tell Us?” UNU-MERIT, Maastricht University School of Business and Economics, Maastricht, the Netherlands. Calvo, Cesar, and Stefan Dercon. 2009. “Chronic Poverty and All That.” Poverty Dynamics: Interdisciplinary Perspectives, edited by Tony Addison, David Hulme, and Ravi Kanbur, 29–59. Oxford, U.K.: Oxford University Press. Carter, Michael R., and Christopher B. Barrett. 2006. “The Economics of Poverty Traps and Persistent Poverty: An Asset-Based Approach.” Journal of Development Studies 42 (2): 178–99. Carter, Michael, and Munenobu Ikegami. 2009. “Looking Forward: Theory-Based Measures of Chronic Poverty and Vulnerability.” In Poverty Dynamics: Interdisciplinary Perspectives, edited by Tony Addison, David Hulme, and Ravi Kanbur, 128–53. Oxford, U.K.: Oxford University Press. Castaneda, A., A. Fruttero, L. F. López-Calva, and M. A. Lugo. 2013. Chronic Poverty in Brazil: Understanding the Patterns and Policy Challenges. Washington, DC: World Bank. Clark, David A., and David Hulme. 2005. “Towards A Unified Framework for Understanding the Depth, Breadth and Duration of Poverty.” GPRG-WPS-020, Global Poverty Research Group, Oxford. Conel, J. Le R. 1939. The Postnatal Development of the Human Cerebral Cortex. Vol. 1. The Cortex of the Newborn. Cambridge, MA: Harvard University Press. Cord, L., and L. F. López-Calva. 2012. “The Equity Agenda in Latin America and the Caribbean: Opportunities.” World Bank, Washington, DC. WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 49 Cruces, Guillermo, Peter Lanjouw, Leonardo Lucchetti, Elizaveta Perova, Renos Vakis, and Mariana Viollaz. 2015. “Intra-Generational Mobility and Repeated Cross- Sections: A Three-Country Validation Exercise.” Journal of Economic Inequality 13 (2): 161–79. Dang, Hai-Anh, and Peter Lanjouw. 2014. “Measuring Poverty Dynamics with Synthetic Panels Based on Cross-Sections.” World Bank, Washington, DC. Dang, Hai-Anh, Peter Lanjouw, Jill Luoto, and David McKenzie. 2014. “Using Repeated Cross-Sections to Explore Movements into and out of Poverty.” Journal of Development Economics 107: 112–28. Davalos, Maria E., and Moritz Meyer. 2015. “Moldova: A Story of Upward Economic Mobility.” Policy Research Working Paper 7167, World Bank, Washington, DC. Duckworth, Angela L., Christopher Peterson, Michael D. Matthews, and Dennis R. Kelly. 2007. “Grit: Perseverance and Passion for Long-Term Goals.” Journal of Personality and Social Psychology 92 (6): 1087–101. Duflo, Esther. 2012. “Human Values and the Design of the Fight against Poverty.” Tanner Lecture, Massachusetts Institute of Technology, Cambridge, MA. https://www .povertyactionlab.org/sites/default/files/documents/TannerLectures_EstherDuflo _draft.pdf. Dweck, Carol. 2006. Mindset: The New Psychology of Success. New York: Random House. Elbers, Chris, Jean O. Lanjouw, and Peter Lanjouw. 2003. “Micro-Level Estimation of Poverty and Inequality.” Econometrica 71 (1): 355–64. Ferreira, Francisco H. G., Julian Messina, Jamele Rigolini, Luis-Felipe López-Calva, Maria Ana Lugo, and Renos Vakis. 2012. Economic Mobility and the Rise of the Latin American Middle Class. Washington, DC: World Bank. Foster, James. 2007. “Report on Multidimensional Poverty Measurement.” El Colegio de Mexico, Mexico City. ———. 2009. “A Class of Chronic Poverty Measures.” Poverty Dynamics: Interdisciplinary Perspectives, edited by Tony Addison, David Hulme, and Ravi Kanbur, 59–76. Oxford, U.K.: Oxford University Press. Foster, James E., and María Emma Santos. 2013. “Measuring Chronic Poverty.” In Poverty and Social Exclusion: New Methods of Analysis, edited by Gianni Betti and Achille Lemmi. New York: Routledge. Gradín, Carlos, Coral del Río, and Olga Cantó. 2012. “Measuring Poverty Accounting for Time.” Review of Income and Wealth 58 (2): 330–54. Hickey, Sam, and Sarah Bracking. 2005. “Exploring the Politics of Chronic Poverty: From Representation to a Politics of Justice?” World Development 33 (6): 851–65. Hoy, Michael, and Buhong Zheng. 2011. “Measuring Lifetime Poverty.” Journal of Economic Theory 146 (6): 2544–62. Hulme, David, and Andrew Shepherd. 2003. “Conceptualizing Chronic Poverty.” World Development 31 (3): 403–23. Jalan, Jyotsna, and Martin Ravallion. 1998. Determinants of Transient and Chronic Poverty: Evidence from Rural China. Washington, DC: World Bank. ———. 2000. “Is Transient Poverty Different? Evidence for Rural China.” Journal of Development Studies 36 (6): 82–99. Jensen, Robert. 2010. “The (Perceived) Returns to Education and the Demand for Schooling.” Quarterly Journal of Economics 125 (2): 515–48. 50 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Kraay, Aart, and David McKenzie. 2014. “Do Poverty Traps Exist? Assessing the Evidence.” Journal of Economic Perspectives 28 (3): 127–48. Lipton, Michael. 1983. “Poverty, Undernutrition, and Hunger.” Working Paper 597. World Bank, Washington, DC. ———. 1986. “Seasonality and Ultrapoverty.” IDS Bulletin 17 (3): 4–8. Locke, Edwin A., and Gary P. Latham. 2002. “Building a Practically Useful Theory of Goal Setting and Task Motivation: A 35-Year Odyssey.” American Psychologist 57 (9): 705–17. López-Calva, L. F., Shabana Mitra, and E. Ortiz-Juarez. 2013. “If Life Gives You Lemons… Using Multidimensional Poverty Indices to Approximate Chronic Poverty in the Absence of Panel Data.” World Bank, Washington, DC. Macours, Karen, and Renos Vakis. 2014. “Changing Households’ Investments and Aspirations through Social Interactions: Evidence from a Randomized Transfer Program.” Economic Journal 124 (576): 607–33. Martinez, Arturo, Jr., Mark Western, Michele Haynes, and Wojtek Tomaszewski. 2013. “Measuring Income Mobility Using Pseudo-Panel Data.” Philippine Statistician 62 (2): 7 1–99. Mendola, Daria, Annalisa Busetta, and Anna Maria Milito. 2011. “Combining the Intensity and Sequencing of the Poverty Experience: A Class of Longitudinal Poverty Indices.” Journal of the Royal Statistical Society: Series A (Statistics in Society) 174 (4): 953–73. Mullainathan, Sendhil, and Eldar Shafir. 2013. Scarcity: Why Having Too Little Means So Much. New York: Macmillan. Perova, Elizaveta, and Renos Vakis. 2013. Improving Gender and Development Outcomes through Agency: Policy Lessons from Three Peruvian Experiences. World Bank Other Operational Studies 16259, Washington, DC. Porter, Catherine, and Natalie Naïri Quinn. 2008. “Intertemporal Poverty Measurement: Tradeoffs and Policy Options.” Working Paper, University of Oxford, Oxford, U.K. Radner, J., and J. Shonkoff. 2012. “Mobilizing Science to Reduce Intergenerational Poverty.” In Investing in What Works for America’s Communities: Essays on People, Place and Purpose, edited by N. O. Andrews, D. J. Erickson, I. J. Galloway, and E. S. Seidman, 338–50. San Francisco: Federal Reserve Bank of San Francisco and the Low Income Investment Fund. Ray, Debraj. 2006. “Aspirations, Poverty, and Economic Change.” In Understanding Poverty, edited by Abhijit Vinayak Banerjee, Roland Bénabou, and Dilip Mookherjee, 409–21. Oxford Scholarship Online. Roemer, John E. 1998. Theories of Distributive Justice. Cambridge, MA: Harvard University Press. Ruggles, Patricia, and Roberton Williams. 1989. “Longitudinal Measures of Poverty: Accounting for Income and Assets over Time.” Review of Income and Wealth 35 (3): 225–43. Sen, Amartya. 1983. Poverty and Famines: An Essay on Entitlement and Deprivation. Oxford, U.K.: Oxford University Press. ———. 1985. “Well-Being, Agency and Freedom: The Dewey Lectures 1984.” Journal of Philosophy 82 (4): 169–221. ———. 1999. Development as Freedom. Oxford, U.K.: Oxford University Press. WHAT IS CHRONIC POVERTY AND HOW IS IT MEASURED? 51 Wilson, William J. 1987. The Truly Disadvantaged: The Inner City, the Underclass, and Public Policy. Chicago: University of Chicago Press. Wood, Geof. 2003. “Staying Secure, Staying Poor: The ‘Faustian Bargain.’” World Development 31 (3): 455–71. World Bank. 2013. Inclusion Matters: The Foundation for Shared Prosperity. Washington, DC: World Bank. Yaqub, Shahin. 2000. Poverty Dynamics in Developing Countries, vol. 16. Institute of Development Studies, Brighton, United Kingdom. 52 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN CHAPTER 3 Five Facts about Chronic Poverty in Latin America and the Caribbean A review of recent trends in chronic poverty reveals five stylized facts. Stylized Fact 1: One of Five People in Latin America and the Caribbean Lives in Chronic Poverty Cross-sectional data reveal that one in four poor people in Latin America and the Caribbean (LAC) was poor in 2012. But how many of them were chronically poor? In 2012, 21.6 percent, or one in five poor people in the region, had also been poor in 2004 (table 3.1). This figure implies that 130 million people—about half of all poor people in 2012—were chronically poor. The rest of the poor— about 8 percent of the population that was nonpoor in 2004—fell into poverty over this period. These figures reveal that despite the extraordinary success in reducing poverty in the region, many people were left behind, either staying or becoming poor. Chronic poverty and poverty dynamics vary considerably across countries in the region. Figure 3.1 shows the share of the poor who were poor in both 2004 and 2012 and the share that fell into poverty by 2012 for each country (table A.2 in the appendix provides the full set of results of the synthetic panel estimates of poverty transitions for each country based on the parametric estimator discussed in chapter 2). It shows wide heterogeneity across countries. Uruguay, Argentina, and Chile have the lowest rates of chronic poverty, with chronic poverty rates of about 10 percent. Nicaragua, Honduras, and Guatemala have the highest rates of chronic poverty, ranging from 37 percent in Nicaragua to 50 percent in Guatemala, significantly higher than the regional average of 21 percent. 53 TABLE 3.1 Movement in and out of poverty in Latin America and the Caribbean between 2004 and 2012 (percent of population) 2012 Poor Nonpoor Total 2004 Poor 21.6 23.4 44.9 Nonpoor 4.2 50.9 55.1 Total 25.7 74.2 100.0 Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are population-weighted averages of country-specific estimates, which are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. FIGURE 3.1 Chronic poverty and downward mobility in selected countries in Latin America and the Caribbean, 2004–12 60 Percent of population 40 20 0 be nd Ca ric ico Ho ua la Ni bia Co ile na Pa ica ma Gu uras Ec lic mi Sa u Ar uay Pa y Re r Co dor ia zil an do ua Do El Per ma liv Ch b an rib a a rag nti a the Am Mex R lom na pu nic lva ua ug rag Br Bo nd ate ge sta ca Ur e tin La Downwardly mobile Chronically poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates of poverty at the regional level are population-weighted averages of country-specific estimates, which are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Total poverty does not necessarily match country-specific poverty rates, which are estimated based on the synthetic panel approach, which uses a subpopulation of households with adult household heads. Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. Chronically poor people are people whose income remained below $4 between 2004 and 2012. Downwardly mobile people are people who were nonpoor in 2004 but poor in 2012. 54 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN In a few countries, downward mobility contributed to poverty generation. In Honduras, for instance, the poverty rate was 57 percent in 2012. Of this 57 percentage points, 42 percentage points were people who were chronically poor and 15 percentage points were people who became poor between 2004 and 2012. Put another way, 45 percent of people who were not poor in 2004 became poor by 2012 (see table A.2 in the appendix). For the region as a whole, the picture is heterogeneous. Overall, 8 per- cent of the nonpoor fell into poverty during this period, but seven countries (Brazil, Chile, Colombia, Ecuador, Panama, Peru, and Uruguay) experienced relatively low levels of downward mobility (less than 5 percent). In six coun- tries (Bolivia, the Dominican Republic, El Salvador, Honduras, Nicaragua, and Paraguay), more than 25 percent of the nonpoor fell into poverty, with more than 40 percent of the nonpoor in Nicaragua and Honduras becoming poor between 2004 and 2012. Despite the strong growth rates most countries in the region experienced, millions of people fell into poverty, especially in noncommodity boom countries (see box 3.1 for a discussion of other dimensions of poverty during this period). BOX 3.1 Monetary versus nonmonetary measures of chronic poverty Chronic poverty is a complex multidimensional phenomenon. In this book we measure it largely by income. To show why doing so makes sense, we compare our monetary chronic poverty measure based on the synthetic panel approach with two related concepts. For all the countries we study, we estimated multidimensional poverty indexes (MPI) using the methods proposed by Alkire and Foster (2011). The MPI is based on three equally weighted poverty dimensions—health, education, and living standards—which are captured by 10 indicators. A person who is deprived in four or more of the weighted attributes is considered poor. The value of the index reflects the number of dimensions in which a person is poor; it ignores the depth of deprivation below the cutoff. This approach is transparent and simple, and as Alkire and Foster (2011) show, the resulting poverty indicator satisfies a number of very desirable axiomatic properties. Panel a of box figure B3.1.1 shows the relationship between the monetary chronic poverty used in this book and the MPI. It shows a remarkably high (R 2 = 0.7) correlation between the two measures, suggesting that these measures capture some (albeit not all) of the dynamics we are interested in understanding. Panel b shows a similar association between the synthetic panel approach we use and the extension by Castaneda and others (2013), which uses the intersection of MPI and income poverty to define chronic poverty. The two measures are highly correlated (R 2 = 0.7), even though the absolute levels of chronic poverty estimated by applying the Castaneda and others approach are lower, driven mainly by the likely influence the MPI has in the classification of the chronically poor. (continued on next page) FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 55 BOX 3.1 Monetary versus nonmonetary measures of chronic poverty (continued) FIGURE B3.1.1 Correlation between estimates of chronic poverty and multidimensional poverty in selected countries in Latin America and the Caribbean a. Chronic poverty and multidimensional poverty index (MPI) poverty estimates 60 50 Guatemala MPI poverty percent of population 40 Nicaragua 30 Honduras Bolivia El Salvador 20 Mexico Dominican Republic Peru 10 Colombia Chile Paraguay Ecuador Brazil Costa Rica 0 Uruguay 0 10 20 30 40 50 60 Chronic poverty percent of population b. Chronic poverty and MPI and monetary poverty estimates 60 Intersection of MPI and monetary poverty percent Guatemala 50 Nicaragua 40 of population 30 Honduras 20 Bolivia El Salvador 10 Mexico Colombia Costa Rica Peru Dominican Republic Uruguay Paraguay Ecuador 0 Brazil 0 10 Chile 20 30 40 50 60 Chronic poverty percent of population Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Averages in panel b use the classification proposed by Castaneda and others (2013). Chronic poverty refers to the proportion of people who are deprived along both monetary and nonmonetary dimensions (people who are multidimensionally poor). Argentina and Panama are excluded, because data limitations prevented MPI poverty rates from being calculated. 56 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Stylized Fact 2: Chronic Poverty Tends to Be Geographically Concentrated Geography is an extremely important factor for understanding chronic poverty. Figure 3.2 presents chronic poverty rates at the level of subnational regions in LAC (dividing each country into subnational regions for a total of 168 regions across LAC).1 Overall, poverty rates in LAC fell over the last decade, but the level of chronic poverty ranges widely both within countries and across LAC. In Brazil, for example, Santa Catarina has a chronic poverty rate of about 5 percent, which is lower than the national average of 20 percent. This rate is close to the average for Uruguay, the LAC country with the lowest rate of chronic poverty. By contrast, about 40 percent of the population in Ceará, Brazil, is chronically poor—twice the Brazilian average. This rate is close to the average for Honduras, one of the coun- tries with the highest chronic poverty rate in the region. This concentration can be further disaggregated using the poverty mapping techniques described in chapter 2. Map 3.1 presents a district-level map of Peru, classifying each of its 1,838 districts according to the change in poverty between 2007 and 2012. About 40 percent of these districts (740) had poverty rates above 50 percent in both periods. Ten percent (179) had persistent poverty rates well FIGURE 3.2 Subnational chronic poverty rates in Latin America and the Caribbean, 2012 100 90 Percent of population that is chronically poor 80 70 60 50 Ceará (Brazil) 40 Latin America and the Caribbean 30 Santa Catarina 20 (Brazil) Tocantins (Brazil) 10 0 0 20 40 60 80 100 Ranking Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Horizontal axis presents a normalized ranking of regions (out of a total of 168) based on their chronic poverty rates. Selection of subnational regions varies depending on level of representativeness of surveys. Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 57 MAP 3.1 Chronic poverty in Peru, by district, 2012 District’s poverty status Below 50 percent in both periods Worsened Improved Chronically poor Sources: 2007 and 2012 Peru Census and ENAHO data. Note: Below 50 percent in both periods = districts in which moderate poverty rate was below 50 percent in both 2007 and 2012. Worsened = districts in which moderate poverty rate was below 50 percent in 2007 but above 50 percent in 2012. Improved = districts in which moderate poverty rate was above 50 percent in 2007 but below 50 percent in 2012. Chronically poor = districts in which moderate poverty rates were above 50 percent in both 2007 and 2012. 58 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN above 75 percent. Twenty-four percent of all districts (439) with poverty rates above 50 percent in 2007 saw poverty rates fall below 50 percent by 2012. Looking exclusively at poverty rates provides only half the picture, how- ever, because a large number of the chronically poor may reside in densely populated areas with relatively low rates of chronic poverty (regional and national capitals). Figure 3.3 presents the cumulative distribution (Lorenz curve) of chroni- cally poor households across subnational regions (the vertical axis), sorted by the contribution of each region to the absolute number of chronically poor (horizontal axis). It shows that a large share of the chronically poor population is concentrated in a few regions, with half of them living in 20 of the 187 regions considered. Moreover, the regions where the majority of the chronically poor reside are not necessarily the regions with the highest rates of chronic poverty. In Mexico, for instance, the incidence of chronic poverty in the Distrito Federal is equal to the LAC regional average, but because of its large population, the capital hosts almost 3 percent of LAC’s chronically poor, making it the sixth largest contributor to chronic poverty in LAC. In contrast, Baja California, where the concentration of chronic poverty is much higher, is home to only 0.2 percent of LAC’s chronically poor. FIGURE 3.3 Concentration of chronic poverty in Latin America and the Caribbean, 2012 100 Distrito Federal 90 (Mexico) Cumulative distribution of percent of population 80 70 Chiapas (Mexico) that is chronically poor 60 50 40 30 20 Baja California (Mexico) 10 0 0 20 40 60 80 100 Ranking Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Selection of subregions varies depending on level of representativeness of surveys. FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 59 Stylized Fact 3: Chronic Poverty Is As Big a Problem in Urban Areas as in Rural Areas The concentration of chronic poverty is greater in rural areas than in urban areas (figure 3.4, panel a).2 In 2012, for example, chronic poverty affected about 20 percent of Bolivia’s entire population, 31 percent of its rural population, and 15 percent of its urban population; in Panama more than 40 percent of the rural pop- ulation is chronically poor—five times more than in urban areas. Despite the much higher rates of chronic poverty in rural areas, chronic pov- erty is as much an urban as a rural issue, because more people live in urban areas. In panel b of figure 3.4, ratios above 1 indicate that the number of chronically poor people is higher in urban areas than in rural areas. In five countries (Chile, Brazil, Mexico, Colombia, and the Dominican Republic), the number is higher in urban areas. In Brazil, for example, for every chronically poor person in a rural area, two live in urban areas. The number of chronically poor is about equal in urban and rural areas in many other countries These insights partly reflect the level of urbanization across the region. As figure 3.4 shows, there is a positive correlation between the urban/rural chronic poverty ratio and the level of urbanization. In countries that are more urbanized, more of the chronically poor live in urban areas. In Chile, for example, which is 87 percent urbanized, there are six chronically poor people in urban areas for every one in a rural area. In contrast, Guatemala, which is 47 percent urbanized, has an urban to rural ratio of just 0.5. There are a few exceptions to this general rule. In Peru and Panama, for example, both of which are highly urbanized, more chroni- cally poor people live in rural areas than in urban areas. Chronically poor households tend to have better economic opportunities in urban areas than in rural areas (table 3.2). On average, heads of households in urban areas have six years of education, more than two years more than their rural counterparts. They also tend to come from smaller families. There are