WORLD BANK TECHNICAL PAPER NO. 518 _~J WTP518 Work in progress December 2001 for public discussion Measurement and Meaning Combining Quantitative and Qualitatiuve .lfethodls for tthe AnaC Asis of Poverty and Social Exclusion in Latin America Edited b, Estunislao (Getriha-.lI,Urio Qluenltitn Ubo1nos Recent World Bank Technical Papers No. 429 Gary McMahon, Jose Luis Evia, Alberto Pasc6-Font, and Jose Miguel Sanchez, An Environmental Study of Artisanal, Small, and Medium Mining in Bolivia, Chile, and Peru No. 430 Maria Dakolias, Court Performance around the World: A Comparative Perspective No. 431 Severin Kodderitzsch, Reforms in Albanian Agriculture: Assessing a Sector in Transition No. 432 Luiz Gabriel Azevedo, Musa Asad, and Larry D. Simpson, Management of Water Resources: Bulk Water Pricing in Brazil No. 433 Malcolm Rowat and Jose Astigarraga, Latin American Insolvency Systems: A Comparative Assessment No. 434 Csaba Csaki and John Nash, eds., Regional and International Trade Policy: Lessonsfor the EU Accession in the Rural Sector-World Bank/FAO Workshop, June 20-23, 1998 No. 435 lain Begg, EU Investment Grants Review No. 436 Roy Prosterman and Tim Hanstad, ed., Legal Irmpediments to Effective Rural Land Relations in Eastern Europe and Central Asia: A Comparative Perspective No. 437 Csaba Csaki, Michel Dabatisse, and Oskar Honisch, Food and Agriculture in the Czech Republic: From a "Velvet" Transition to the Challenges of EU Accession No. 438 George J. Borjas, Economic Research on the Determinants of Immigration: Lessons for the Europeant Union No. 439 Mustapha Nabli, Financial Integration, Vulnerabilities to Crisis, and EU Accession in Five Central European Countries No. 440 Robert Bruce, loannis Kessides, and Lothar Kneifel, Overcorming Obstacles to Liberalization of the Telecom Sector in Estonia, Poland, the Czech Republic, Slovenia, and Hungary: An Overview of Key Policy Concerns and Potential Initiatives to Facilitate the Transition Process No. 441 Bartlomiej Kaminski, Hungary: Foreign Trade Issues in the Context of Accession to the EU No. 442 Bartlomiej Kaminski, The Role of Foreign Direct Investment and Trade Policy in Poland's Accession to the EIuropean Union No. 443 Luc Lecuit, John Elder, Christian Hurtado, Fran,ois Rantrua, Kamal Siblini, and Maurizia Tovo, DeMIStifying MIS: Guidelinesfor Management Information Systems in Social Funds No. 444 Robert F. Townsend, Agricultural Incentives in Sub-Saharan Africa: Policy Challenges No. 445 Ian Hill, Forest Management in Nepal: Economics of Ecology No. 446 Gordon Hughes and Magda Lovei, Economic Reform and Environmental Performance in Transition Economics No. 447 R. Maria Saleth and Ariel Dinar, Evaluating Water Institutions and Water Sector Performance No. 449 Keith Oblitas and J. Raymond Peter in association with Gautam Pingle, Halla M. Qaddumi, and Jayantha Perera, Transferring Irrigation Management to Farmers in Andhra Pradesh, India No. 450 Andres Rigo Sureda and Waleed Haider Malik, eds., Judicial Challenges in the New Millennium: Proceedings of the Second Summit of the Ibero-American Supreme Courts No. 451 World Bank, Privatization of the Power and Natural Gas Industries in Hungary and Kazakhstan No. 452 Lev Freinkman, Daniel Treisman, and Stephen Titov, Subnational Budgeting in Russia: Preempting a Potential Crisis No. 453 Bartlomiej Kaminski and Michelle Riboud, Foreign Investment and Restructuring: The Evidence from Hungary No. 454 Gordon Hughes and Julia Bucknall, Poland: Complying with EU Environmental Legislature No. 455 Dale F. Gray, Assessment of Corporate Sector Value and Vulnerability: Links to Exchange Rate and Financial Crises No. 456 Salman M.A. Salman, ed., Groundwater: Legal and Policy Perspectives: Proceedings of a World Bank Semillar No. 457 Mary Canning, Peter Moock, and Timothy Heleniak, Reforming Education in the Regions of Russia No. 458 John Gray, Kazakhstan: A Review of Farm Restructuring No. 459 Zvi Lerman and Csaba Csaki, Ukraine: Review of Farm Restructuring Experiences No. 460 Gloria La Cava and Rafaella Y. Nanetti, Albania: Filling the Vulnerability Gap No. 461 Ayse Kudat, Stan Peabody, and Caglar Keyder, eds., Social Assessment and Agricultural Reform in Central Asia and Turkey No. 462 T. Rand, J. Haukohl, and U. Marxen, Municipal Solid Waste Incineration: Requirementsfor a Successful Project No. 463 Stephen Foster, John Chilton, Marcus Moench, Franklin