no large differences in the age composition of household heads in urban and rural areas. Per capita income in urban areas is about 20 percent higher than in rural areas. In 2004 the gap was 30 percent, suggesting some catching up by rural households. Chronically poor households in urban areas are more likely to have a head of household working as an employee (58 percent), whereas their rural counterparts are more likely to have a self-employed head of household (37 percent [results not shown]). Urban household heads are more likely to work in commerce (27 percent), whereas rural household heads are substantially more likely to work in agriculture (74 percent). Among chronically poor households, women account for a larger share of household income in urban households (33 percent) than in rural house- holds (9 percent), denoting better labor market opportunities for women. These findings have some implications for the design of targeting schemes and antipoverty strategies. Countries with high concentrations of chronic poverty 60 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 3.4 Chronic poverty in rural and urban areas in selected countries in Latin America and the Caribbean, 2012 a. Share of rural and urban population that is chronically poor 80 70 60 Percent of urban population that is chronically poor 50 40 Guatemala Nicaragua Honduras 30 Dominican Republic Colombia 20 El Salvador Ecuador Paraguay Peru Brazil 10 Mexico Bolivia Chile Costa Rica Panama 0 0 10 20 30 40 50 60 70 Percent of rural population that is chronically poor b. Urbanization rate and ratio of number of people who are chronically poor in urban versus rural areas 7 6 Chile Ratio of number of chronically poor 5 people in urban to rural areas 4 3 2 Brazil Costa Rica Dominican Republic Mexico Paraguay Bolivia Colombia 1 Honduras Nicaragua Ecuador Guatemala El Salvador Peru 0 Panama –1 –2 45 55 65 75 85 95 Urbanization rate (percent) Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Vertical line indicates average urbanization rate in the region. FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 61 TABLE 3.2 Characteristics of the chronically poor in urban and rural areas of Latin America and the Caribbean, 2012 Characteristic Urban Rural Demographics Household size 4.5 5.0 Age of household head 45.4 45.9 Education Years of education of household head 6.4 3.9 Income Initial median per capita income, 2004 (dollars) 2.6 1.9 Initial median per capita income, 2012 (dollars) 2.9 2.4 Labor income as share of nonlabor income 0.3 0.3 Income generated by women as share of total income 32.7 18.7 Labor market (percent of people working in sector) Primary (agriculture and livestock) 15.7 72.9 Manufacturing 11.8 6.1 Construction 10.9 3.5 Services 14.7 4.1 Retail 30.6 8.9 Utilities 5.7 1.4 Domestic services 10.5 3.1 Source: Calculations based on SEDLAC data (CEDLAS and World Bank). in rural areas may require a different approach than countries in which chronic poverty is largely an urban phenomenon. Context-specific policy instruments and approaches are needed. Stylized Fact 4: Economic Growth Was Not Sufficient to Lift the Chronically Poor out of Poverty LAC witnessed impressive and unprecedented economic growth during the 2000s, which was key for reducing poverty. The annual rate of poverty reduction was 1.8 percentage points between 2004 and 2012, about 70 percent of which can be attributed to growth (the remaining 30 percent came from the decline in income inequality) (World Bank 2014). Growth did not help the chronically poor much, however, for two reasons. First, countries with the highest rates of chronic poverty grew the least (figure 3.5). 62 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Economic growth in Guatemala, where about half the population was stuck in poverty, was less than 1 percent a year. In contrast, in Panama, where 20 percent of the population was stuck in poverty, the economy grew 6 percent a year. Second, and perhaps more important, the chronically poor did not benefit as much from economic growth. The distribution of income among people who escaped poverty in Peru is considerably closer to the poverty line (panel b in figure 3.6) than the distribution of income of people who remained poor (panel a in figure 3.6), which suggests that it is the “richer” among the poor who were more likely to exit poverty. In this sense, the sustained rates of economic growth experi- enced in the past decade dealt with the “easy” part of the poor—people that were arguably closer to the poverty line, for whom a small increase in incomes allowed them to exit poverty. These results extend to the rest of the region. People who were poor in 2004 but escaped poverty by 2012 initially earned $3 a day (in 2005 purchasing power parity terms); by 2012 their median income had doubled to $6 a day (figure 3.7). The incomes of the chronically poor also increased, from $1.50 to $2.70 a day, but the smaller increase was not enough to lift them out of poverty. For the region as a whole, annual income growth was systematically lower among the chroni- cally poor (7.9 percent) than among people who escaped poverty (9.0 percent). This trend also holds within countries, where income growth among people who FIGURE 3.5 Correlation between per capita GDP growth and chronic poverty in selected countries in Latin America and the Caribbean, 2012 60 Percent of population that is chronically poor 50 Guatemala 40 Honduras Nicaragua Colombia 30 Ecuador Dominican Republic El Salvador Bolivia Peru 20 Latin America and the Caribbean Mexico Brazil Panama Paraguay Costa Rica 10 Chile Uruguay 0 0 1 2 3 4 5 6 7 Annualized percentage change in GDP per capita Source: Calculations based on SEDLAC data (CEDLAS and World Bank). FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 63 FIGURE 3.6 Initial (2004) income distribution of chronically poor and people who escaped poverty by 2012 in Peru a. Chronically poor b. Escaped poverty 0.8 0.8 0.6 0.6 Percent Percent 0.4 0.4 0.2 0.2 0 0 0 2 4 0 2 4 Daily per capita income Daily per capita income (2005 purchasing power parity dollars) (2005 purchasing power parity dollars) Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Figure shows lower-bound mobility estimates in Peru using the Dang and others (2014) technique. FIGURE 3.7 Median income in selected countries in Latin America and the Caribbean, 2012 a. People who exited poverty 8 Daily per capita income power parity dollars) (2005 purchasing 6 4 2 0 Ho ala Ni ico mi Co le Ec a na ia Re ia Pa a ma Co lic Me s Pa eru Sa y y ate r Ur or il a Gu do Ric u El ua ua i liv b az Ch b ur d nti rag x m nic lom na P pu ua lva Bo rag ug Br nd ge sta ca Ar an Do (continued on next page) 64 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 3.7 Median income in selected countries in Latin America and the Caribbean, 2012 (continued) b. People who remained poor 8.0 Daily per capita income 6.0 power parity dollars) (2005 purchasing 4.0 2.0 0 –2.0 Co hile Re a Ho ala Ni ico a mi Cos bia ia Pa a ma Ec lic as Pa eru Sa y y Gu dor Ur or il an Ric tin u El ua ua liv az b ur d rag x m lom na P C pu ua lva en Bo rag ug Me Br nd ate nic ta ca g Ar Do Initial income (median) Change in median income Median income in Latin America Median income in Latin America and the Caribbean, 2012 and the Caribbean, 2004 Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Figure is updated version of figure 4.5 in Ferreira and others (2012). Note: Figure shows lower-bound mobility estimates using the Dang and others (2014) technique. Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. escaped poverty exceeded that of the chronically poor in all but three countries (Argentina, Brazil, and the Dominican Republic), in some cases by large margins. In Paraguay, for example, the incomes of people who escaped poverty grew by an annualized rate of 8 percent—25 percent more than the 6 percent increase among the chronically poor. In Nicaragua, incomes of people who escaped poverty grew by 6 percent, compared with only 4 percent among the chronically poor. Growth thus played a useful but limited role for the chronically poor. Sustained growth helped lift millions of individuals out of poverty—but it also left many behind. In the context of modest prospects for future economic growth in the region and still high levels of inequality, it is therefore unlikely that growth alone will help the chronically poor escape poverty in the near future. FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 65 Stylized Fact 5: The Chronically Poor Have Limited Income Opportunities Labor income was by far the biggest driver behind the significant reduction in poverty between 2004 and 2012. We explore prospects going forward by look- ing at the income-generation potential of the chronically poor, people who exited poverty, and people who were never poor during the past decade. Figure 3.8 shows the number of people in a household with labor income in the three groups during 2012. In every country there were fewer labor income earners among chronically poor households (about 1.5) in 2004 than in either of the other two groups (about 1.8). Chronically poor households thus had 20 percent less human resources to generate income than the rest of the population. A similar story emerges for female labor market participation. Gender equality is crucial for poverty reduction, because greater economic oppor- tunities for women can enhance broader productivity gains and households’ economic perspectives (World Bank 2012). Recent trends are encouraging. Since the 1980s, there has been a significant increase in female labor market participation in developing countries, with LAC enjoying the largest increase. FIGURE 3.8 Number of labor income earners in household in selected countries in Latin America and the Caribbean, by poverty group, 2012 2.5 Average number of labor income 2.0 earners in household 1.5 1.0 0.5 0 a ua Ca ica via nic Ecu a Pa xico Ho ala bia ile Ar blic na ca s El Peru Me r Co ay Pa uay Re or Co azil be nd do Ni ura Ric m Ch an ad rag nti u i tem lom na rib a l pu Ur n lva Br the Ame Bo rag ug nd a sta ge Sa a Gu r tin mi La Do Never poor Exited poverty Chronically poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are based on 2004 and 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. 66 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 3.9 Female labor force participation and chronicity of poverty in selected countries in Latin America and the Caribbean, 2012 80 70 women 25–65 (percent of labor force) Labor force participation among 60 50 40 30 20 10 0 ru Ur ia Co uay Pa bia y co Ca ric zil Gu ean d Ec ala Ni dor Pa a ge a Sa a Ho dor as mi ost ile Re a c ua bli u m El ntin an Ric b n liv Pe the Am Bra Ch ur i rag rib a a x m lom na ua lva pu ug rag Bo Me nd ate nic a ca Ar e C tin La Do Never poor Exited poverty Chronically poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are based on 2004 and 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. Indeed, more than 70 million women entered the labor force in LAC since the 1980s (World Bank 2012). Colombia witnessed an increase of almost 20 percentage points. The chronically poor benefitted little from such trends, however. In 2012, female labor force participation lagged male participation in almost every coun- try in the region among chronically poor households (figure 3.9). Average female labor force participation was 48 percent among chronically poor households, 16 percentage points lower than among the nonpoor. It is similar (at about 58 per- cent) among the nonpoor and the poor who escaped poverty. Limited labor market participation is only one characteristic of the chroni- cally poor. An equally important characteristic is their concentration in lower- productivity sectors. Figure 3.10 relates the proportion of chronically poor people in the 168 subnational regions of LAC with the share of the labor force employed in various sectors. The results are consistent with limited opportuni- ties for the chronically poor, especially in sectors that tend to be associated with higher income potential. Regions with high concentrations of people employed in agriculture tend to have higher rates of chronic poverty, while high-tech industry, FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 67 FIGURE 3.10 Correlation between chronic poverty and sector of employment in selected countries in Latin America and the Caribbean, 2012 a. Primary b. Low-tech industry Percent of population that Percent of population that 1.0 1.0 is chronically poor is chronically poor 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 20 40 60 80 0 10 20 Percent of labor force employed in sector Percent of labor force employed in sector c. High-tech industry d. Utilities and service Percent of population that Percent of population that 1.0 1.0 is chronically poor is chronically poor 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 20 40 0 20 40 60 80 Percent of labor force employed in sector Percent of labor force employed in sector e. Construction f. Retail Percent of population that Percent of population that 1.0 1.0 is chronically poor is chronically poor 0.8 0.8 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 20 40 0 20 40 Percent of labor force employed in sector Percent of labor force employed in sector 95% Confidence Interval Fitted values Andinos Cono Sur Brazil Mexico Central America Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are based on 2004 and 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Cono Sur includes Argentina, Chile, Paraguay, and Uruguay. Andinos includes Bolivia, Colombia, Ecuador, and Peru. 68 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 3.11 Correlation between chronic poverty and rate of growth of formal sector in Latin America and the Caribbean, 2012 0.20 Percent of population that is chronically poor 0.15 0.10 0.05 0 –0.05 –0.10 –0.15 –0.20 –0.25 –0.30 –0.3 –0.2 –0.1 0 0.1 0.2 0.3 Change in formality between 2004 and 2012 (percent) Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. FIGURE 3.12 Correlation between chronic poverty and dependence on nonlabor income in Latin America and the Caribbean, 2012 50 45 Share of nonlabor income (percent) 40 35 30 25 20 15 10 5 0 il Ca a an lic Co bea the Sa o ma ua Me la na Co dor Pa ay Am R ru Pa dor Ni uay Ar ile Ec as Gu bia ia Ho ica az El xic ma c b liv sta n e rag u nti Ch ur na Br eri epu ua lom tin ica P lva R ug rib d rag Bo ate nd ge ca Ur n La inm Do Chronically poor Exited poverty Never poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. FIVE FACTS ABOUT CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN 69 services, construction, and retail are more likely to employ more people in regions with lower chronic poverty. Informality is also associated with chronic poverty (figure 3.11). People who are chronically poor are also more dependent on nonlabor income than people who escape poverty and the nonpoor (figure 3.12). On average about a third of the income of the chronically poor comes from nonlabor sources. In the majority of countries in LAC, this share is twice as large as it is for people who escaped poverty and people who were never poor. The chronically poor thus have a harder time generating income. Notes 1. The level of representativeness differs considerably across countries. For example, the ENAHO (Encuesta Nacional de Hogares) survey in Peru is representative at the level of departments (25 departments in total), whereas the ENEMDU (Encuesta de Empleo, Desempleo y Subempleo) labor force survey in Ecuador is representative at the regional level (six regions in total). 2 . Argentina and Uruguay are excluded from this analysis because only urban data were available. See table A.1 in the appendix for data sources. References Alkire, Sabina, and James Foster. 2011. “Counting and Multidimensional Poverty Measurement.” Journal of Public Economics 95 (7): 476–87. Castaneda, A., A. Fruttero, L. F. López-Calva, and M. A. Lugo. 2013. Chronic Poverty in Brazil: Understanding the Patterns and Policy Challenges. Washington, DC: World Bank. Dang, Hai-Anh, Peter Lanjouw, Jill Luoto, and David McKenzie. 2014. “Using Repeated Cross-Sections to Explore Movements into and out of Poverty.” Journal of Development Economics 107: 112–28. Ferreira, Francisco H. G., Julian Messina, Jamele Rigolini, Luis-Felipe López-Calva, Maria Ana Lugo, and Renos Vakis. 2012. Economic Mobility and the Rise of the Latin American Middle Class. Washington, DC: World Bank. World Bank. 2012. World Development Report 2013: Jobs. Washington, DC: World Bank. ———. 2014. “Social Gains in the Balance: A Fiscal Challenge for Latin America and the Caribbean.” LCSPP Poverty and Labor Brief 6, World Bank, Latin America and the Caribbean Poverty Gender and Equity Group, Washington, DC. 70 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN CHAPTER 4 Unraveling the Complexities of Chronic Poverty P eople who escape poverty seem to be better integrated and share more of the characteristics of the nonpoor than people who remain poor. Identifying some of the potential correlates of chronic poverty is helpful in understanding why some people stay poor. We first focus on the endowments and the context, in order to determine what accounts for the concentration and persistence of poverty in specific geo- graphical areas within a country. Why did some areas remain poor, even after a period of sustained economic growth across Latin America and the Caribbean (LAC)? Under one hypothesis, spatial pockets of high poverty persistence reflect concentrations of people with unfavorable endowments (low levels of skills and assets) (Ravallion 1998). Under another, local conditions affect returns to endowments, so that a household could be poor in one region and not poor in another (because returns to education vary across locations, for example). The uneven distribution of geographic and local factors (for example, climate, public services, access to markets) may account for geographical differences in chronic poverty. Understanding such differences is critical, because endowments and contexts imply different types of policies to address chronic poverty. Under the first hypoth- eses, the starting point should be improving individuals’ endowments (improving the quality of education). Under the second, a territorial approach may be more effective (improving institutions or infrastructure to promote trade, and therefore employment opportunities, or facilitating migration). The first part of this chapter reviews how the local context affects returns to endowments and identifies some of the local factors that affect returns. It focuses on factors that are relevant for social policy: institutions and uninsured risk. 71 The second part of the chapter reviews the role of the state of mind and its rela- tionship with the first two inputs and the process of escaping poverty. It shows that aspirations and stress can be greater barriers to upward mobility than endowments or the context. Policies that do not account for the state of mind could therefore be less effective. Endowments, the Context, and Chronic Poverty Where a child is born should not affect his or her future. But differences in initial endowments and access to services have enormous effects on the probability of being or staying poor. Endowments Table 4.1 presents descriptive characteristics of the chronically poor in 2004. To put them in perspective, we compare them with two other relevant groups: people who escaped poverty during the following decade and people who remained nonpoor throughout the 2004–12 period. Many initial similarities are apparent between the chronically poor and people who escaped poverty, and many difference are apparent between both groups and the nonpoor. Relative to the nonpoor, the chronically poor had larger households in 2004 and higher dependency ratios of younger children (more than two), and they lived in households in which the head was younger and had significantly fewer years of education (six versus nine). Ethnicity Ethnicity is associated with chronic poverty (box 4.1; see World Bank 2014a for an extensive review of issues pertaining to indigenous people in Latin America). Differences in cognitive skills Schady and others (2014) find large socioeconomic gradients of cognitive devel- opment skills in early childhood. They examine test scores of children age three to seven from five Latin American countries on the Test de Vocabulario en Imagenes Peabody (TVIP), a measure of receptive language known to be strongly corre- lated with success in adulthood. The gaps between children from the richest and poorest wealth quartiles are large (figure 4.1). In Nicaragua and Peru, the average child in the poorest quartile in rural areas has a TVIP score that is more than two standard deviations below the international reference population used to norm the test.1 Children in urban areas have somewhat higher scores than children in rural areas, but even in urban areas the gap between the top and bottom quartiles are large (6 points in Chile and 26 points [more than 1.5 standard deviations] in Colombia). These gaps exist across countries. A child in the richest quartile 72 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN TABLE 4.1 Characteristics of people in Latin America and the Caribbean who were chronically poor, who escaped poverty, and who were never poor, 2004 Characteristic Chronically poor Escaped poverty Never poor Demographics Household size 4.8 4.9 3.6 Age 0–12 2.0 1.9 0.9 0–15 2.4 2.3 1.1 13–18 0.7 0.8 0.5 19–70 2.1 2.2 2.2 70+ 0.1 0.1 0.1 Age of household head 37.8 40.0 40.8 Education Years of education of household head 6.0 5.1 9.2 Access to services Drinking water 0.8 0.9 1.0 Shelter 0.7 0.8 0.9 School attendance 0.9 1.0 1.0 At least five years of schooling 0.9 0.9 1.0 Sanitation 0.5 0.7 0.9 Electricity 0.9 1.0 1.0 Assets 0.6 0.8 0.9 Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Estimates are population-weighted averages of country-specific estimates, which are based on 2004 or 2012 surveys (or nearest year in cases in which 2004 or 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. Chronically poor = poor in both 2004 and 2012. Escaped from poverty = poor in 2004 but not 2012. Never poor = not poor in either year. residing in rural Nicaragua scores below the child in the poorest quartile in all other countries but Peru. Differences in school outcomes In all countries, children from chronically poor households are severely disad- vantaged in schooling outcomes (figure 4.2). In Brazil, for example, a quarter of children from chronically poor households finished primary school on time in 2012, compared with half of all children from households that were never poor. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 73 BOX 4.1 Ethnicity and chronic poverty in rural Guatemala Indigenous people in Guatemala have historically faced economic and social exclusion. In 2011 three out of four Guatemalans living in chronically poor rural areas (municipalities in which poverty rates were above 75 percent in 2000 and 2011) were indigenous (figure B4.1.1, panel a). By contrast, in rural municipalities that significantly reduced their poverty rates between 2000 and 2011, half of the population was indigenous. Adults living in chronically poor municipalities with higher concentrations of indigenous populations have lower education levels, and their children have lower school attendance and higher malnutrition rates. Within indigenous communities location is a key factor determining the probability of living in a chronically poor area: Indigenous municipalities at higher altitudes and farther from Guatemala City are significantly more likely to be chronically poor. Some progress has occurred. The share of people living in municipalities that significantly reduced their poverty rates between 2000 and 2011 was similar in indigenous and nonindigenous communities (figure B4.1.1, panel b). The indigenous communities that improved had lower initial poverty rates and poverty gaps than the chronically poor indigenous communities, suggesting that endowments and context indeed matter in the dynamics of poverty persistence. FIGURE B4.1.1 Chronic poverty and ethnicity in rural Guatemala, 2011 a. Share of indigenous population in chronically b. Rate of chronic poverty and improvement poor and improved municipalities in majority indigenous and majority nonindigenous municipalities 100 60 Percent of population Percent of population 80 40 60 40 20 20 0 0 Chronically poor Improved Majority Majority indigenous nonindigenous Indigenous Nonindigenous Chronically poor Improved Sources: Poverty maps for 2000 and 2011, based on Encuesta Nacional de Condiciones de Vida (ENCOVI) and population census. Note: Chronically poor municipalities = municipalities in which poverty rate was above 75 percent in 2000 and 2011. Improved municipalities = municipalities in which poverty rate was above 75 percent in 2000 and below 75 percent in 2011. 74 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 4.1 Mean standardized test scores of richest and poorest children in urban and rural areas of five countries in Latin America 120 Mean score on the Test de Vocabulario 90 en Imagenes Peabody (TVIP) 60 30 0 Urban Rural Urban Rural Urban Rural Rural Urban Rural Chile Colombia Ecuador Nicaragua Peru Richest quintile Poorest quintile Source: Schady and others 2014. Note: Scores are externally standardized to allow comparisons with the international norm. A score of 100 indicates that a child is performing at the normal age-appropriate level. Data are from various years in the 2000s. The differences between chronically poor children and children from households that escaped poverty during this period are also large. In 2012, 34 percent of Peruvian children from chronically poor households finished middle school on time. The figure was almost 50 percent among children from households that escaped poverty and nearly 70 percent among children from households that were never poor. These gaps indicate a strong transmission of chronic poverty across generations. The Disabling Context The chronically poor had less access to a range of basic services and fewer assets than both the nonpoor and people who escaped poverty by 2012 (see table 4.1).2 In 2004 only 79 percent of people who were still poor in 2012 had access to water, compared with 95 percent of the nonpoor and 89 percent of people who escaped poverty. Only 58 percent of the chronically poor had a minimum level of assets, compared with 90 percent of the nonpoor and 78 percent of people who escaped poverty. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 75 FIGURE 4.2 Intergenerational transmission of chronic poverty in selected countries in Latin America and the Caribbean, 2012 a. Finished 6th grade on time 100 School completion rate (percent) 80 60 40 20 0 Ec o a Co ama Ho bia ile Pa ia Ur a Sa s Gu agua la tin Pa ublic ru Ca eric ay y Bo r Ni dor Re il be nd do El ura xic Ric tin ua an raz liv ma Pe Ch the m gu lom Co an rib a a ua lva en ug Me n nd nic B p ate sta r a ca g r Ar A mi La Do Chronically poor Never poor b. Finished 10th grade on time 100 School completion rate (percent) 80 60 40 20 0 na Me e Ec o Co ma a Ho bia Gu gua la Pa ia c Sa s ru ay Ca eric ay Ni dor Bo r Re il be nd El ura tin Pa ubli il xic Ric do ma an raz liv Pe Ch nti u the Am ragu lom na Co an rib a a lva ua ra ug nd nic B p ate ge sta ca Ur Ar mi La Do Chronically poor Never poor Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Figures come from 2012 surveys (or nearest year in cases in which 2012 data were not available). Access to services Regions with higher rates of chronic poverty offer limited services, including access to clean water, sewerage systems, and sanitation facilities (figure 4.3). The chronically poor are also more likely to reside in regions with lower coverage of electricity and mobile telephony, although the gaps for these services narrowed 76 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 4.3 Correlation between chronic poverty and access to services at subnational level in Latin America and the Caribbean, 2012 a. Access to clean water b. Access to sanitantion 1.0 1.0 Regional access to Regional access to 0.8 0.8 clean water sanitantion 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Chronically poor (%) Chronically poor (%) c. Access to sewer d. Access to mobile phone 1.0 Regional access to 1.0 Regional access to 0.8 0.8 mobile phone 0.6 0.6 sewer 0.4 0.4 0.2 0.2 0 0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Chronically poor (%) Chronically poor (%) e. Access to electricity f. Overall access to services Regional lacks of at least 1.0 1.0 Regional access to 0.8 0.8 3 services (%) electricity 0.6 0.6 0.4 0.4 0.2 0.2 0 0 0 0.2 0.4 0.6 0.8 0 0.2 0.4 0.6 0.8 Chronically poor (%) Chronically poor (%) 95% Confidence interval Fitted values Cronically poor by region (%) Source: Calculations based on SEDLAC data (CEDLAS and World Bank) Note: Figures come from 2012 surveys (or nearest year in cases in which 2012 data were not available). Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. Selection of regions varies depending on level of representativeness of surveys. over the last decade (box 4.2). Services appear to complement one another as drivers out of chronic poverty: As panel f in figure 4.3 shows, households with no access to three or more basic services are more likely to reside in regions with higher levels of chronic poverty. Chronically poor districts in rural Guatemala were significantly less likely to have access to a range of services than districts with low rates of poverty or rates of poverty that had fallen (panels a–c in figure 4.4). They were also significantly more UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 77 BOX 4.2 Connectivity pays off: Reducing poverty in rural Peru Beuermann, McKelvey, and Vakis (2012) estimate the effects of mobile phone expansion on poverty in rural Peru over the past 15 years. They exploit the timing of the arrival of mobile coverage at the village level, which allows them to test causally whether poverty in villages that received mobile coverage early on is lower. Their findings are striking: Mobile phone expansion increased real household consumption by 11 percentage points and decreased poverty by more than 5 percentage points (figure B4.2.1). Moreover, these benefits increased over time: Receiving mobile coverage nine years earlier was associated with a 15 percentage point difference in poverty at the village level, and the benefits appear to have been shared by all households in the villages, regardless of mobile phone ownership, suggesting strong spillover effects. The authors hypothesize a number of channels through which mobile connectivity reduced poverty, including reducing information asymmetries and search costs, potentially leading to higher productivity in agriculture (see also Marchionni and Glüzmann 2012). Connectivity pays off. FIGURE B4.2.1 Effect of mobile phone coverage on poverty in rural Peru 0.2 Poverty rate in villages with mobile signal minus poverty rate in villages without mobile signal 0.1 0 –0.1 –0.2 –5 –4 –3 –2 –1 0 1 2 3 5 4 6 7 8 9 Years since cellphone coverage began Note: Dashed lines show 95 percent confidence intervals. isolated (as measured by road access). The fact that these differences strongly cor- relate with poverty persistence reinforces how geographic isolation limits oppor- tunities and access to markets, fostering chronic poverty. Remoteness Differences in access to services and connectivity are also evident at the regional level. For example, both chronically poor and still poor but improved districts in 78 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 4.4 Access to services in rural Guatemala, 2011 a. Access to water b. Access to sanitation 100 100 Percent of households with Percent of households with 80 80 access to service access to service 60 60 40 40 20 20 0 0 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 Poverty rate (percent of population) Poverty rate (percent of population) c. Access to electricity d. Road density index 1.0 100 Percent of households with 0.8 Road density index 80 access to service 0.6 60 0.4 40 20 0.2 0 0 0 0.2 0.4 0.6 0.8 1.0 0 0.2 0.4 0.6 0.8 1.0 Poverty rate (percent of population) Poverty rate (percent of population) Chronically poor Below mean Improved Linear fit Sources: Poverty maps for 2000 and 2011, based on Encuesta Nacional de Condiciones de Vida (ENCOVI) and population census. Note: Chronically poor (tan circles) = areas in which poverty rate was above 75 percent in 2000 and 2011. Improved (blue circles) = areas in which poverty rate was above 75 percent in 2000 and below 75 percent in 2011. Below mean (+) = area in which poverty rate was below 75 percent in 2000 and 2011. Size of symbols reflects size the population. Peru had similar initial poverty levels. But between 2007 and 2012, average pov- erty fell by 9 percentage point in chronically poor districts (to 64 percent) and by 33 percentage points (to 35 percent) in districts that improved (table 4.2). Location seems to explain much of the difference. The majority of chronically poor districts in Peru are located in the highlands (table 4.2). They tend to be in far, remote areas: More than 15 percent are located more than five hours from the provincial capital, compared with only 1 percent of districts with poverty rates below 50 percent. Size of district The chronically poor are also more likely than other groups to live in mid-size districts: More than two-thirds reside in districts with populations of 4,000–25,000, UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 79 TABLE 4.2 Population size and remoteness between districts that were never poor, escaped poverty, and were chronically poor in Peru Poverty level Variable Never poor Escaped poverty Chronically poor Poverty headcount (percent) 2007 24 68 73 2012 17 35 64 Distribution of population by size of village, town, or city (percent in 2007) Less than 1,000 0 1 1 1,000–1,999 1 4 3 2,000–3,999 1 10 11 4,000–9,999 4 24 33 10,000–25,000 9 28 34 More than 25,000 85 32 18 Average altitude (meters) 641 2,372 2,293 Percent of population living more than five 1 9 15 hours from provincial capital (2009) All districts in Peru 641 439 740 Sources: Poverty maps for 2007 and 2012, census data for 2007, and El Registro Nacional de Municipalidades (RENAMU). Note: Below 50 percent = districts with poverty rates below 50 percent in 2007 and 2012. Improved = districts with poverty rates above 50 percent in 2007 but below 50 percent in 2012. Chronically poor = districts with poverty rates above 50 percent in both years. with the other third split equally between small and large districts (see table 4.2). In contrast, districts in which the poverty rate was less than 50 percent in 2012 tended to have more than 25,000 inhabitants. Institutions The type and quality of national and local institutions that people have access to have a strong impact on welfare in general and on chronic poverty in particular. It is easier for the poor to escape poverty if their voices are heard and taken into consideration. The poor need to be represented by leaders who understand their needs and the challenges they face. Local governments, services, and social pro- grams must be staffed by qualified civil servants who seek to make the life of the poor easier rather than create additional obstacles. The starting point must be an inclusive social contract that recognizes every citizen as equal and supports efforts to provide equal opportunities for all. It is impossible to eradicate poverty without establishing such a basis. Robinson 80 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN (2014) looks at the evolution of the social contract in Colombia, both over time and across regions, documenting the channels through which poor governance and elite capture of the political process can perpetuate poverty (box 4.3). But a good social contract is not enough. To function effectively and support citi- zens, institutions must work well at three levels. First, there must be sufficient bureau- cratic capacity to deliver good services. Second, the state must be able to enforce the legal framework and administer the justice system. Third, institutional processes must be transparent and follow clear accountability rules (Fukuyama 2011). The importance of capacity has been regularly underestimated. Social pro- grams often have little impact on the chronically poor because they live in regions where the capacity to implement such programs is weak. Loayza, Rigolini, and Calvo-González (2014) study the ability of Peruvian districts to spend additional budget streams from mining revenues. They find that although poorer districts BOX 4.3 Improving institutions, reducing poverty in Colombia Robinson (2014) argues that poverty-reduction efforts in Colombia have been hindered by weak democratic process and that the recent drop in poverty was to some extent associated with improvements in those processes. The remarkable decline in the poverty rate, which fell from 47.4 percent in 2004 to 32.7 percent in 2012, contrasts with the stagnant state of poverty reduction during the last decades of the 20th century, when poverty fluctuated around slightly more than 50 percent. Deficiencies in the democratic process in some regions that prevented the poor from exercising political power may have prevented them from getting the state to provide them with basic services. It also affected their ability to gain proper titles to their land and secure property rights; made their human rights highly insecure and continually violated; and kept them from accessing the same opportunities as the nonpoor. As a consequence, the poor had few assets and low returns on them. Robinson argues that this powerlessness was a consequence of two features that are the crux of what Acemoglu and Robinson (2012) call “extractive political institutions”: the lack of pluralism and the lack of political centralization. The lack of pluralism means that political power was narrowly concentrated. The lack of political centralization means that the state lacked capacity and was not as effective as it could have been. Although Colombia has a vibrant electoral tradition, violence, vote buying, and clientelism may have prevented the poor from exerting political power in some regions. Robinson shows that at times a few thousand votes were sufficient to win a seat in the Senate and that widespread clientelism made policy makers more accountable to the people to whom they owed favors than to their constituencies. Gradual changes and a sequence of shocks obliged elites to enact reforms that eventually led to remarkable poverty-reduction efforts. There were three drivers. First, the government gradually acquired the monopoly on the legitimate use of physical force, a feature, according to Max Weber, that lies at the core of the definition of a state (Weber 1946). As a result, the civilian state spread into areas that had previously been off-limits. Second, the constitutional reform of 1991 facilitated the implementation of health reform, the expansion of basic education, and an increase in fiscal revenues that helped empower the poor. Third, clientelistic networks gradually lost some of their powers. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 81 spend larger shares of the additional budget allocated to them than wealthier dis- tricts (a sign of greater needs), districts with lower average educational attainments have less ability to spend additional funds. Galasso (2011) studies the impact of Chile Solidario, a social intermediation service that targets the extremely poor. She finds that the impacts of the program are significantly greater when beneficiaries are assisted by social workers with good per- formance assessments (potentially proxying for institutional quality and capacity). There is widespread recognition that the quality of governance affects the per- formance of social services and programs. Some of the channels through which governance works are not clear, however. Empirical research shows that bad gov- ernance weakens service delivery by reducing the ability and incentives of policy makers and beneficiaries to monitor providers (World Bank 2004). In its extreme form, bad governance (as captured by high levels of corruption) also raises the cost of services. Using Peruvian data, Hunt and Laszlo (2005) find that bribing does not necessarily help “smooth” processes; rather, it helps beneficiaries obtain from a corrupt official the same service an honest official would have provided for free. They find that the rich pay more bribes than the poor (because they tend to have greater contact with government officials). Using the same data, Hunt (2007) finds that bribes increase with misfortune—a strong driver of chronic poverty—because it increases the demand for public services. Bad governance in the delivery of services disproportionately affects the poorest, because they have no escape route. Kaufmann, Montoriol-Garriga, and Recanatini (2008) construct measures of governance using data from users of pub- lic services from 13 government agencies in Peru. For some basic services, they find that low-income users pay a larger share of their income than wealthier ones do (that is, the bribery tax is regressive). Where there are substitute private pro- viders, low-income users also appear to be discouraged more often and to stop seeking basic services. Bribery may thus penalize poorer users twice—acting as a regressive tax and discouraging use of basic services. Uninsured risk The presence of risk affects households’ welfare through costly risk-management strategies (Fafchamps 2003; Elbers, Gunning, and Kinsey 2007). In the absence of complete insurance markets, households typically skew their income portfolio toward low-risk/low-return assets or activities (Deaton 1992; Rosenzweig and Binswanger 1993; Morduch 1995; Dercon 1998). Shocks—the realization of risk—have been singled out as a key determinant of welfare dynamics. They can have heterogeneous dynamic impacts on welfare, affecting households’ mobility in the short and long term. The chronically poor in LAC face disproportionate levels of uninsured risk. Figure 4.5 shows a positive correlation between countries with higher chronic poverty and the prevalence of natural disasters. In the absence of some form of insurance and savings, shocks—especially repeated ones—can have long-term 82 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 4.5 Correlation between chronic poverty and incidence of weather shocks in Latin America and the Caribbean, 2012 45 Guatemala 40 35 Honduras Percent of population living Nicaragua 30 in chronic poverty El Salvador 25 Colombia Ecuador Paraguay 20 Peru Dominican Republic 15 Mexico Brazil Bolivia Panama 10 Argentina Costa Rica 5 Chile Uruguay 0 0 0.5 1.0 1.5 2.0 Percent of population affected by droughts, floods, and extreme temperatures Sources: SEDLAC data (CEDLAS and World Bank) for chronic poverty; Guha-Sapir, Below, and Hoyois n.d. for data on disasters. implications and exacerbate chronic poverty. Shocks can generate declines in welfare by directly depleting assets (Carter and others 2004; McPeak 2004) or by triggering coping mechanisms (such as selling productive assets or curtailing invest- ments in human capital) that reduce welfare in the long run (Rosenzweig and Wolpin 1993; Zimmerman and Carter 2003) or human capital (Jacoby and Skoufias 1997; Dercon and Hoddinott 2005). In addition, shocks can slow recovery and lead to poverty traps (Carter and Barrett 2006). The effect of shocks on welfare dynamics can extend beyond the short term (Dercon 2004; Lokshin and Ravallion 2004). The relationship between shocks and chronic poverty can also be observed within countries. Data from Guatemala suggest that overall exposure to weather risk and poverty are not correlated (panel a of figure 4.6). In contrast, the potential impact of weather risk varies greatly. The risk of food shortages is close to 100 per- cent for the majority of chronically poor municipalities, whereas it is close to zero for rural municipalities with low poverty rates (panel b of figure 4.6). For munici- palities in which poverty declined over the decade, vulnerability to food shortages is distributed evenly across municipalities. Premand and Vakis (2010) use three rounds of panel data in Nicaragua to show that weather shocks have large impacts on poverty persistence. Using matching techniques, they find that being affected by a weather shock increases UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 83 FIGURE 4.6 Location of and food shortages from weather shocks in rural Guatemala, 2011 a. Correlation between weather shocks and geographic location 1.0 0.8 Weather risk index 0.6 0.4 0.2 0 0 20 40 60 80 100 Poverty rate (percent) b. Correlation between municipalities’ level of chronic poverty and vulernability to food shortages from weather shocks 1.0 Index of vulnerabilty to weather events 0.8 0.6 0.4 0.2 0 0 20 40 60 80 100 Poverty rate (percent) Improved Chronically poor Below mean Sources: Poverty maps for 2000 and 2011, based on Encuesta Nacional de Condiciones de Vida (ENCOVI), population census, and Ministry of Agriculture of Guatemala. Note: The weather risk and the vulnerability risk indexes were created by the Ministry of Agriculture of Guatemala. The weather risk index ranks municipalities based on the probability of experiencing droughts, floods, or frosts based on historical records. The vulnerability risk index ranks municipalities based on the probability of experiencing droughts, floods, or frosts adjusted by the capacity of the municipality to respond and the risk of food shortages. Chronically poor (tan circles) = area in which poverty rate was above 75 percent in 2000 and 2011. Improved (blue circles) = area in which poverty rate was above 75 percent in 2000 and below 75 percent in 2011. Below mean (+) = area in which poverty rate was below 75 percent in 2000 and 2011. Size of symbols reflects size of population. 