Cardy, and Manuel Schiffler, Groundwater in Rural Development: Facing the Challenges of Supply and Resource Sustainability (List continues on the inside back cover) WORLD BANK TECHNICAL PAPER NO. 518 Measurement and Meaning Combining Quantitative and Qualitative Methodsfor the Analysis of Poverty and Social Exclusion in Latin America Edited by Estanislao Gacitua-Mario Quentin Wodon The World Bank Washington, D.C. Copyright © 2001 The International Bank for Reconstruction and Development/THE WORLD BANK 1818 H Street, N.W. Washington, D.C. 20433, U.S.A. All rights reserved Manufactured in the United States of Arnerica First printing December 2001 1 23404030201 Technical Papers are published to communicate the results of the Bank's work to the development com- munity with the least possible delay. The typescript of this paper therefore has not been prepared in accor- dance with the procedures appropriate to formal printed texts, and the World Bank accepts no responsibility for errors. Some sources cited in this paper may be informal documents that are not readily available. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations, or to mem- bers of its Board of Executive Directors or the countries they represent. The World Bank does not guaran- tee the accuracy of the data included in this publication and accepts no responsibility for any consequence of their use. The boundaries, colors, denominations, and other information shown on any map in this vol- ume do not imply on the part of the World Bank Group any judgment on the legal status of any territory or the endorsement or acceptance of such boundaries. The material in this publication is copyrighted. The World Bank encourages dissemination of its work and will normally grant permission promptly. Permission to photocopy items for internal or personal use, for the internal or personal use of specific clients, or for educational classroom use, is granted by the World Bank, provided that the appropriate fee is paid directly to Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, U.S.A., telephone 978-750-8400, fax 978-750-4470. Please contact the Copyright Clearance Center before photocopying items. For permission to reprint individual articles or chapters, please fax your request with complete information to the Republication Department, Copyright Clearance Center, fax 978-750-4470. All other queries on rights and licenses should be addressed to the World Bank at the address above or faxed to 202-522-2422. ISBN: 0-8213-5054-4 ISSN: 0253-7494 Estanislao Gacituia-Mari6 is a Senior Social Scientist in the Environmentally and Socially Sustainable Development Unit of the World Bank's Latin American and the Caribbean Region. Quentin Wodon is a Senior Economist in the Poverty Reduction and Economic Management Division of the World Bank's Latin America and the Caribbean Region. Library of Congress Cataloging-in-Publication has been applied for. iii CONTENTS FIGURES ...............v FOREWORD ..............V ABSTRACT. ...............Vii ACKNOWLEDGMENTS .......Vii; CHAPTER 1. COMBINING QUALITATIVE AND QUANTITATIVE METHODS FOR POLICY RESEARCH ON POVERTY WITHIN A SOCIAL EXCLUSION FRAMEWORK Carine Clert, Estanislao Gacituia-Mari6, and Quentin Wodon ................................................................................ 1 INTRODUCTION .1.......... ARGUMENTS FOR COMBINING QUANTITATIVE AND QUALITATIVE METHODS ........................ ...................................2 ARGUMENTS FOR USING A SOCIAL EXCLUSION FRAMEWORK ................................................................................. 3 QUANTITATIVE AND QUALITATIVE METHODS IN A SOCIAL EXCLUSION FRAMEWORK ............... ................................4 PRESENTATION OF THE CASE STUDIES .................................................................................5 CONCLUSION .................................................................................7 REFERENCES ....... .........................................................................9 CHAPTER 2. REPRODUCTIVE HEALTH IN ARGENTINA'S POOR RURAL AREA Estanislao Gacitzua-Mari6, Corinne Siaens, and Quentin Wodon ............................................................................... 11 INTRODUCTION ................................................................................ I I REPRODUCTIVE HEALTH IN RURAL ARGENTINA: BASIC STATISTICS . ....................................................................... 