84 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN the probability of remaining in chronic poverty by up to 10 percent and that the effects persist for as long as seven years. Barriga, Vakis, and Rigolini (2014) study the effect of repeated shocks (natural disasters) in Peru on household welfare using a four-year panel. They find that early shocks can permanently shift income trajectories, increasing the probability of poverty persistence by 13 percent and creating a long-term increase in the poverty gap. Shocks that push households deeper into poverty make it harder to recover. The fact that (uninsured) risk may directly cause poverty persistence implies that specific policies should be designed to address it (Gardiner and Hills 1999; de Janvry and others 2006). Although insurance and income-stabilization schemes that protect households against idiosyncratic economic shocks may be easier to implement when poverty is transient, there could be large potential gains from risk management policies that target the chronically poor. Risks are not only related to weather, health, and macroeconomic shocks. Other factors, such as crime, are drivers of risk in LAC. Box 4.4 shows how a targeted approach to crime can reduce chronic poverty. BOX 4.4 Does increased public expenditures reduce crime and chronic poverty? Evidence from Mexico A recent study looks at the linkages between chronic poverty and crime in Mexico. Martínez-Cruz and Rodríguez-Castelán (2014) compile municipality-level panel data on income, poverty, and crime for 1990, 2000, 2005, and 2010. They define a municipality as chronically poor if the percentage of people in food consumption poverty remains above the national average during two consecutive periods. They measure crime in terms of the number of homicides per 100,000 inhabitants. They find that crime contributed to chronic poverty. Everything else being equal, an urban municipality with 225 homicides per 100,000 people was roughly 2.5 times more likely to remain chronically poor than a rural municipality with no crime. The authors also explore the role of municipal-level public expenditures and their link with crime. Average public spending per capita in 2010 was 30 percent lower in chronically poor municipalities than in nonchronically poor municipalities. The effect on chronic poverty of increasing public spending was greater than the effect of reducing crime: On average $Mex0.16.5 of per capita public spending were needed to compensate for the impact of one homicide per 100,000 inhabitants in terms of perpetuating poverty. In municipalities with average public spending of $Mex3,000 per capita (about the third quartile of the spending distribution) and homicide rates above the national average, an additional peso of investment per capita was associated with a reduction in the probability of being chronically poor of 0.27 percent, compared with 0.17 for a municipality with the same level of spending but a below-average homicide rate. These results showcase the significant cost of crime on welfare and poverty persistence. They suggest that well-targeted public spending in chronically poor municipalities, especially municipalities with higher crime rates, could simultaneously help address the persistence of both poverty and crime. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 85 Linkages between Endowments and Context Both endowments and context affect chronic poverty, but to what extent do they do so? To understand differences in the income-generation processes, we carry out some basic Oaxaca-Blinder decompositions, closely following Blinder (1973). The goal is to estimate the proportions of the difference in income in 2012 between people who were poor in 2004 and stayed poor and people who were poor and escaped poverty that can be attributed to differences in endowments (for example, education and household composition); to returns to those endowments; and to other factors (the unexplained portion of the differential). The results are remarkable. Endowments matter, but in most settings the returns to endowments, which depend on the context in which endowments can be used, matter more: Context explains a significant part of the variation in income differences between the chronically poor and people who escape poverty in every country except Colombia. In El Salvador, for example, out of the 10 percentage point difference in incomes between the chronically poor and people who escaped poverty, almost 8 percentage points can be attributed to returns to endowments; endowments themselves explain a little more than 2 percentage points (figure 4.7). FIGURE 4.7 Role of endowments in explaining income differences between the chronically poor and people who escaped poverty in Latin America and the Caribbean, 2012 30 25 Percent of income differential explained 20 15 10 5 0 –5 –10 –15 Ni ivia ua Sa zil r o r as ay bia ma na Re le Co blic Pa ica y ru do do ua xic i a Pe Ch u ur rag nti R lom na Br lva ua l pu ug rag Bo Me nd ge sta Pa Ec ca Ur Co Ho Ar El an nic mi Do Endowments Returns Unexplained component Source: Calculations based on SEDLAC data (CEDLAS and World Bank). Note: Endowments include household composition, assets, and education. Figures are based on 2012 surveys (or nearest year in cases in which 2012 data were not available). 86 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN These differences vary across the region: In Uruguay, returns explain 9 percent more of the income differential, while in Brazil they explain three times as much. Geographic factors and the local context—access to infrastructure, markets, and services—therefore influence income-generation capabilities in a big way.3 These insights have important implications for policy design. Supporting individuals is necessary, but it may not be sufficient if the enabling context does not provide them with the scope to take advantage of opportunities. Without a good context, chronic poverty may prevail because people are unable to use their endowments effectively. Social policies and regional development therefore need to go hand in hand. Some caveats on interpreting these results are warranted. First, the exer- cise does not reveal causation. Second, the decomposition ignores the role of intensity in the use of endowments. For example, lack of a land title could prevent property owners from accessing credit (because of lack of collateral), constraining their ability to fully exploit their capacity (Bussolo and López- Calva 2014). Arguably, knowing whether the constraints lie in improving endowments or improving their use could lead to different policy implica- tions (investing in skills or increasing access to credit). Third, comparing the appropriateness of territorial (agglomeration) policies with policies that directly support people demands that one consider the relative effectiveness of specific policies as well factors that go beyond the economic dimension. In this sense, rather than an either-or approach, our findings strongly point toward the complementarities of both approaches and the need to coordinate local development and social assistance. Our findings remain silent on the appropriateness of policies favoring local development versus policies facilitating migration toward regions that offer higher returns to endowments. This book sheds little light on this complex question. How much weight policies should put on either approach depends on more than pure economic analysis. Cultural, socioeconomic, and political factors need to be considered and case-specific decisions made. State of Mind and the Process of Emerging from Poverty Behavior under Stress Policy makers rarely consider state of mind and behaviors in designing policies and programs. Many social programs address material obstacles standing in the way of a better future (lack of capital, assets, skills), but very few explicitly target emotional obstacles, such as lack of self-esteem, stress, or the ability to aspire. Most programs forget that chronic poverty can have psychological and even neurobio- logical consequences, which in turn affect people’s state of mind, behaviors, and the process of making economic decisions. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 87 Behavioral economists have long argued that people deviate from the standard economic model, which ignores the process of decision making (Kahneman 2011). Because the ability to analyze information is limited, people engage in “bounded rationality” (Simon 1955). They use quick rules of thumb to make decisions, not complex analytical models that maximize welfare. They often select default options rather than choose among many alternatives they perceive as complicated. Willpower is also bounded. People do not always make choices that are in their best interest in the long run, valuing small instant gratification instead of larger delayed rewards. Selfishness is also sometimes bounded: People act altruistically (think of the decision to be an organ donor). Finally, everyone has a limited ability to process information. The need to address immediate concerns may constrain the ability to strategically plan for the future (Mullainathan and Shafir 2013). As Kahneman (2011) argues, a common element across these insights is that they are not driven by emotional patterns but are instead linked to cognitive capacities. People’s own beliefs (intuition), the beliefs of others (cultural norms), and the way people interact with one another (social norms) also shape the process of decision making (World Bank 2014b).4 A key insight that emerges from this literature is that (chronic) poverty is both an outcome and a condition. The condition of poverty puts pressure on people, influencing their state of mind and decision-making process in a profound way (box 4.5). On a daily basis, people in poverty need to deal with many issues (find- ing clean water, paying bills in person, walking to work) that richer people do not. They are likely to put more weight on solving these daily problems than on making strategic decisions about their future. The day-to-day planning that poor people have to engage in requires precious attention and time that they often do not have. Poor people are also less likely than the nonpoor to draw or infer from the positive experience of their peers and more likely to be affected by negative or suboptimal experiences of the surrounding context, which may determine how they perceive and take advantage of opportunities. One’s state of mind can thus be a strong element in the perpetuation of poverty (Haushofer and Fehr 2014). BOX 4.5 Does poverty cause stress? Haushofer and Shapiro (2013) randomly selected households in Kenya to receive gifts of $0, $400, or $1,500. After a year, they found that people receiving $400 or $1,500 reported significant improvement in measures of happiness, life satisfaction, stress, and depression. People who received $1,500 also had significantly lower levels of cortisol (the stress hormone). Chemin, de Laat, and Haushofer (2013) studied farmers in Kenya who experienced negative income shocks as a result of drought. They observed increases in both self-reported stress and cortisol. 88 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Chronic Poverty, Aspirations, and Forward-Looking Attitudes Poverty can smother dreams, forward-looking behaviors, and aspirations. Appadurai (2004) and Ray (2006) suggest that the lack of capacity to aspire— one manifestation of lack of agency—may hinder upward mobility among the poor by preventing them from investing in human capital and production technology. The role of aspirations in decision making has also been incorporated into the theoretical microeconomic literature (Börgers and Sarin 2000; Diecidue and Van de Ven 2008), and aspirations-based learning has been introduced into game theory (Bendor, Mookherjee, and Ray 2001). Lack of aspiration can affect economic outcomes at different stages in the decision-making process. It can affect how people perceive opportunities and the returns to opportunities (affecting choice), whether they choose to take them up (influencing actions), and how much effort they may put into actions (affect- ing upward mobility). These behavioral channels can themselves contribute to poverty persistence by limiting attention and favoring habitual behaviors at the expense of goal-directed ones, despite the potential presence of the best “econom- ically designed” public or private interventions. To study aspirations, we complement the chronic poverty estimates with infor- mation from perception surveys conducted by the Latin American Public Opinion Project (LAPOP) on expectations of future well-being. A number of interesting insights emerge. At the country level, expectations about the future are lower in countries in which the incidence of chronic poverty is higher (figure 4.8, panel a). Expectations about the future are very low in the countries with the highest chronic poverty rates (Guatemala, Nicaragua, and Honduras). People living in provinces in which the incidence of chronic poverty is higher also tend to be more pessimistic about their economic prospects (figure 4.8, panel b). Differences in perceptions about the future are also striking across socioeco- nomic groups. The chronically poor are the most pessimistic about their future, with one out of every five people expecting his or her economic situation to worsen in the next year—twice as many as among people who exited poverty or who were never poor (figure 4.9). The chronically poor are half as likely as people who exited poverty to expect their situation to improve in the future. People who exited poverty are the most optimistic, consistent with the causal evidence on how positive events can affect people’s outlook in life. Our results are also consistent with data from the World Values Surveys that suggest that both within and across countries, people with lower incomes are more likely to report that life is meaning- less, to agree that it is better to live day to day because of the uncertainty of the future, and to reject adventure and risk (Haushofer and Fehr 2014). To be sure, cultural factors also matter, and forward-looking attitudes of the chronically poor are heterogeneous across countries. The chronically poor in UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 89 FIGURE 4.8 Correlation between chronic poverty and expectations in Latin America and the Caribbean, 2010 a. Country level Brazil Uruguay Paraguay 2.6 Average level of expected well-being Colombia Dominican Republic 2.4 Costa Rica Panama Peru Chile Bolivia 2.2 Mexico El Salvador Ecuador Nicaragua Honduras Argentina Guatemala 2.0 0 10 20 30 40 50 60 70 Percent of population living in chronic poverty b. Provincial level 3 2.5 Average expected well-being 2 1.5 1 15 25 35 Percent of population living in chronic poverty Source: Chronic poverty data come from SEDLAC (CEDLAS and World Bank). Note: Expected well-being data are based on responses (on a scale of 1–5) to the question, “Do you think that in the next 12 months your economic situation will be better/same/worse?” in 2010 Latin American Public Opinion Project survey. 90 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 4.9 Level of optimism about next 12 months by the chronically poor, people who escaped poverty, and people who were never poor in Latin America and the Caribbean, 2010 100 80 Percent of respondents 60 40 20 0 Chronically poor Never poor Escaped poverty Better The same Worse Source: Latin American Public Opinion Project 2010. Brazil, Uruguay, and Bolivia are the most optimistic, while the chronically poor in Argentina, Guatemala, or Mexico are among the most pessimistic (figure 4.10). In addition to culture, dynamic poverty patterns could also explain this variation across countries. Brazil, Uruguay, and Bolivia experienced high rates of poverty reduction over the decade; the positive experiences of other poor people exiting poverty may have helped people still living in poverty retain hope (or aspire more). In contrast, Guatemala, Mexico, and Nicaragua experienced small improvements or even increases in poverty. Social norms and peer effects also affect state of mind, behaviors, and aspira- tions. Psychologists have long emphasized that social comparisons can influ- ence both attitudes and actions (Festinger 1954; Karlsson and others 2004) and that leaders who communicate an inspiring vision and behave supportively can enhance the performance of others (Latham and Saari 1979; Bass 1985). In economics some theoretical microeconomic models (Dixon 2000; Bendor, Mookherjee, and Ray 2001) and overlapping generation macroeconomic mod- els (Genicot and Ray 2014; Mookherjee, Ray, and Napel 2010) look at aspiration levels and investments as depending on the experiences of others. Paraphrasing Appadurai (2004) and Ray (2006), the capacity to aspire is a specific instance of culture that is future oriented. Aspirations are socially determined. The capac- ity to aspire is thus inherently unequal between the rich and the poor, who lack the aspirational resources (or capacity) to act and change the conditions of their own poverty. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 91 FIGURE 4.10 Expectations about next 12 months by the chronically poor in selected countries in Latin America and the Caribbean, 2010 100 80 Percent of respondents 60 40 20 0 Ur zil y Pa livia y Co ile C o i ca Pa ia ma Ho ru Ec as Sa or mi Nic or Re a Gu tina o la Ar lic ua ua an gu xic b ma d d a Pe b Ch ur R lom na ua Br lva pu ug rag n nic ara Bo Me nd ge ate sta El Do Better The same Worse Source: Latin American Public Opinion Project 2010. The poor may fail to invest in the future because the people close to them may suggest that escaping poverty is not feasible. Breaking such a transmission chan- nel is essential. Like the overall context, the behavioral context matters. Learning about the positive experiences of others can be beneficial to one’s own aspirations. Aspirations may affect how other people perceive opportunities. For example, parents’ aspirations can affect the aspirations of their children, potentially creat- ing a vicious cycle that may affect the intergenerational transmission of poverty. Using data from Peru, we find a positive and robust association between children’s educational aspirations and the aspirations their parents have for them (although children tend to have slightly higher aspirations than their parents) (figure 4.11). Evidence of such channels can be extrapolated by exploring the association of aspirations among peers. The Young Lives Project in Peru also collects panel data about children, whom it tracks every four to five years (the project started in 2001). Figure 4.12 plots children’s educational aspirations against those of their peers. 92 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 4.11 Educational aspirations of Peruvian children and their parents 20 Children´s educational aspirations 15 10 5 0 0 5 10 15 20 Parent’s educational aspirations for their children Children-parent aspiration Source: Data from Peru’s Young Lives Project. Note: Scores are based on responses to the question, “Which education level would you like to achieve?” (scale 1-5) FIGURE 4.12 Correlation between educational aspirations of children ages 12–15 and their peers in Latin America and the Caribbean, 2011 20 Years of education aspired to by peers 15 10 5 0 5 10 15 20 Years of education aspired to by children Source: Young Lives Surveys Peru. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 93 The relationship is strong: Children living in communities in which the educa- tional aspirations of their peers are high have high aspirations (and vice versa). This evidence suggests that social interactions can trigger positive changes in the decision-making process. Notes 1. Results should be treated with caution, as norms reflect international norms, which are not necessarily appropriate in these countries. 2. Assets include a landline or mobile phone; a bicycle, motorcycle, or car; and a refrigerator and television. 3. Using data for 1985–97, Escobal and Torero (2000) show that what seem to be sizable geographic differences in living standards in Peru can be almost fully explained by the spatial concentration of households with readily observable nongeographic characteristics, in particular public and private assets. 4. For a wide variety of examples in health, finance, education, and other domains, see Thaler and Sunstein (2008) and Mullainathan and Shafir (2013). References Acemoglu, Daron, and James Robinson. 2012. Why Nations Fail: The Origins of Power, Prosperity and Poverty. New York: Crown Business. Appadurai, Arjun. 2004. “The Capacity to Aspire: Culture and the Terms of Recognition.” In Culture and Public Action, edited by Vijayendra Rao and Michael Walton, 59–84. Stanford: Stanford University Press. Barriga, Oscar C., Renos Vakis, and Jamele Rigolini. 2014. “Poverty Persistence and Shocks in Peru.” World Bank, Washington, DC. Bass, Bernard M. 1985. Leadership and Performance beyond Expectations. New York: Free Press. Bendor, Jonathan, Dilip Mookherjee, and Debraj Ray. 2001. “Aspiration-Based Reinforcement Learning in Repeated Interaction Games: An Overview.” International Game Theory Review 3 (2–3): 159–74. Beuermann, Diether W., Christopher McKelvey, and Renos Vakis. 2012. “Mobile Phones and Economic Development in Rural Peru.” Journal of Development Studies 48 (11): 1617–28. Blinder, Alan S. 1973. “Wage Discrimination: Reduced Form and Structural Estimates.” Journal of Human Resources 8 (4): 436–55. Börgers, Tilman, and Rajiv Sarin. 2000. “Naive Reinforcement Learning with Endogenous Aspirations.” International Economic Review 41 (4): 921–50. Bussolo, Maurizio, and Luis F. López-Calva. 2014. Shared Prosperity: Paving the Way in Europe and Central Asia. Washington, DC: World Bank. Carter, Michael R., and Christopher B. Barrett. 2006. “The Economics of Poverty Traps and Persistent Poverty: An Asset-Based Approach.” Journal of Development Studies 42 (2): 178–99. 94 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Carter, Michael R., Peter D. Little, Tewodaj Mogues, and Workneh Negatu. 2004. “Shocks, Sensitivity and Resilience: Tracking the Economic Impacts of Environmental Disaster on Assets in Ethiopia and Honduras.” BASIS Research Program on Poverty, Inequality and Development, U.S. Agency for International Development, Washington, DC. Chemin, Matthieu, Joost de Laat, and Johannes Haushofer. 2013. “Negative Rainfall Shocks Increase Levels of the Stress Hormone Cortisol among Poor Farmers in Kenya.” SSRN 2294171. Deaton, Angus. 1992. Household Savings in LDCs: Credit Markets, Insurance and Welfare. Scandinavian Journal of Economics 94 (2): 253–73. De Janvry, Alain, Elisabeth Sadoulet, Pantelis Solomon, and Renos Vakis. 2006. “Uninsured Risk and Asset Protection: Can Conditional Cash Transfer Programs Serve as Safety Nets.” Social Protection Working Paper, World Bank, Washington, DC. Dercon, Stefan. 1998. “Wealth, Risk and Activity Choice: Cattle in Western Tanzania.” Journal of Development Economics 55 (1): 1–42. ———. 2004. “Growth and Shocks: Evidence from Rural Ethiopia.” Journal of Development Economics 74 (2): 309–29. Dercon, Stefan, and John Hoddinott. 2005. “Livelihoods, Growth, and Links to Market Towns in 15 Ethiopian Villages.” FCND Discussion Paper 194, Food Consumption and Nutrition Division, International Food Policy Research Institute, Washington, DC. Diecidue, Enrico, and Jeroen Van De Ven. 2008. “Aspiration Level, Probability of Success and Failure, and Expected Utility.” International Economic Review 49 (2): 683–700. Dixon, Huw D. 2000. “New Keynesian Macroeconomics: The Role of Theory and Evidence.” In Macroeconomics and the Real World, vol. 2. Keynesian Economics, Unemployment, and Policy, edited by Roger E. Backhouse and Andrea Salanti. Oxford, U.K.: Oxford University Press. Elbers, Chris, Jan Willem Gunning, and Bill Kinsey. 2007. “Growth and Risk: Methodology and Micro Evidence.” World Bank Economic Review 21 (1): 1–20. Escobal, Javier, and Máximo Torero. 2000. Does Geography Explain Differences in Economic Growth in Peru? Inter-American Development Bank, Washington, DC. Fafchamps, Marcel. 2003. Rural Poverty, Risk and Development. Cheltenham, United Kingdom: Edward Elgar Publishing. Festinger, Leon. 1954. “A Theory of Social Comparison Processes.” Human Relations 7 (2): 117–40. Fukuyama, Francis. 2011. The Origins of Political Order: From Prehuman Times to the French Revolution. New York: Farrar, Straus and Giroux. Galasso, Emanuela. 2011. “Alleviating Extreme Poverty in Chile: The Short Term Effects of Chile Solidario.” Estudios de Economía 38 (1): 101–27. Gardiner, Karen, and John Hills. 1999. “Policy Implications of New Data on Income Mobility.” Economic Journal 109 (453): 91–111. Genicot, Garance, and Debraj Ray. 2014. “Aspirations and Inequality.” NBER Working Paper 19976, National Bureau of Economic Research, Cambridge, MA. Guha-Sapir, D., R. Below, and Ph. Hoyois. n.d. EM-DAT: International Disaster Database, Université Catholique de Louvain, Brussels. www.emdat.be. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 95 Haushofer, Johannes, and Ernst Fehr. 2014. “On the Psychology of Poverty.” Science 344 (6186): 862–67. Haushofer, Johannes, and Jeremy Shapiro. 2013. “Household Response to Income Changes: Evidence from an Unconditional Cash Transfer Program in Kenya.” https:// www.princeton.edu/~joha/publications/Haushofer_Shapiro_UCT_2013.pdf. Hunt, Jennifer. 2007. “How Corruption Hits People When They Are Down.” Journal of Development Economics 84 (2): 574–89. Hunt, Jennifer, and Sonia Laszlo. 2005. “Bribery: Who Pays, Who Refuses, What Are the Payoffs?” NBER Working Paper 11635, National Bureau of Economic Research, Cambridge, MA. Jacoby, Hanan G., and Emmanuel Skoufias. 1997. “Risk, Financial Markets, and Human Capital in a Developing Country.” Review of Economic Studies 64 (3): 311–35. Kahneman, Daniel. 2011. Thinking, Fast and Slow. New York: Macmillan. Karlsson, Niklas, Peter Dellgran, Birgitta Klingander, and Tommy Gärling. 2004. “Household Consumption: Influences of Aspiration Level, Social Comparison, and Money Management.” Journal of Economic Psychology 25 (6): 753–69. Kaufmann, Daniel, Judit Montoriol-Garriga, and Francesca Recanatini. 2008. “How Does Bribery Affect Public Service Delivery? Micro-Evidence from Service Users and Public Officials in Peru.” Policy Research Working Paper, World Bank, Washington, DC. LAPOP (Latin American Public Opinion Project). 2010. Vanderbilt University, Nashville, TN. Latham, Gary P., and Lise M. Saari. 1979. “Importance of Supportive Relationships in Goal Setting.” Journal of Applied Psychology 64 (2): 151–56. Loayza, Norman V., Jamele Rigolini, and Oscar Calvo-González. 2014. “More Than You Can Handle: Decentralization and Spending Ability of Peruvian Municipalities.” Economics & Politics 26 (1): 56–78. Lokshin, Michael, and Martin Ravallion. 2004. “Household Income Dynamics in Two Transition Economies.” Studies in Nonlinear Dynamics and Econometrics 8 (3): 1–31. Marchionni, M., and P. Glüzmann. 2012. “Distributional Incidence of Social, Infrastructure, and Telecommunication Services in Latin America.” CEDLAS-UNLP, CONICET, Munich Personal RePEc Archive 42931. Martínez-Cruz, A. L., and C. Rodríguez-Castelán. 2014. “Crime and Chronic Poverty: A Long-Run Perspective for Mexico.” Working Paper. World Bank, Washington, DC. McPeak, John. 2004. “Contrasting Income Shocks with Asset Shocks: Livestock Sales in Northern Kenya.” Oxford Economic Papers 56 (2): 263–84. Mookherjee, Dilip, Debraj Ray, and Stefan Napel. 2010. “Aspirations, Segregation, and Occupational Choice.” Journal of the European Economic Association 8 (1): 139–68. Morduch, Jonathan. 1995. “Income Smoothing and Consumption Smoothing.” Journal of Economic Perspectives 9 (3): 103–14. Mullainathan, Sendhil, and Eldar Shafir. 2013. Scarcity: Why Having Too Little Means So Much. New York: Macmillan. Premand, Patrick, and Renos Vakis. 2010. “Do Shocks Affect Poverty Persistence? Evidence Using Welfare Trajectories from Nicaragua.” Well-Being and Social Policy 6 (1): 95–129. 96 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Ravallion, Martin. 1998. “Poor Areas.” In The Handbook of Applied Economic Statistics, edited by Aman Ullah and David E. A. Giles. New York: Marcel Dekker. Ray, Debraj. 2006. “Aspirations, Poverty, and Economic Change.” In Understanding Poverty, edited by Abhijit Vinayak Banerjee, Roland Bénabou, and Dilip Mookherjee, 409–21. Oxford Scholarship Online. Robinson, J. 2014. “The Misery in Colombia: Political Roots of Persistent Poverty.” World Bank, Washington, DC. Rosenzweig, Mark R., and Hans P. Binswanger. 1993. “Wealth, Weather Risk and the Composition and Profitability of Agricultural Investments.” Economic Journal 103 (416): 56–78. Rosenzweig, Mark R., and Kenneth I. Wolpin. 1993. “Credit Market Constraints, Consumption Smoothing, and the Accumulation of Durable Production Assets in Low-Income Countries: Investments in Bullocks in India.” Journal of Political Economy 1 (April): 223–44. Schady, Norbert, Jere Behrman, Maria Caridad Araujo, Rodrigo Azuero, Raquel Bernal, David Bravo, Florencia Lopez-Boo, Karen Macours, Daniela Marshall, Christina Paxson, and Renos Vakis. 2014. “Wealth Gradients in Early Childhood Cognitive Development in Five Latin American Countries.” Policy Research Working Paper 6779, World Bank, Washington, DC. Simon, Herbert A. 1955. “A Behavioral Model of Rational Choice.” Quarterly Journal of Economics 69 (1): 99–118. Thaler, Richard H., and Cass R. Sunstein. 2008. Nudge. New Haven, CT: Yale University Press. Weber, Max. 1946. “Bureaucracy.” In From Max Weber: Essays in Sociology, edited by H. H. Gerth and C. Wright Mills, 196–244. Oxford, U.K.: Oxford University Press. World Bank. 2004. World Development Report 2004: Making Services Work for Poor People. Washington, DC: World Bank. ———. 2014a. “Indigenous Peoples in Latin America: Political, Economic, and Social Overview at the End of the Second International Decade of the World’s Indigenous Peoples.” Washington, DC. ———. 2014b. World Development Report 2015: Mind, Society, and Behavior. Washington, DC: World Bank. Zimmerman, Frederick J., and Michael R. Carter. 2003. “Asset Smoothing, Consumption Smoothing and the Reproduction of Inequality under Risk and Subsistence Constraints.” Journal of Development Economics 71 (2): 233–60. UNRAVELING THE COMPLEXITIES OF CHRONIC POVERTY 97 CHAPTER 5 From Diagnosis to Policies: Crafting Coordinated Policies That Reduce Chronic Poverty C hronic poverty is a complex phenomenon that no single approach can deal with effectively. Policies have to be consistent with the social contract and political visions of society, budgets, capacity, and the overall institutional setting, including programs that already exist (Acemoglu, Johnson, and Robinson 2004). This chapter examines policy design. The first section synthesizes relevant findings from the diagnosis. The second section investigates the coordination of poverty-reduction efforts to build social protection systems, an emerging area of policy interest that can substantially increase the effectiveness of social assistance programs, particularly along the behavioral dimension. The third section makes the case for recognizing the state of mind of the poor in crafting policy. The last section argues that social intermediation services are critical for reducing chronic poverty. Because of the complexity and country specificities of policy design, we refrain from making concrete recommendations for the design of policies to deal with chronic poverty. Instead, we identify design elements that need to be taken into consideration to address some of the findings of our analysis. Improving Endowments and the Enabling Context The interaction between people’s ability to use their resources in the context in which they live is a key element of policy that needs to be understood to address chronic poverty. The diagnostics in chapter 4 suggest a few areas for policy. 99 Addressing the Intergenerational Transmission of Poverty The context in which a child is born should not affect his or her chances to suc- ceed in life. But, as this book showcases, children born in contexts of chronic pov- erty are less likely to have access to the same opportunities as luckier children. Malnutrition, poor stimulation, fragile health, absent parents, and a risky or violent environment prevent children from fully developing their potential and keep them in poverty. These differences keep growing over time. By the time social programs finally reach the adult population, some individuals may lack the skills and mind-set required to sustainably escape poverty through better and more stable employ- ment. The emerging policy dialogue and integration of early childhood develop- ment in the social development agenda in the region is therefore welcome and should be expanded. Improving the Environments in Which Poor People Live The environment in which people live matters as much, if not more, than people’s skills and characteristics. Because returns to individuals’ characteristics depend very much on the environment in which they live and the opportunities available, a family could be chronically poor if it lived in a remote district of the Andean sierra or Amazon selva but not poor if it lived in São Paulo, Bogota, or Lima. For a given level of skills, people with better cellphone coverage are likely to have better information and greater bargaining power. People who live in areas with paved roads are better able to sell their goods. People who work in manufacturing are more likely to earn higher wages than people who work in agriculture. People who benefit from universal health coverage are better able to cope with health crises, and people who live in areas with better policing are safer and more likely to earn higher profits from their business. Social policy should therefore balance direct support to the chronically poor with broader investments that improve the envi- ronment in which they live. Indeed, no social policy can be sustainable if it does not provide opportunities to improve incomes and living conditions. Bringing infrastructure and services to remote regions can be very expensive and trigger new challenges, such as cultural and environmental ones. And some regions may always be economically more depressed than others. For this reason, voluntary migration processes should be an integral part of social policies. Increasing Opportunities to Earn Labor Income Chronically poor households have fewer income earners than other households. Boosting their labor income is among the few known ways to sustainably pull them out of poverty. Comprehensive poverty-reduction programs should therefore include strategies that promote labor income, such as training and labor insertion programs. 100 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN Expectations about the impacts of such programs should be realistic, however: As a result of inequalities early in life, poor people often reach adulthood with fewer skills and aspirations than the nonpoor, severely weakening the potential impacts of income-generation policies and programs. Limited labor income is thus not only a cause but also a consequence of chronic poverty. Moreover, even if an income promotion program boosts income by 50 percent (close to the largest impacts that have been measured), many chronically poor families will remain in poverty and need continued assistance. Using the Number of Poor People Rather than the Incidence of Poverty as a Basis for Targeting Resources Many social programs base geographic coverage decisions on the incidence of poverty. Although the incidence of poverty is often highest in rural areas, the num- ber of poor people is often greater in urban areas. Targeting areas with the highest incidence of poverty thus fails to reach the largest number of poor people. Expanding programs to reach urban areas requires some serious rethinking that may go beyond coverage and identification. The urban poor are more mobile than the rural poor, making it more difficult to identify and support them. The rural and urban poor also have different sources of income, and the shocks and sources of vulnerabilities that keep them in poverty differ. Overall, supporting the urban poor may be more challenging than supporting the rural poor, which is why some programs, such as conditional cash transfers, have focused on rural areas. But greater complexity should not be a reason for not trying to address the grow- ing number of chronically poor people in urban areas. Coordinating Poverty-Reduction Efforts After decades in which the social contract ignored the poor, Latin America and the Caribbean (LAC) has become more inclusive. Over the last decade, the living con- ditions of the poor have improved, partly as a result of better services and social programs. The surge in social spending has been stunning: Colombia increased the number of new programs by a factor of seven, and El Salvador raised the num- ber from 1 to 30 (figure 5.1).1 Although the increase remains more modest in some countries, such as Argentina (largely because it started from a relatively high base), the trend was positive throughout the region. Social spending also increased dra- matically (figure 5.2). There is some evidence that increased spending reduces poverty (box 5.1). The increase in spending contributed to the reduction in income inequality and reduction during the decade and the rise of an emergent middle class (see Ferreira and others 2012), although economic growth, which resulted in higher labor incomes for the poor, played a much more significant role. By international FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 101 FIGURE 5.1 Number of new social assistance programs in selected countries in Latin America and the Caribbean, 1990–99 and 2000–11 40 Number of new programs 30 20 10 0 a il ile bia r or as o ru ay do az tin xic ad Pe Ch gu ur lom Br ua en Me nd lv u Ec Sa Ur g Co Ho Ar El 1990–99 2000–11 Source: World Bank LAC Social Protection database. FIGURE 5.2 Spending on social assistance in selected countries in Latin America and the Caribbean, 2000–10 3.0 2.5 2.0 Percent of GDP 1.5 1.0 0.5 0 or r ras ico ru y a il ile bia do az ua tin d Pe Ch x lom du Br ua lva ug n Me ge n Ec Sa Ur Co Ho Ar El 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Source: World Bank LAC Social Protection database. 102 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN BOX 5.1 Does increased public spending help reduce poverty? Evidence from Peru District data on public expenditures combined with information on poverty for 2007–12 in Peru sheds some light on the relationship between spending and poverty reduction (figure B5.1.1). In 2007, districts with poverty rates above 50 percent received similar amounts of public spending per capita (about 400 nuevos soles [$150]). By 2012, spending in districts that had reduced poverty below 50 percent had risen to almost 1,100 nuevos soles ($400), an increase of almost 200 percent. Spending also increased in districts in which poverty rates remained above 50 percent, but the increase was smaller (100 percent). Although this evidence is not causal, it may suggest that increased spending is a factor in reducing chronic poverty. FIGURE B5.1.1 Median public spending per capita in “improved” and chronically poor districts of Peru 1,200 1,000 Per capita spending (nuevos soles) 800 600 400 200 0 Improved districts Chronically poor districts 2007 2012 Sources: Data from Sistema Integrado de Administración Financiera (SIAF) and Instituto Nacional de Estatistica e Informatica (INEI) and poverty maps for 2007 and 2012. Note: Improved districts = districts in which poverty rate was above 50 percent in 2007 but below 50 percent in 2012. Chronically poor districts = districts in which poverty rate was above 50 percent in both 2007 and 2012. Values are in 2007 nuevos soles. standards, however the social budget of most LAC countries remains modest: Many middle-income countries in Eastern Europe, such as Croatia, Romania, the Russian Federation, and Ukraine, spend 3 percent of GDP or more on social assis- tance (Tesliuc and others 2014). LAC also lags behind other regions in terms of tax collection, which has remained relatively low and often regressive, making it more difficult for governments to address the poverty challenge. FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 103 Indeed, despite improvements, a fifth of the population in LAC remains chron- ically poor. For them, neither the social contract nor economic growth have been enough. The existing policy toolkit is not sufficient; something more needs to happen. A more systematic and coordinated approach is needed. Programs and initia- tives need to have not only clear, measurable, individual goals, they also have to fit into a comprehensive poverty-reduction framework. Comprehensive poverty- reduction programs should include income promotion strategies, such as training and labor insertion programs, in addition to social assistance and improvements in local infrastructure and connectivity. Rather than being implemented as stand- alone programs, these income promotion strategies should have conceptual and physical connections to social assistance programs. Coordination needs to happen at all levels. At the level of the state, there must be broad agreement that social programs are not populist initiatives but effective social inclusion tools. To achieve such a consensus, the generosity of benefits and type of programs may have to be adjusted to society’s perceptions of the scope of mutual help and solidarity. At the level of the executive, ministries and pro- grams must be given adequate resources. Technically competent program staff who are shielded from political pressures should be appointed. Poverty-reduction strategies should ensure that ministries and agencies are not only given clear and measurable goals but that they work together toward achieving those goals in a coordinated manner. Lack of systematic coordination is a pervasive feature in many countries in the region (Camacho and others 2014). The lack of dialogue between conditional cash transfer and health service programs often prevents the early detection of malnour- ished children. Without the integration of programs’ databases, it is not possible to know which benefits poor families are receiving. The result is excess cover- age for some families and lack of adequate assistance for others. Uncoordinated expansions of social programs can drastically reduce the effectiveness of the whole system, as beneficiaries receive only bits and pieces of a package. Well-conceived poverty-reduction strategies often fail because each ministry or agency focuses only on its own objectives, missing the bigger picture and the need for coordina- tion not only at the ministerial