13 Statistics for the survey sample ................................................................................ 13 Comparing the survey sample to other data sets ................................................................................ 19 QUANTITATIVE ANALYSIS: CONTRACEPTION, DELIVERIES, AND WORK PATTERNS .................................................. 21 Impact of contraception on the probability of a delivery ................................................................................ 21 Impact of a recent delivery on work ................................................................................ 23 QUALITATIVE ANALYSIS: OBSTACLES TO CONTRACEPTION AND REPRODUCTIVE HEALTH ............ .......................... 26 Conceptualframework ................................................................................ 26 Reproductive health awareness ................................................................................ 30 Family planning ................................................................................ 30 Prenatal care and abortions. .................... ............................................................ 31 Sexuality, sexual abuse, and domestic violence ................................................................................ 32 Social Networking ................................................................................ 33 Health care and coverage ................................................................................ 34 CONCLUSION ........ ........................................................................ 35 REFERENCES ........ ........................................................................ 38 CHAPTER 3. THE TARGETING OF GOVERNMENT PROGRAMS IN CHILE Carine Clert, and Quentin Wodon ................................................................................ 43 INTRODUCTION ................................................................................ 43 BACKGROUND ................................................................................ 44 The Ficha CAS ................................................................................ 44 The Targeted Programs: Their Role in Fostering Security and Alleviating Poverty ......................................... 45 QUANTITATIVE EVALUATION ................................................................................ 47 QUALITATIVE EVALUATION: AN ACTOR-ORIENTED APPROACH ............................................................................. 52 The Viewfrom the Households ................................................................................ 52 The View From Social Workers and Welfare Assistants ................................................................................ 57 The Changes introduced in September 1999: Achievements and Persisting challenges ........... ....................... 58 The impact of additional eligibility criteria: Housing and child care ............................................................... 59 CONCLUSION AND POLICY IMPLICATIONS ................................................................................ 61 REFERENCES ................................................................................ 63 APPENDIX ONE: DESCRIPTION OF SOME OF THE MAIN PROGRAMS TARGETED ACCORDING TO THE FICHA CAS ....... 64 APPENDIX Two: CAS FORM AT THE TIME OF THE QUALITATIVE FIELDWORK (DECEMBER 1997) - ENCUESTA CAS. 66 iv CHAPTER 4. SOCIAL EXCLUSION IN URBAN URUGUAY Judy L. Baker ............................................................................. 69 INTRODUCTION .............................................................................. 69 QUANTITATIVE ANALYSIS .............................................................................. 70 Methodology ............................................................................. 70 Geographic Polarization ............................................................................ 72 Characteristics of Marginal Neighborhoods ............................................................................ 74 Vulnerable Groups ............................................................................ 76 QUALITATIVE ANALYSIS .............................................................................. 