level but on the ground (Trivelli 2014). Spending more on social assistance may not be enough to lift the chronically poor out of poverty; there is also a need to spend better. The efficiency of indi- vidual programs is increasingly coming under scrutiny; there is less discussion of how individual programs should coordinate with one another to build an effective social protection system. Many countries in LAC have established coordinating agencies, reflecting a growing awareness of the need to align institutions in the fight against poverty. Such agencies take many forms, including supra-ministries (Ecuador), ministries 104 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN (Peru), and agencies within a ministry or the presidency (such as the National Agency for Overcoming Extreme Poverty [ANSPE] in Colombia). Nine out of 19 surveyed countries in LAC have an agency in charge of coordinating the social development agenda. The scope and nature of these agencies varies (table 5.1). Some, such as the Ministerio de Desarollo e Inclusion Social (MIDIS) in Peru and the Ministerio de Desarollo Social (MDS) in Chile, have been given a monitoring and evaluation role. Others have a more limited scope. Some have legislative power; others imple- ment their own programs. In all countries, the finance ministry remains ultimately responsible for financing social development programs. TABLE 5.1 Social development and institutional arrangements in Latin America and the Caribbean Number of agencies Unique coordination Planning and Executing Country agency objectives Financing agencies Monitoring Evaluation Argentina Yes 1 1 1 0 0 Bolivia No 2 1 2 0 0 Brazil No 2 1 2 0 0 Chile Yes 1 1 2 1 1 Colombia No 1 1 1 1 0 Costa Rica Yes 1 1 1 0 0 Dominican Republic Yes 1 1 3 1 1 Ecuador No 2 1 2 1 1 El Salvador No 1 1 3 0 0 Guatemala Yes 1 1 1 1 1 Honduras No 1 1 3 0 0 Mexico Yes 1 1 1 1 1 Nicaragua Yes 1 1 1 1 1 Panama No 2 1 2 0 0 Paraguay No 1 1 3 0 1 Peru Yes 1 1 2 1 1 Uruguay Yes 1 1 1 1 1 Venezuela, R. B. No 1 1 2 0 0 Source: Szekely 2013. Note: Figures indicate number of agencies. FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 105 To complement these efforts, some countries have also launched large initiatives to fight poverty. Examples include Brasil Sem Miséria (Brazil without Poverty) and Mexico’s Cruzada Nacional contra el Hambre (National Crusade against Hunger) (box 5.2). The main difference between these initiatives and coordination agencies is their more temporary nature; they fall somewhere between large programs and agencies. Do these efforts work? The extent to which coordination improves the effec- tiveness of the social protection system remains unclear and is probably country specific. It is much more difficult to evaluate a system than a program, because traditional impact evaluation methodologies cannot be used. Qualitative reviews, however, suggest that the absence of a coordinating agency with some authority over the decision-making process of other ministries and agencies and some con- trol over the budget reduces the effectiveness of poverty-reduction efforts. A recent review of the establishment of Peru’s Ministry of Development and Social Inclusion (MIDIS) provides some insights about the challenges that imple- mentation efforts may face and draws some lessons for improving the effectiveness of coordination efforts (Trivelli 2014). Despite differences across the region in terms of how these models operate, it yields some lessons that seem generalizable (see box 5.2). BOX 5.2 Coordinating poverty-reduction efforts in Brazil, Mexico, and Peru To accelerate their poverty-reduction efforts, Brazil, Mexico, and Peru have launched ambitious interministerial strategies coordinated by their ministries of social development. These efforts integrate existing social programs, with the aim of improving coordination and refocusing coverage toward the neediest districts. Brasil Sem Miséria During Luis Lula da Silva’s administration, Brazil implemented Bolsa Família, a conditional cash transfer program. Although the program had substantial impacts on poverty-reduction, 1.5 million families in extreme poverty were still not receiving assistance in 2010. To reach and support these families, in June 2011 Dilma Rousseff’s government set up Brasil Sem Miséria (Brazil without Misery [BSM]), with the goal of eradicating extreme poverty and generating opportunities for the poor. Upon implementation, the government discovered that more than 1 million families in extreme poverty were not included in the Cadastro Único (Unified Registry), the entrance door to Bolsa Família and BSM. BSM includes about 100 lines of action managed by 22 ministries. Efforts are coordinated by the Ministry of Social Development and the Fight against Hunger (MDS), through a temporary Special Secretariat for Overcoming Extreme Poverty (SESEP). The strategy is based on three pillars: guaranteed income for immediate alleviation of extreme poverty; access to public services; (continued on next page) 106 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN BOX 5.2 Coordinating poverty-reduction efforts in Brazil, Mexico, and Peru (continued) and productive inclusion in urban and rural areas to improve employment opportunities, income generation, and welfare. BSM targets the extremely poor population, with a focus on children and adolescents. It has national coverage, with the northeastern regions, where extreme poverty rates are higher, receiving special attention. Like Bolsa Família, it has a decentralized structure in order to enable more agile execution. With respect to the first pillar (guaranteed income), in 2011 BSM increased benefits to children and adolescents and included an additional payment for pregnant and lactating mothers. In 2012 BSM launched a new program, Ação Brasil Carinhoso (Actions for Loving Brazil), which provides new cash benefits for Bolsa Família beneficiaries with children up to 15 years old and per capita income of less than R$70 a month (the national poverty threshold). In March 2013 the program’s benefits were extended to all Brazilian families in extreme poverty and its name was changed to Benefício para Superação da Extrema Pobreza (BSP). Expansion helped another 2.5 million people reach the national poverty threshold through cash transfer support. The total value of cash transfers of Bolsa Família increased by 60 percent between 2010 and 2014, from R$17.37 billion to R$24.7 billion. Coverage also increased, from 13.4 million families in 2010 to 14.1 million in 2014. With respect to the second pillar (access to public services), BSM is coordinating and expanding primary care services for the extreme poor, supplying, among others, basic health units, doctors, and boats to cover remote areas in the Amazon and the Pantanal. Ação Brasil Carinhoso finances the prevention and treatment of illnesses affecting development in early childhood, expanding distribution of vitamin A and ferrous sulfate and providing free medicines for asthma. It provides financial incentives to municipalities to build childcare facilities for extremely poor children and offers daycare, day-long education, and school meals in schools with high rates of beneficiaries of Bolsa Família. With respect to the third pillar (productive inclusion), the Programa Nacional de Acesso ao Ensino Técnico e Emprego (PRONATEC) has qualified 1 million workers to date. It expanded entrepreneurship support for the extremely poor, providing technical and managerial assistance through the Serviço Brasileiro de Apoio às Micro e Pequenas Empresas (Sebrae). BSM also reoriented microcredits from federal public banks through the CRESCER program, with reduced annual interest rates. In rural areas, BSM provides families with individualized technical assistance to increase the production, quality, and value of their products. Poor farmers receive inputs, seeds, and R$2,400 to implement agricultural production projects. Through Programa de Aquisição de Alimentos (Food Procurement Program), municipalities and states buy produce from farmers to supply public universities, hospitals, prisons, and other public institutions. Two new programs, Água para Todos (Water for All) and Luz para Todos (Light for All), expanded access to water and electricity in rural areas. Mexico’s Cruzada Nacional contra el Hambre President Peña Nieto announced the launching of the Cruzada Nacional contra el Hambre (National Crusade against Hunger [CNCH]) shortly after taking office in January 2013. The program is (continued on next page) FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 107 BOX 5.2 Coordinating poverty-reduction efforts in Brazil, Mexico, and Peru (continued) designed to reduce extreme poverty levels by raising living standards in Mexico’s least developed municipalities and building a coordinated social network from the bottom up. CNCH seeks to ensure food security and nutrition for 7.4 million Mexicans living in extreme poverty through the coordination and reorientation of 70 social assistance programs (mostly preexisting) toward priority municipalities. In a first stage, 400 municipalities were prioritized as beneficiaries; the program has since been expanded to include more than 1,000 municipalities. Led by SEDESOL (the Secretariat for Social Development), the program involves 16 ministries, civil society, and the private sector. CNCH aims to achieve five goals: eradicate hunger, eliminate child malnutrition and improve children’s growth indicators, increase farmers’ production and income, minimize food losses after harvest and during storage and transportation, and promote civic engagement. One of its most important roles is the coordination of all actors involved in the crusade. In building a bottom-up social network, CNCH acts as a liaison between the central government, regions, and municipalities. The program created local committees for community development in all CNCH municipalities to strengthen citizen engagement in identifying community needs and social gaps and to monitor the accomplishment of CNCH objectives and the transparency of program actions. It also established an interministerial commission and a committee of experts. To ensure buy-in and accountability at the ministerial level, the government holds every minister personally responsible for coordinating and implementing CNCH. In addition to Prospera (Mexico’s conditional cash transfers program, previously called Oportunidades), social assistance programs embraced by CNCH include the food support program Pal Sin Hambre, which seeks to improve food supply through stores, dairies, canteens, schools, and community kitchens; the Programa de Abasto Rural; the Programa de Abasto Social de Leche, which supplies milk to vulnerable communities; and the nutritional tracking of 1.1 million children under the age of five. Other programs aim at decreasing the number of people with incomes below the minimum level of well-being (Mex$824.56 in rural areas, Mex$1,161 in urban areas) from 7.4 million to 5.6 million. They include the PROCAMPO Productivo; the Programa de Empleo Temporal; the Programa de Fondos Nacionales Indígenas; the Pensión para Adultos Mayores; the Programa Nacional de Financiamiento al Microempresario; and the Fondo de Microfinanciamiento a Mujeres Rurales. In education, CNCH funds scholarship programs in order to increase school attendance by children ages 3–15 and to certify people born after 1982 who did not complete elementary or secondary school. In health, the Caravanas de la Salud and Seguro Médico Siglo XXI provides vulnerable people with access to health services and medicines and builds health centers. These programs supplement Seguro Popular, which grants free health coverage to 3.8 million Mexicans. Other health programs monitor children’s weight and height, and provide nutrients to pregnant women and children under age five. The CNCH also funds infrastructure, including housing (the Programa de Vivienda Digna, Programa de Vivienda Rural, and Programa de Garantías para la Vivienda Popular), water, electricity, and sanitation programs. (continued on next page) 108 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN BOX 5.2 Coordinating poverty-reduction efforts in Brazil, Mexico, and Peru (continued) Peru’s Ministry of Development and Social Inclusion The Ministry of Development and Social Inclusion (MIDIS) was established in 2011 with the objectives of coordinating Peru’s poverty-reduction efforts and implementing cross-sectoral poverty-reduction programs. Beginning in the 1990s, Peru made several attempts to build programs and policies to assist the poorest and most vulnerable segments of the population. Various state agencies—from the defunct Ministry of the Presidency during the 1990s to the Presidency of the Council of Ministers until very recently—administered social programs. Each sector or agency was responsible for one or more programs to help poor families, but there was little coordination among agencies or social service providers. Results were fragmented. The administration of President Ollanta Moisés Humala Tasso created the MIDIS, with a mandate to serve as the governing body for development and social inclusion policies. It seeks to ensure that the benefits of more than a decade of economic growth trickle down to the poor and to enable the poorest households to overcome the conditions of exclusion and vulnerability that affect them. Implementation of the MIDIS involved major challenges. It required reorganizing sectors, introducing new practices in social programs, evaluating each program, and proposing a comprehensive strategy for poverty reduction and social inclusion that coordinated efforts of other programs and sectors of the national and subnational governments. Creation of the MIDIS also demanded justifying the need for a fresh look at social policy. The widespread perception that social programs were a political instrument also had to change. To do so, it was crucial to set clear objectives, goals, and measurable indicators of success. In consultation with the government and other actors, the MIDIS proposed a strategy based on six goals, with measurable indicators for each of them. These goals were adopted by the public sector in its entirety; in 2012 they were included as the social goals of Peru’s multiannual macroeconomic framework. Sources: Program information and Trivelli 2014. Designing Policies with Clear, Specific, and Measurable Objectives In designing a strategy, it is very important to set goals to be achieved by each intervention and define clear responsibilities and competencies in the implemen- tation of each intervention. Doing so requires identifying who has the legal compe- tence and the financial and operational resources to act and who is responsible for leading thematic areas. It is important to specify clear rules of operations, inputs, and expected outputs and to ensure that implementation is transparent, which requires, among other things, a strong monitoring and evaluation component and independent evaluations. Clear rules and transparency are important not only to ensure effective coordination but also to avoid social programs becoming hostages of political cycles. Programs must be perceived as serving a critical need on the long-term path to development and social inclusion. Constant work is needed to FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 109 raise the quality of programs, by defining quality standards, establishing proce- dures for quality monitoring and improvements, and creating mechanisms for lis- tening and responding to beneficiaries’ feedback. Providing Incentives to Coordinate Incentives and opportunities must be given to people who contribute to the process, and people who do not contribute must be sanctioned. The design of incentives, sanctions, and institutional commitments must keep inno- vating by creating new mechanisms, assessing them, and improving them. Incentives that make it is easier and less expensive to operate jointly than to work in isolation must be crafted. There is also a need to develop tools that facilitate the process, such as data- bases that can be shared and validated by different sectors; transparent indica- tors that can be easily measured and updated; common targeting schemes that allow the identification of geographic areas and social groups for which coordi- nated action would increase impact; joint prioritization of budget and geographic areas across sectors to ensure that vulnerable populations receive a comprehensive package of assistance; and establishment of an accountability system that sanctions failure to perform. Showing Evidence of the Benefits of Coordination The benefits of coordination should be demonstrated in order to ensure that social policy stops being perceived as an inefficient expense and instead becomes a central pillar of a country’s development efforts. Social sectors must demonstrate that they have created rigorous processes for evaluating programs, generating assessment plans, and establishing systems for moni- toring their social programs. They must also develop strategies with mea- surable objectives that include records of the cost of social programs, so that budget forecasts can be made and programs evaluated for cost-effectiveness. Increases in budgets should be tied to results—through, for instance, a process of results-based budgeting. A central aim of coordination is to reduce the costs to the government of service provision by eliminating duplicative processes or programming. Boosting Coordination at the Local Level Effective coordination at the local level—among education, health, and coordinat- ing agencies, for example—can be difficult if local stakeholders do not support coordination. Outputs by civil servants, teachers, doctors, and suppliers (student performance, health improvements), not inputs (number of workshops conducted, number of patients treated), must be benchmarked and rewarded. At the same time, 110 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN coordination is costly across time and space because of its complexity. Policy mak- ers should thus seek the degree of coordination that maximizes net benefits (ben- efits less costs). Recognizing the State of Mind of the Poor in Crafting Policy The psychosocial aspects of chronic poverty affect every step of the process of upward mobility, from the mere consideration of engaging in an activity to taking action and having it bear results. Policy design therefore needs to incorporate the mind-set of the chronically poor, including their often lowered aspirations. If these issues remain unaddressed, the poor may fail to register for social programs or comply with their co-responsibilities. This section showcases the importance of addressing the state of mind of poor people in social programs. It describes a range of areas where the link between behavioral aspects and policy design matters. (For an extensive set of examples in LAC and the rest of the world, see the 2015 World Development Report.) Taking up Opportunities The first step in the process of escaping poverty is engaging with the process and taking up opportunities. Individuals’ state of mind can severely affect their readi- ness to engage: If they are stressed or have low aspirations, they may not even consider taking up an opportunity. Productive interventions such as vocational training, microfinance, and business grants are widespread in the developing world and tend to include self-targeting mechanisms. These programs may fail to reach the very people who stand to benefit most from them if their low aspirations, stress, and in some cases depression make them perceive returns as lower than they are likely to be. Carneiro, Galasso, and Ginja (2014) look at participation in Chile’s Subsidio Familiar (SUF), a household allowance scheme. They show that along most of the income distribution, there is a negative relationship between income and partici- pation (as there should be): The richer a household is, the less likely it is to par- ticipate. The poorest Chileans, however—who would benefit the most from the household allowance—are about 20 percent less likely to participate in the SUF than families that are less poor. An experimental study from Nicaragua provides insights on the use of defaults (Macours, Premand, and Vakis 2014). A pilot provided business grants or voca- tional skills training to randomly selected eligible households in poor rural com- munities. Data collected before the pilot was offered suggested that the poorest households had little interest in engaging in these income-generation activities. To attract these households, the program bundled business grants and training FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 111 with a conditional cash transfer linked to early childhood outcomes. This bun- dling induced all households to sign up. The results were striking: Compared with similar households in the control group, the poorest households that were offered the business grant were almost 15 percentage points more likely to engage in self- employment, and the poorest households that were offered vocational training were 10 percentage points more likely to engage in self-employment. Even more surprisingly, the poorest households—who before the program showed no interest in joining such schemes—earned more in nonagricultural self-employment than other beneficiaries. The authors suggest that aspirational and motivational changes may have driven these results. They show that the program made participants more optimis- tic and motivated about their futures. Signing up to receive the conditional cash transfer forced them to spend time exploring the income-generation component. Once they did, they recognized the potential of engaging with the business grant and training components. This example highlights two potential insights for policy design in a context of low aspirations: Offering simple choices can increase the likelihood that people take up an intervention, and programs that explicitly seek to change aspirations as a complement to other types of support can help change people’s assessment of the returns to different opportunities. Another example comes from Peru’s renowned tuberculosis (TB) preven- tion and treatment