79 CONCLUSIONS .............................................................................. 86 REFERENCES .............................................................................. 88 BOXES Box 2.1: QUANTITATIVE ANALYSIS: SURVEY, BASIC STATISTICS, AND ECONOMETRIC MODELS .................. .1 4 Box 2.2: QUALITATIVE ANALYSIS-METHODOLOGY .............................................................................. 28 Box 3.1: METHODOLOGY FOR THE QUANTITATIVE EVALUATION .............................................................................. 50 Box 3.2: METHODOLOGY FOR THE QUALITATIVE EVALUATION .............................................................................. 53 Box 4.1: METHODOLOGY FOR THE QUANTITATIVE ANALYSIS: POLARIZATION ANALYSIS ...................... . 71 Box 4.2: METHODOLOGY FOR THE QUALITATIVE ANALYSIS .............................................................................. 81 FIGURES FIGURE 2.1 DIMENSIONS OF WOMEN'S ACTIVITIES AND ROLES .............................................................................. 27 FIGURE 3.1: SOURCES OF INFORMATION ON INCOME TRANSFERS AMONG POOR HOUSEHOLDS . . 55 TABLES TABLE 2.1: REPRODUCTIVE HEALTH INDICATORS FOR POOR RURAL AND URBAN WOMEN IN ARGENTINA ....... 1 9 TABLE 2.2: RURAL REPRODUCTIVE HEALTH INDICATORS FOR BOTTOM TWO QUINTILES, SELECTED COUNTRIES ... 20 TABLE 2.3: IMPACT OF FAMILY PLANNING ON DELIVERY IN LAST 3 YEARS, RURAL ARGENTINA, 2000 .. 22 TABLE 2.4: PROBABILITY OF DELIVERY CONDITIONAL TO THE USE OF MODERN CONTRACEPTION .. 23 TABLE 2.5: IMPACT OF DELIVERY ON PROBABILITY OF WORKING AND NUMBER OF HOURS WORKED .. 25 TABLE 3.1: GINI INCOME ELASTICITY OF SOCIAL PROGRAMS TARGETED ACCORDING TO THE FICHA CAS .. 49 TABLE 4.1: CHARACTERISTICS OF HIGH AND Low RISK URBAN NEIGHBORHOODS . ................................................... 72 TABLE 4.2: SPATIAL CHANGES IN MONTEVIDEO BY NEIGHBORHOOD CLUSTERS . ...................................................... 74 TABLE 4.3: CHARACTERISTICS OF URBAN VULNERABLE GROUPS .............................................................................. 78 v FOREWORD The World Development Report on Attacking Poverty recommends looking at the multi- dimensionality of poverty, including social disadvantage, vulnerability, and powerlessness. To deal with these issues, the World Bank has enriched its traditional quantitative analysis of poverty with qualitative and participatory research. The Bank's evolution towards broader forms of assessment of disadvantage and the use of qualitative as well as quantitative methods results in part from the lessons of the existing evidence in the development literature. But it is also a response to the challenges faced in its lending operations. Quantitative methods have long been used for project design, for example in improving the targeting and implementation of interventions on the basis of monitoring and evaluation studies. But qualitative methods have also proved essential in identifying key social issues, assessing stakeholder interests and interactions, their likely effect on proposed Bank operation and the potential consequences for individuals and groups. In Poverty Reduction Strategies Papers as well, qualitative research has been essential to complement the traditional focus on quantitative analysis. The need for a richer understanding of poverty and disadvantage through the combination of quantitative and qualitative methods has particular resonance in Latin America. With democracy consolidated in most countries, the challenge of empowerment goes along with that of citizenship and the right of the poor to have their voices heard and faithfully reflected by development researchers. Persistent inequities have contributed to the use of the concepts of exclusion and inclusion in both civil society and government circles, and the region's socio- cultural diversity is an additional reason to systematize the used of mixed research techniques in applied work. Qualitative techniques are specially suited to understand the subjective meanings of poverty; the perceived barriers to escaping it; the political, socio-cultural factors determining it; and the intra- household dimensions and dynamics. The use of quantitative and qualitative techniques under an integrated framework (as the social exclusion framework proposed by the authors in the introductory chapter) is specially helpful to identify premonitory signs of poverty, inequality