program, which provides treatment, helps patients register for health insurance, screens people living in the same household as patients and peo- ple undergoing HIV testing, and provides vaccination shots to children, all free of charge. Despite these efforts, Peru has one of the highest incidences of TB in the world. The two big challenges are the failure of patients to seek diagnosis early on, which increases both the severity of the disease and the probability of infecting others, and low adherence to treatment, which is fairly long (a daily pill needs to be taken for at least six months for the basic strand of TB), which increases the chances of recurrence and multidrug resistance. A majority of the population affected by TB resides in the slums of Lima, where chronic poverty, stress, and depressed aspirations are the norm. TB-affected households are more likely to feel stigmatized, cry, and think about suicide; tests for depression revealed that almost a third of TB patients were moderately or severely depressed and that 1 in 30 admitted to thinking about committing suicide (Karlin and others 2011). Depression has a big impact on behavior: TB patients who were depressed at the time of diagnosis had a 43 percent greater likelihood of abandoning treatment before they were cured (figure 5.3). Attempts to treat depression with drugs at the health centers seemed to be ineffective, unsustainable, and difficult to integrate with TB treatment. A local nongovernmental organization adopted an alternative approach. The Innovative Socioeconomic Interventions against Tuberculosis (ISIAT) project team designed 112 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN FIGURE 5.3 Adherence to treatment for tuberculosis among depressed and nondepressed patients in Peru 20 Percent of patients not adhering to treatment 10 0 0 1 2 3 4 5 6 Months in treatment Depressed Not depressed Source: Karlin and others 2011. a package of support activities aimed at helping TB patients overcome the stigma of the illness while exploring income-generation opportunities. As part of the proj- ect, clinical psychologists worked with TB patients in small groups or community workshop activities. These activities led to a reduction in depression. The effects on TB prevention were dramatic: Households that received the ISIAT intervention were far more likely to vaccinate their children against TB. Before the intervention, less than a quarter of TB-affected families did so, even though vaccination was recommended and offered for free; the rate of take-up was lowest among the poorest households, at just 17 percent (figure 5.4). In commu- nities receiving the ISIAT intervention, the share of the poorest households vac- cinating their children more than doubled, to 41 percent. Integrating stress and depression into the design of this TB reduction program was critical. FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 113 FIGURE 5.4 Effect of socioeconomic intervention on vaccination against tuberculosis of children in Peru 50 45 Percent of children vaccinated against tuberculosis 40 35 30 25 20 15 10 5 0 Control TB-affected TB-affected families families (without receiving socio- intervention) economic intervention Poorer Less poor Source: Rocha and others 2011. Note: All households with a member with tuberculosis (TB). Shows % of children that completed chemoprophylaxis. Dealing with Reduced Mental Resources People in poverty have to devote much of their time and mental resources to tack- ling survival problems. This obligation causes stress and limits the time available for other tasks, such as long-term planning. A pilot program from Bogota, Colombia, attempted to address the fact that stress can reduce the value people place on the future. Beneficiaries of a condi- tional cash transfer program who received bimonthly payments were randomly split into two groups. The first group received the full amount each payment day. The second group received only two-thirds of the payment, with the last third set aside by the program team in a savings account. The total savings were given to beneficiaries as a lump sum in December, just before children enroll in school. While both payment schemes had similar impacts on school attendance, the “save for when you need it” approach increased reenrollment rates (Barrera-Osorio 114 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN and others 2011). By simply changing the timing of the payments, the program made savings available to parents when they most needed them, resulting in higher reenrollment rates. Evidence from Bolivia and Peru on the use of telephones to remind people to save also showcases how powerful such simple interventions can be. For exam- ple, compared with a control group, participants who received reminders to save through text messages improved savings rates by almost 5 percent in Peru (Karlan and others 2010). The type of communication also mattered: Reminders that emphasized a specific goal, such as saving to reach a particular amount, almost tripled the impact compared with the generic reminder. These examples highlight the potential of these types of interventions and their applicability in a wide range of policy areas. Adding simple solutions to existing interventions may go a long way, at little cost. Raising Aspirations Another channel through which the mental state operates is its impact on aspirations, which can be influenced by local social norms and peer effects. The earlier example of combining productive interventions with a condi- tional cash transfer in Nicaragua provides some evidence of how this channel works. The program targeted the vast majority of women in each community (on average 90 percent of households in a community were eligible), explic- itly encouraging group formation by creating opportunities for social interac- tions among beneficiaries. Because local leaders were among the program’s beneficiaries, the authors were able to measure both the overall impact of the program and whether social inter- actions between leaders and other beneficiaries led to additional impacts. They found that interactions with leaders (and role models in general) matter a great deal: Beneficiaries who interacted more with local leaders (measured as proximity to the leader) invested more in their children. Among beneficiaries who received a business grant, social interactions with community leaders increased school attendance rates by additional 9.7 percentage points (above the overall program impact) and expenditures on food with higher nutritional value by an additional 2 percentage points. Social interactions also amplified income effects (figure 5.5). Income from nonagricultural activities among beneficiaries who received the business grant and who lived near a local leader rose by an additional $3.30, and the average value of a household’s animal stock increased by an additional $12. Given that the average baseline income from nonagricultural activities was $8.75 per capita, social inter- actions alone raised income by 40 percent. In Brazil, soap operas are used to transmit positive messages. In a pilot pro- gram, a key recurrent character was a woman who opted for a smaller household FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 115 FIGURE 5.5 Effect on income of social interaction with local leaders in Nicaragua 600 500 Income (córdobas) 400 300 200 100 0 Control group Beneficiaries receiving Beneficiaries receiving business grant business grant and interacting socially with local leaders Source: Macours and Vakis 2014. in order to take up opportunities that could put her on an upward mobility trajec- tory. La Ferrara, Chong, and Duryea (2012) find causal evidence suggesting that fertility rates started declining among women who watched the show for a year, particularly women closer in age to the character. Among women 35–44, fertil- ity declined by more than 10 percent—equivalent to the impact of two additional years of education. Interventions from the United States also show how well-designed programs can change mind-sets. Blackwell, Trzesniewski, and Sorich Dweck (2007) assess the impact of including an eight-week module on the malleability of the brain and the growth mind-set (discussed in chapter 2) on an existing study skills workshop for seventh graders in New York City. While the math scores of students in the control group—who participated only in a study skills workshop—spiraled down during the school year, the scores of the treatment group climbed. A randomized trial by Duckworth and others (2013) studied fifth graders of low socioeconomic status in the United States who acquired mental tools to avoid distraction and temptation and thus persevere in their studies. It finds better grades, school attendance, and conduct. These illustrations show that it is possible to shift households’ aspirations and behavior, particularly when interventions that aim to influence attitudinal shifts are complemented by resources to allow beneficiaries to seek opportunities. There is not necessarily a need to develop brand new programs; small changes to exist- ing programs that exploit such behavioral insights are often a cost-effective way to improve the impacts of programs that address chronic poverty. Simply taking into account existing interventions and redesigning them to simplify processes, social norms, or social interactions can yield large returns. Natural leaders and 116 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN role models living in the community can also be important vehicles for change by motivating and encouraging others and providing examples that people aspire to follow. Teachers, who already serve as role models, can help children change their attitudes toward learning by incorporating the mind-set approaches that seem to be working in some settings. Developing Coordinated Solutions that Address Behavioral Constraints: Social Intermediation Services In their worst forms, behavioral barriers can induce the chronically poor to exclude themselves from the very social programs designed to assist them. The presence of numerous uncoordinated social programs, all with different eligibility rules, does not help reach them. In the classical “passive” social assistance approach, budget- ary considerations and poverty status dictate coverage, and the poor are assumed to seek out and enroll in programs on their own. This approach represents a shift away from the traditional social assistance paradigm of providing the poor with a broad range of services and goods. It features a more personalized approach that aims at giving people the building blocks needed to overcome the specific challenges they face. In contrast to the classical approach in which families are required to apply for benefits, social intermediation services put families at the center by actively identi- fying and approaching them and guaranteeing them priority access to existing and newly created programs. Efforts to offer a more systematic response to both behavioral and coordination constraints have been emerging in many LAC countries. These social intermedia- tion services are intended to help the poor overcome information and other barriers through a holistic, systemic, and household-based approach (Camacho and others 2014). These services have the dual objective of facilitating access of the chroni- cally poor to existing programs and increasing their chances of exiting poverty by addressing their state of mind and specific constraints. A family support component—in which one-on-one support is provided—is the central pillar of social intermediation services. In its most sophisticated form, social workers work with the family to develop a personalized development plan, which is then monitored and discussed on a regular basis. Such an active approach requires bending the classical institutional design of social policies and requires heavy institutional coordination, so that existing social programs reach and priori- tize the chronically poor. Social intermediation services do not bring material benefits to families directly; instead they facilitate access to other programs. They must therefore be well inte- grated within the social assistance system, have interoperable information systems to track the supply of social services and population demand, and employ well- trained social workers to match families to the social programs that will address their needs. FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 117 Social intermediation services stand between the demand for and supply of social services. Coordination with the supply side is therefore essential for their success. Ensuring the quality of supply is as important as addressing the demand for social services: Granting access to services of poor quality or services that are poorly tailored to the needs of the chronically poor may lead to little or no impact. At times the greatest increase in access and impact can be achieved by solving supply-side constraints, such as simplifying enrollment procedures or improving the quality of health services, rather than implementing a whole new program. Social workers are the backbone of family support. They need to meet mini- mum qualifications and be well trained. In addition to having knowledge of the rules and procedures of all social assistance programs, they need to understand the informational and psychosocial barriers faced by the extreme poor and be able to dialogue with families to help them overcome these barriers. Visits to families must take place on a regular basis and be tailored to individual needs—behavioral barriers can be overcome only if families feel that their constraints are understood and that social workers are willing and able to help them overcome them. Social intermediation services can vary in sophistication. Chile Solidario seeks to do more than just provide access to social programs; it also includes psycho- social support and programs to fill supply gaps. Where resources and capacity are more limited, simpler programs that focus mainly on access can also lead to positive impacts. The very essence of social intermediation services—overcoming psychosocial barriers to ensure access to other programs—makes it difficult to assess their effectiveness, which depends on the extent to which these services facilitate the poverty-reduction efforts of other programs. These challenges not- withstanding, Camacho and others (2014) review two of these programs: Chile Solidario, the first such program in the region, and Red Unidos (in Colombia). The two programs are similar in design but have different implementation models. Chile Solidario is an intensive and highly coordinated intervention. Beneficiary families receive regular visits from social workers, who are trained and evaluated on a regular basis. A database that is integrated across programs and ministries informs monitoring and implementation. The intervention has resources available to influence the expansion of other social programs and the supply of social services to cover its beneficiaries. Red Unidos is less comprehensive. Visits to beneficiary families occur less often, social workers are not evaluated on a regular basis, databases are not inte- grated with other programs and ministries, and the program does not have the resources to finance the expansion of other social programs and the supply of social services to cover its beneficiaries. Camacho and others argue that social intermediation services can be powerful and cost-effective tools for supporting poor and marginalized families by facilitat- ing their access to social programs; improving their socioemotional well-being; and, if the right conditions are provided, raising their employment prospects. 118 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN These services show that psychosocial constraints are not insurmountable barri- ers and that, at least along the access dimension, it is possible to reach out to the chronically poor and ensure that they benefit from assistance. Social intermedia- tion services also seem to be cost-effective: In both Chile and Colombia, it costs less than $100 a year to cover most beneficiaries, who tend to be located in regions with easy access and high population density. Coordinating efforts and dealing with the state of mind of the poor adds complexity and costs. But social intermediation services hold great promise. Unless countries actively reach out to the chronically poor, they are likely to fall through the cracks of the social assistance system, perpetuating the vicious cycle of poverty. Note 1. The World Bank’s LAC Social Protection database includes information about the year a program was created for each program that existed in or after 2003. We use this information (and the assumption that few programs were closed) to compare the creation of programs in the 1990s and the 2000s. The data may miss some programs that opened in the 1990s and closed before 2003, but it is easier to open a program than to close it, and it is unlikely that the closure of a few programs would alter the overall picture. Social assistance consists of noncontributory programs usually targeted to the poor or vulnerable. Some programs focus on ameliorating chronic poverty or providing equality of opportunity; others focus on protecting families from shocks and the long-term losses they can impose on the unprotected poor. These safety net or social welfare programs include cash transfers (conditional and unconditional) and in-kind transfers, such as school meal programs and targeted food assistance (for a description of such programs see Cerutti and others 2014). References Acemoglu, D., S. Johnson, and J. Robinson. 2004. “Institutions as the Fundamental Cause of Long-Run Growth.” In Handbook of Economic Growth, edited by Phillipe Aghion and Steven F. Durlauf. Amsterdam: North Holland. Barrera-Osorio, Felipe, Marianne Bertrand, Leigh L. Linden, and Francisco Perez-Calle. 2011. “Improving the Design of Conditional Transfer Programs: Evidence from a Randomized Education Experiment in Colombia.” American Economic Journal: Applied Economics 3 (2): 167–95. Blackwell, Lisa S., Kali H. Trzesniewski, and Carol Sorich Dweck. 2007. “Implicit Theories of Intelligence Predict Achievement across an Adolescent Transition: A Longitudinal Study and an Intervention.” Child Development 78 (1): 246–63. Camacho, Adriana, Wendy Cunningham, Jamele Rigolini, and Veronica Silva. 2014. “Addressing Access and Behavioral Constraints through Social Intermediation Services: A Review of Chile Solidario and Red Unidos.” Policy Research Working Paper 7136, World Bank, Washington, DC. FROM DIAGNOSIS TO POLICIES: CRAFTING COORDINATED POLICIES THAT REDUCE CHRONIC POVERTY 119 Carneiro, Pedro Manuel, Emanuela Galasso, and Rita Ginja. 2014. “Tackling Social Exclusion: Evidence from Chile.” IZA Discussion Paper 8209, Institute for the Study of Labor, Bonn. Cerutti, Paula, Anna Fruttero, Margaret Grosh, Silvana Kostenbaum, Maria Laura Oliveri, Claudia Rodriguez-Alas, and Victoria Strokova. 2014. “Social Assistance and Labor Market Programs in Latin America: Methodology and Key Findings from the Social Protection Database.” Working Paper 88769, World Bank, Washington, DC. Duckworth, Angela Lee, Teri A. Kirby, Anton Gollwitzer, and Gabriele Oettingen. 2013. “From Fantasy to Action: Mental Contrasting with Implementation Intentions (MCII) Improves Academic Performance in Children.” Social Psychological and Personality Science: 1948550613476307. Ferreira, Francisco H. G., Julian Messina, Jamele Rigolini, Luis-Felipe López-Calva, Maria Ana Lugo, and Renos Vakis. 2012. Economic Mobility and the Rise of the Latin American Middle Class. Washington, DC: World Bank. Karlan, Dean, Margaret McConnell, Sendhil Mullainathan, and Jonathan Zinman. 2010. “Getting to the Top of Mind: How Reminders Increase Saving.” NBER Working Paper 16205, National Bureau of Economic Research, Cambridge, MA. Karlin, D. M., C. Evans, C. Loiselle, K. Zevallos, F. Fernandez, N. Allen, R. Montoya, and D. Boccia. 2011. “Depression Is Common in TB Patients and Is Associated with Treatment Abandonment.” Journal of Investigative Medicine 60: 167, 197. La Ferrara, Eliana, Alberto Chong, and Suzanne Duryea. 2012. “Soap Operas and Fertility: Evidence from Brazil.” American Economic Journal: Applied Economics 4 (4): 1–31. Macours, K., P. Premand, and R. Vakis. 2014. “Demand versus Returns? Pro-poor Targeting of Business Grants and Vocational Skills.” Policy Research Working Paper 6389, World Bank, Washington, DC. Macours, Karen, and Renos Vakis. 2014. “Changing Households’ Investments and Aspirations through Social Interactions: Evidence from a Randomized Transfer Program.” Economic Journal 124 (576): 607–33. Rocha, C., R. Montoya, K. Zevallos, A. Curatola, W. Ynga, J. Franco, F. Fernandez, N. Becerra, M. Sabaduche, M. A. Tovar, E. Ramos, A. Tapley, N. R. Allen, D. A. Onifade, C. D. Acosta, M. Maritz, D. F. Concha, S. G. Schumacher, and C. A. Evans. 2011. “The Innovative Socio-economic Interventions against Tuberculosis (ISIAT) Project: An Operational Assessment.” International Journal of Tuberculosis and Lung Disease: The Official Journal of the International Union against Tuberculosis and Lung Disease 15 (Suppl 2): S50–57. Szekely, Miguel. 2013. “Arreglos institucionales en el sector de desarollo social en America Latina: análisis e implicaciones para el caso Peruano.” World Bank, Washington, DC. Tesliuc, Emil, Lucian Pop, Margaret Grosh, and Ruslan Yemtsov. 2014. Income Support for the Poorest: A Review of Experience in Eastern Europe and Central Asia. Washington, DC: World Bank. Trivelli, Carolina. 2014. “Articular, única opción para movernos de programas sociales aislados a una estrategia de desarrollo e inclusión social que enfrente la pobreza.” World Bank, Washington, DC. 120 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN APPEN DIX TABLE A.1 Household surveys used in the analysis Country Name of survey Acronym Closest year Closest year Coverage to 2004 to 2012 Argentina Encuesta Permamente de EPH-C 2004 2012 Urban (31 cities) Hogares Continua Bolivia Encuesta Permamente de ECH 2004 2012 National Hogares-MECOVI Brazil Pesquisa Nacional por PAND 2004 2012 National Amostra de Domicilios Chile Encuesta de Caracterizacion CASEN 2003 2011 National Socioeconomica Nacional Colombia Gran Encuesta Integrada de GEIH 2004 2012 National Hogares Costa Rica Encuesta de Hogares de EHPM 2003 2009 National Propositos Multiples Dominican Encuesta Nacional de Fuerza ENFT 2004 2012 National Republic de Trabajo Ecuador Encuesta de Empleo, ENEMDU 2004 2012 National Desempleo y Subempleo El Salvador Encuesta de Hogares de EHPM 2004 2012 National Propositos Multiples Guatemala Encuesta Nacional de ENCOVI 2004 2011 National Condiciones de Vida (continued on next page) 121 TABLE A.1 Household surveys used in the analysis (continued) Country Name of survey Acronym Closest year Closest year Coverage to 2004 to 2012 Honduras Encuesta Permanente de EHPM 2004 2011 National Hogares de Propositos Multiples Mexico Encuesta Nacional de ENIGH 2004 2012 National Ingresos y Gastos de los Hogares Nicaragua Encuesta Nacional de EMNV 2001 2009 National Hogares sobre Medicion de Niveles de Vida Panama Encuesta de Hogares de EH 2007 2012 National Propositos Multiples Paraguay Encuesta Permanente de EPH 2003 2011 National Hogares Peru Encuesta Nacional de ENAHO 2004 2012 National Hogares sobre Medicion de Niveles de Vida Uruguay Encuesta Continua de ECH 2004 2012 Urban Hogares (Montevideo and areas in the interior with at least 5,000 inhabitants) 122 LEFT BEHIND: CHRONIC POVERTY IN LATIN AMERICA AND THE CARIBBEAN TABLE A.2 Movement into and out of poverty in Latin America and the Caribbean, by country, 2004–12 (population shares in 2012) Poverty status Round Nonpoor in first Poor in first Poor in both round, poor round, nonpoor Nonpoor in Country rounds in second in second both rounds 1 2 Argentina 9.0 3.3 28.1 59.6 2004 2012 Bolivia 22.5 14.1 27.1 36.3 2004 2012 Brazil 19.6 1.9 26.5 52.0 2004 2012 Chile 10.7 2.0 17.6 69.7 2003 2011 Colombia 32.4 2.0 22.6 43.0 2004 2012 Costa Rica 12.5 7.7 16.9 62.9 2003 2009 Dominican 25.6 8.6 28.3 37.5 2004 2012 Republic Ecuador 28.2 1.2 25.2 45.4 2004 2012 El Salvador 25.3 10.4 34.8 29.5 2004 2012 Guatemala 50.0 10.9 12.9 26.2 2000 2011 Honduras 41.5 15.0 24.9 18.6 2004 2011 Mexico 19.7 6.5 17.0 56.8 2004 2012 Nicaragua 36.9 15.7 25.3 22.1 2001 2009 Panama 19.5 2.1 11.7 66.7 2007 2012 Paraguay 17.9 13.4 28.4 40.3 2003 2011 Peru 24.4 0.8 25.0 49.8 2004 2012 Uruguay 7.8 0.2 29.5 62.5 2004 2012 Averagea 21.6 4.2 23.4 51.0 2004 2012 Source: Calculations based on SEDLAC data (CEDLAS and the World Bank). Note: Poverty is defined as per capita income of less than $4 a day in 2005 purchasing power parity dollars. a. Population-weighted average of country estimates. APPENDIX 123 I N DE X Boxes, figures, tables, maps, and notes are indicated by b, f, t, m, and n following the page number. A strategies for improving, 20, 112, 115–116 upward mobility and, 16, 38, 89 academics. See education asset-based approach to chronic poverty, 35 access to services, 12, 14f, 76–78, 77–79f, 78b accountability of institutional processes, 15, 81 Acemoglu, Daron, 81b B Actions for Loving Brazil, 107b Bane, Mary Jo, 41 active social assistance approach, 23, 117 banking, xviii, 39–40b adolescents and children Barrett, Christopher B., 35 cognitive development skills in, 72–73, 75f Barriga, Oscar C., 85 educational attainment aspirations of, 17, Beuermann, Diether W., 78b 18, 18f, 92–94, 93f Bierbaum, Mira, 45 Alkire, Sabina, 43, 55b Blackwell, Lisa S., 116 Appadurai, Arjun, 17, 89, 91 Blinder, Alan S., 86 Argentina Bolivia chronic poverty in, 6, 6f, 53, 54f expectations and optimism for future in, expectations and optimism for future in, 91, 92f 91, 92f telephone interventions in, 115 social assistance programs in, 101, 102f urban vs. rural poverty in, 60, 61f aspirations Bolsa Familia (Brazil), 106b, 107b in capabilities approach, 36–37 bounded rationality, 88 cultural influences on, 89, 91 brain development, 37 decision making and, 89 Brazil defined, 4, 36 access to services in, 107b developmental factors influencing, 37 cash transfer programs in, 106b educational, 17, 18, 18–19f, 92–94, 93f coordination of poverty-reduction efforts expectations and optimism for future, in, 106, 106–107b 16–17, 89, 90–92f educational attainment of children in, 73, 76f intergenerational transmission of poverty endowments and income differences in, and, 92 86f, 87 social norm and peer influences on, 17, 18f, expectations and optimism for future in, 38, 91, 92–94, 93f 91, 92f 125 geographical concentration of chronic employment sectors and, 11, 12, 67–70, poverty in, 6, 7f, 57, 57f 68–69f social assistance programs in, 106, 106–107b ethnicity and, 72, 74b, 74f transmission of positive messages through framework for study of, 3–4, 3f, 33–38, soap operas in, 115–116 39–40b urban vs. rural poverty in, 8, 9f, 60, 61f geographical concentrations of, 6–7, 7–8f, Brazil without Misery (BSM), 106, 106–107b 57–59, 57f, 58m, 59f bribes, 82 gross domestic product and, 62, 63f BSM (Brazil without Misery), 106, 106–107b income-earning opportunities and, 11–17, budgetary considerations, 103, 106, 110, 117 11f, 13t, 14–18f bureaucratic capacity of institutions, 15, 81–82 meta-dimensions of, 33 business grant programs, 20, 111–112, 115 policies for reduction of. See policy design prevalence of, xvii, 5–6, 5t, 6f, 53–55, 54f, 54t C prospects for escape from, 2, 28–29 Calvo-González, Oscar, 15, 81–82 risk factors for, xvii Camacho, Adriana, 23, 118 time dimension of, 33, 40 capabilities approach to chronic poverty, 3, 35, in urban vs. rural areas, 7–8, 9f, 60–62, 36–37 61f, 62t Caribbean. See Latin America and the civil servants, 14, 23, 80 Caribbean Clark, David A., 33 Carneiro, Pedro Manuel, 111 clientelism, 81b Carter, Michael R., 35, 36 cognitive development skills, 72–73, 75f cash transfer programs, xviii, 20, 101, 104, Colombia 106b, 114–115 cash transfer programs in, xviii, 20, Castaneda, A., 43, 55b 114–115 Center for Distributional Labor and Social evolution of social contract in, 81, 81b Studies (CEDLAS), 47 poverty reduction in, 81b centralization, political, 81b social assistance programs in, 21, 101, Chemin, Matthieu, 88b 102f, 118–119 children and adolescents women in labor force in, 67 cognitive development skills in, 72–73, 75f communications, mobile, 12, 14f, 76, 77–78f, educational attainment aspirations of, 17, 78b, 100 18, 18f, 92–94, 93f components approach to measurement of Chile poverty, 40–41, 42 chronic poverty in, 6, 6f, 53, 54f coping mechanisms, 16, 83 social assistance programs in, 15, 23–24, Cord, L., 35 82, 105, 111, 118–119 Corredor Puno Cusco (financial literacy urban vs. rural poverty in, 60, 61f program), 39–40b validation of nonparametric corruption, 82 technique in, 45 cortisol, 88b Chile Solidario program, 15, 23, 24, 82, 118 crime, impact on poverty persistence, 85, 85b Chong, Alberto, 116 cross-sectional data sets, 5, 42, 43, 44–45, 53 chronic component of income, 41 Cruces, Guillermo, 45, 48n3 chronic poverty. See also intergenerational transmission of poverty; measurement of chronic poverty; social assistance D programs Dang, Hai-Anh, 5, 43, 44, 45, 46, 48n3 asset-based approach to, 35 Davalos, Maria E., 45 capabilities approach to, 3, 35, 36–37 decision making definitions of, 33, 34, 41 aspirations and, 89 economic growth and, 8, 10, 10f bounded rationality in, 88 126 INDEX social influences on, 17, 94 constraints on accumulation and use of, 3, 35 state of mind and, 3–4, 36, 88 context and, 3, 12, 35, 86–87, 87f in upward mobility process, 4, 37–38 educational. See education de Laat, Joost, 88b ethnicity, 72, 74b, 74f depression, 16, 20, 88b, 111, 112–113, 113f income differences and, 86–87, 86f disabling context, 75–85. See also enabling income-earning opportunities and, 12–16, context 13t, 14–15f access to services in, 76–78, 77–79f, 78b of nonpoor vs. chronically poor institutions in, 80–82, 81b populations, 72, 73t overview, 75 policy design and, 87 remoteness and, 78–79, 80t strategies for improving, 99–101 size of districts in, 79–80, 80t in upward mobility process, 4, 38 uninsured risk and, 82–85, 83–84f, 85b environmental considerations, 20, 100 downward mobility, geographical variations in, equality of opportunity, 34 6, 6f, 54f, 55 Escobal, Javier, 94n3 drought, 88b ethnicity and chronic poverty, 72, 74b, 74f Duckworth, Angela L., 37, 116 extractive political institutions, 81b Duflo, Esther, 4, 37 Duryea, Suzanne, 116 Dweck, Carol, 37, 116 F Faustian bargains, 34 females. See women E Ferreira, Francisco H. G., 24n2, 31n2 economic security, 24n2, 31n2, 34 FGT (Foster-Greer-Thorbecke) class, 42 economic shocks, 85 financial literacy programs, 39–40b education fixed mind-set, 37 aspirations for, 17, 18, 18–19f, 92–94, 93f food shortages, 83, 84f chronic poverty impacting levels of, 73, Foster, James, 41, 42, 43, 55b 75, 76f Foster-Greer-Thorbecke (FGT) class, 42 intergenerational transmission of poverty and, 18, 19f, 75, 76f social assistance programs for, 108b G in urban vs. rural areas, 60, 62t Galasso, Emanuela, 15, 82, 111 Elbers, Chris, 47 game theory, 89 electricity services, access to, 12, 14f, 76, 77f, 79f Gassmann, Franziska, 45 elite capture, 81 GDP. See gross domestic product Ellwood, David T., 41 gender equality, 11, 66. See also women El Salvador Gini coefficient, 1, 27 endowments and income differences in, Ginja, Rita, 111 86, 86f gross domestic product (GDP) social assistance programs in, 21, 101, 102f correlation with chronic poverty, 62, 63f employment. See labor and labor force growth of, 1, 2, 27, 28–29, 28f enabling context. See also disabling context growth mind-set, 37 in chronic poverty framework, 3, 3f, 35–36 Guatemala endowments and, 3, 12, 35 access to services in, 77–78, 79f income-earning opportunities and, 12–13 chronic poverty in, 6, 6f, 8, 16, 53, 54f institutions and, 36 economic growth in, 8, 63 strategies for improving, 99–101 expectations and optimism for future in, in upward mobility process, 4, 38 16–17, 91, 92f endowments indigenous populations in, 74b, 74f in chronic poverty framework, 3, 3f, 35 urban vs. rural poverty in, 60, 61f cognitive development skills, 72–73, 75f weather shocks in, 83, 84f INDEX 127 H civil servants in, 14, 23, 80 corruption in, 82 Haushofer, Johannes, 88b in disabling context, 80–82, 81b health service programs, 104 enabling context and, 36 Honduras extractive political, 81b chronic poverty in, 6, 6f, 16, 53, 57 leadership in, 13–14 downward mobility in, 54f, 55 social contract as foundation for, 14–15, expectations and optimism for future in, 80–81, 81b 16–17 transparency in, 15, 81 hopelessness, feelings of, 4, 37 upward mobility facilitated by, 36 household allowance schemes, 111 intergenerational transmission of poverty household assets, 24n3 aspirations and, 92 Hoy, Michael, 42 educational outcomes and, 18, 19f, 75, 76f Hulme, David, 33, 34 equality of opportunity in, 34 human capital, 34, 83, 89 policy design for, 21, 37, 100 human rights, 81b state of mind and, 18, 36 Hunt, Jennifer, 82 investment in human capital, 34 I productivity of, 38 in social assistance programs, 2, 29 Ikegami, Munenobu, 36 ISIAT (Innovative Socioeconomic income Interventions against Tuberculosis), components of, 41 112–113 earning opportunities, 11–18, 11f, 13t, 14–18f, 28, 100–101 economic growth and, 8, 10, 10f, 63–65, J 64–65f Jalan, Jyotsna, 33, 40, 41, 42 endowments and differences in, jobs. See labor and labor force 86–87, 86f nonlabor, 11, 69f, 70 policy design and, 100–101 K in poor vs. nonpoor groups, 11, 11f, 66, Kahneman, Daniel, 88 66f Kaufmann, Daniel, 82 reduction in inequality, xvii, 1, 27, Kyrgyz Republic, validation of nonparametric 28f, 101 technique in, 45 social interactions and, 115, 116f strategies for promoting growth of, 21, 104 in urban vs. rural areas, 60, 62t L indigenous populations, 74b, 74f labor and labor force Indonesia, validation of nonparametric employment sectors related to chronic technique in, 45 poverty, 11, 12, 67–70, 68–69f Innovative Socioeconomic Interventions outlook for, 2, 29 against Tuberculosis (ISIAT), 112–113 women in, 11–12, 28, 66–67, 67f inputs. See also enabling context; endowments; LAC. See Latin America and the Caribbean state of mind La Ferrara, Eliana, 116 in chronic poverty framework, 3–4, 3f, Lanjouw, Jean O., 47 35–37 Lanjouw, Peter, 5, 43, 44, 45, 46, 47 in upward mobility process, 3f, 4, 16, LAPOP (Latin American Public Opinion 37–38, 39f Project), 89 institutions Laszlo, Sonia, 82 accountability of, 15, 81 Latin America and the Caribbean (LAC). bureaucratic capacity of, 15, 81–82 See also specific countries 128 INDEX access to services in, 12, 14f, 76–78, Meyer, Moritz, 45 77–79f, 78b middle class economic growth in, 1, 2, 8, 10, 10f, 27, economic classification of, 24n2, 31n2 62–63 emergence of, xvii, 1, 2f, 27, 29f, 101 employment in. See labor and labor force Ministry of Development and Social Inclusion gross domestic product in, 1, 2, 27, (Peru), 106, 109b 28–29, 28f Mitra, Shabana, 43 income in. See income mobile communications, access to, 12, 14f, 76, living conditions in, 101 77–78f, 78b, 100 poverty in. See poverty monetary vs. nonmonetary measures of poverty, rural areas of. See rural areas 55–56b, 56f urban areas of. See urban areas Montoriol-Garriga, Judit, 82 Latin American Public Opinion Project Multidimensional Poverty Index (MPI), 43, (LAPOP), 89 55b, 56f leadership, institutional, 13–14 Lipton, Michael, 34 Loayza, Norman V., 15, 81–82 N longitudinal data sets, 5, 30 National Crusade against Hunger (Mexico), López-Calva, L. F., 35, 43 106, 107–108b natural disasters, 16, 82, 85 Nicaragua M business grant programs in, 20, malnutrition, 21, 74b, 100, 104, 108b 111–112, 115 Martinez, Arturo, Jr., 45 chronic poverty in, 6, 6f, 16, 53, 54f Martínez-Cruz, A. L., 85b downward mobility in, 54f, 55 McKelvey, Christopher, 78b expectations and optimism for future in, measurement of chronic poverty 16–17 census data in, 47, 48 income growth in, 64–65f, 65 components approach to, 40–41, 42 TVIP scores in, 72, 73, 75f cross-sectional data sets in, 5, 42, 43, validation of nonparametric 44–45, 53 technique in, 45 empirical challenges in, 40–48 weather shocks in, 83, 85 household surveys in, 47, 48, 121–122t nonlabor income, 11, 69f, 70 literature review, 40–43 nonmonetary vs. monetary measures of poverty, longitudinal data, lack of, 5, 30 55–56b, 56f mapping techniques for, 46, 47–48, nonparametric approach, 44, 45 57–59, 58m monetary vs. nonmonetary measures, 55–56b, 56f O pseudo-panels in, 42 Ortiz-Juarez, E., 43 spells approach to, 40, 41–42 synthetic panels in, 5, 31, 43–47, 55b meta-dimensions of chronic poverty, 33 P Mexico Panama coordination of poverty-reduction efforts chronic poverty in, 8 in, 106, 106–107b economic growth in, 8, 63 expectations and optimism for future in, urban vs. rural poverty in, 60, 61f 91, 92f Paraguay, income growth in, 64–65f, 65 geographical concentration of chronic parametric approach, 44–45, 46–47 poverty in, 7, 8f, 59, 59f passive social assistance approach, 23, 117 social assistance programs in, 106, peer influences on aspirations, 17, 18f, 38, 91, 107–108b 92–94, 93f INDEX 129 pensions, 28 perpetuation of, 4, 71 Peru reduction of, 1, 2f, 27, 29f, 62 banking for women in, xviii, 39–40b state of mind and emergence from, 87–94, coordination of poverty-reduction efforts 88b, 90–93f in, 106, 109b stress and, 88b education level and aspirations of children ultra-poverty, 34 in, 17, 18, 75, 76f, 92–94, 93f upward mobility process for emergence geographical concentration of chronic from, 3f, 4 poverty in, 57–59, 58m, 94n3 poverty lines, 24n1, 31n1, 42 income distribution in, 63, 64f poverty mapping techniques, 46, 47–48, Ministry of Development and Social 57–59, 58m Inclusion in, 106, 109b poverty traps, 3, 16, 34, 35, 83 mobile phone expansion in, 78b, 78f Premand, Patrick, 83 natural disaster shocks in, 85 present biases, 36 poverty estimates in, 45, 46t pseudo-panels, 42 remoteness of districts in, 79, 80t social assistance programs in, 15, 81–82, 103b, 103f, 105 R telephone interventions in, 115 Ravallion, Martin, 33, 40, 41, 42 tuberculosis prevention and treatment Ray, Debraj, 17, 89, 91 program, 20, 112–113, 113–114f Recanatini, Francesca, 82 TVIP scores in, 72, 73, 75f Red Unidos program (Colombia), 23, 118 validation of nonparametric remoteness of districts, 20, 78–79, 80t technique in, 45 rich, economic classification of, 24n2, 31n2 Philippines, validation of nonparametric Rigolini, Jamele, 15, 81–82, 85 technique in, 45 road services, access to, 20, 35, 78, 79f, 100 pluralism, 81b Robinson, J., 80–81, 81b policy design, 99–119. See also social assistance Rodríguez-Castelán, C., 85b programs role models, promotion of social interactions coordination in, 101–111, 102–103f, 103b, with, 20, 115, 116–117 105t, 106–109b rural areas elements supporting elimination of chronic chronic poverty in, 7–8, 9f, 60–62, 61f, 62t poverty, 18–24 education levels in, 60, 62t endowments and, 87 income in, 60, 62t environmental, 100 policy design for, 101 goal development in, 109 social assistance programs in, 21 for income opportunities, 100–101 TVIP scores in, 72, 75f for intergenerational transmission of poverty, 21, 37, 100 overview, 99 S of social intermediation services, 23–24, 82, sanitation facilities, access to, 12, 14f, 76, 117–119 77f, 79f state of mind and, 111–117, Santos, María Emma, 41 113–114f, 116f savings accounts, 39–40b transparency in, 109 Schady, Norbert, 72 in urban vs. rural areas, 101 school. See education political centralization, 81b SEDLAC (Socio-Economic Database for Latin poor, economic classification of, 24n2, 31n2 America and the Caribbean), 47 poverty. See also chronic poverty; Sen, Amartya, 3, 4, 34, 35, 36–37 intergenerational transmission of services, access to, 12, 14f, 76–78, 77–79f, 78b poverty; transitional poverty sewerage systems, access to, 12, 14f, 76, 77f crime and, 85, 85b Shapiro, Jeremy, 88b 130 INDEX Shepherd, Andrew, 34 social norm and peer influences on, 17, 91 soap operas, transmission of positive messages strategies for improving, 20 through, 115–116 stress and, 87–88 social assistance programs upward mobility and, 4, 16, 37–38 bureaucratic capacity of, 15, 81–82 stress, 87–88, 88b, 114 business grant programs, 20, 111–112, 115 structural poverty. See chronic poverty cash transfer, xviii, 20, 101, 104, 106b, Subsidio Familiar (Chile), 111 114–115 synthetic panels, 5, 31, 43–47, 55b coordination of, 21–23, 104–109, 105t, 106–109b defined, 119n1 T for education, 108b telephone interventions, 115 growth of, 21, 22f, 101, 102f Test de Vocabulario en Imagenes Peabody for health services, 104 (TVIP), 72–73, 75f household allowance schemes, 111 time dimension of chronic poverty, 33, 40 investment in, 2, 29 Torero, Máximo, 94n3 passive vs. active approaches to, 23, 117 transient component of income, 41 spending on, 101–103, 102–103f, 103b transitional poverty in urban vs. rural areas, 21 measurement of, 41, 44, 45–46 social contracts, institutional, 14–15, prevalence of, 5, 5t, 53, 54t, 123t 80–81, 81b transparency social interactions of institutional processes, 15, 81 in decision making process, 17, 94 in policy design, 109 income and, 115, 116f Trzesniewski, Kali H., 116 with role models, 20, 115, 116–117 tuberculosis prevention and treatment social intermediation services, 23–24, 82, programs, 20, 112–113, 113–114f 117–119 TVIP (Test de Vocabulario en Imagenes social norms, aspirations and, 17, 91 Peabody), 72–73, 75f Social Protection database (World Bank), 119n1 social workers, 23–24, 82, 117, 118 U Socio-Economic Database for Latin America ultra-poverty, 34 and the Caribbean (SEDLAC), 47 uninsured risk, 16, 82–85, 83–84f, 85b spells approach to measurement of poverty, 40, upward mobility 41–42 aspirations and, 16, 38, 89 state of mind. See also aspirations decision making and, 4, 37–38 in chronic poverty framework, 3–4, 3f, inputs and, 3f, 4, 16, 37–38, 39f 36–37 institutions in facilitation of, 36 in decision making, 3–4, 36, 88 process for, 3f, 4, 35, 37–38, 39f depression and, 16, 20, 88b, 111, urban areas 112–113, 113f chronic poverty in, 7–8, 9f, 60–62, 61f, 62t emergence from poverty and, 87–94, 88b, education levels in, 60, 62t 90–93f income in, 60, 62t fixed vs. growth mind-set, 37 policy design for, 101 income-earning opportunities and, 16–17, social assistance programs in, 21 16–18f TVIP scores in, 72, 75f intergenerational transmission of poverty Uruguay and, 18, 36 chronic poverty in, 6, 6f, 53, 54f, 57 opportunities for engagement and, endowments and income differences in, 111–113 86f, 87 policy design and, 111–117, expectations and optimism for future in, 113–114f, 116f 91, 92f INDEX 131 V income-generation opportunities for, 36 in labor force, 11–12, 28, 66–67, 67f Vakis, Renos, 78b, 83, 85 workforce. See labor and labor force Vietnam, validation of nonparametric World Bank, Social Protection database, 119n1 technique in, 45 World Development Report on Mind, Society, vulnerable class and Behavior (2015), xviii economic classification of, 24n2, 31n2 World Values Surveys, 89 impact on social policy dialogue, 27 Y W Yaqub, Shahin, 40 water services, access to, 12, 14f, 76, 77f, 79f Young Lives Project (Peru), 92 weather shocks, 4, 38, 83, 83–84f youth. See children and adolescents Weber, Max, 81b women banking for, xviii, 39–40b Z gender equality for, 11, 66 Zheng, Buhong, 42 ECO-AUDIT Environmental Benefits Statement The World Bank Group is committed to reducing its environmental footprint. In support of this commitment, the Publishing and Knowledge Division leverages electronic publishing options and print-on-demand technology, which is located in regional hubs worldwide. Together, these initiatives enable print runs to be lowered and shipping distances decreased, resulting in reduced paper consumption, chemical use, greenhouse gas emissions, and waste. The Publishing and Knowledge Division follows the recommended standards for paper use set by the Green Press Initiative. The majority of our books are printed on Forest Stewardship Council (FSC)–certified paper, with nearly all containing 50–100 percent recycled content. The recycled fiber in our book paper is either unbleached or bleached using totally chlorine-free (TCF), processed chlorine-free (PCF), or enhanced elemental chlorine-free (EECF) processes. More information about the Bank’s environmental philosophy can be found at http://www.worldbank.org/corporateresponsibility. 132 INDEX