and marginality and to assess the likelihood that new groups may fall under poverty or the conditions under which it may be reproduced in the future; to understand the processes conducive to poverty and, thus, to better evaluate the effectiveness of anti-poverty programs; and to recognize the most serious poverty and exclusion risk factors present in a given place or territory and thus to better design adequate social safety nets. vi This report presents three case studies drawn from World Bank economic and sectoral work recently completed in Latin America. Each study takes a social exclusion approach and relies on both types of research methods to analyze the factors and processes contributing to poverty and social disadvantage. The quantitative methods include statistical and regression-based analysis. The qualitative methods range from key informant interviews and semi-structured interviews to focus groups. I hope that this report will contribute to promote a broader use of mixed research techniques not only by World Bank staff, but also by other parties interested in development policy and poverty reduction. Guillermo Perry Chief Economist Latin America and the Caribbean Region vii ABSTRACT This report consists of a collection of case studies from Latin America combining qualitative and quantitative research methods for the analysis of poverty within a social exclusion framework. The first chapter provides an overview of the differences between quantitative and qualitative methods, and the gains from using both types of methods in applied work. The other chapters are devoted to three case studies on reproductive health in rural Argentina, the targeting of social programs in Chile, and social exclusion in urban Uruguay. Each case study was prepared within the broader context of country-specific economic and sectoral work at the World Bank. viii ACKNOWLEDGMENTS This report is a joint product of the LCSEO Group, Environmentally and Socially Sustainable Development Unit (ESSD), and the LCSPP (Poverty) Group, Poverty Reduction and Economic Management Unit (PREM), in the Latin America and the Caribbean Region at The World Bank. The work received financial support from the Regional Studies Program in the Office of the Chief Economist (Guillermo Perry) for the Latin America and the Caribbean Region. The case studies were originally written as part of country-specific economic and sectoral work at the World Bank. They were presented in seminars where the authors benefited from valuable feedback. Chapter 2 on reproductive health in poor rural areas was presented at the August 2001 meeting of the Society for rural sociology in Albuquerque, New Mexico, and in discussions with Government officials in Argentina. Chapter 3 on the targeting of Government programs was presented in the World Bank's May 2001 Economists Forum and in discussions with Government officials in Chile. Chapter 4 on social exclusion in poor urban neighborhoods was presented at the October 2000 LACEA conference in Rio de Janeiro, at a January 2001 brown bag seminar in the World Bank, and in discussions with Government officials in Uruguay. The editors are grateful to Shelton Davis, Norman Hicks, Guillermo Perry, and Luis Serven for their continuing support and advise. Anne Pillay provided editorial assistance. Although the World Bank sponsored this work, the opinions expressed by the authors are theirs only, and should not be attributed to the World Bank, its Executive Directors, or the countries they represent. CHAPTER 1 COMBINING QUALITATIVE AND QUANTITATIVE METHODS FOR POLICY RESEARCH ON POVERTY WITHIN A SOCIAL EXCLUSION FRAMEWORK Carine Clert, Estanislao Gacitzia-Mari6 and Quentin Wodon INTRODUCTION There is a wide consensus on the need to look at the multi-dimensionality of poverty, including issues of social disadvantage, vulnerability, and powerlessness. Within the World Bank, this evolution in the analysis of poverty is reflected in the shift of the World Development Reports on poverty from a focus on low-consumption and low achievement in human capital in 1990, to a broader approach dealing with opportunity, security, and empowerment in 2000 (World Bank, 2001). Although vulnerability is not a synonym for poverty, it contributes to it. It implies both an exposure to risk (whether short-term or long running), and a difficulty in coping with these risks (Chambers, 1989). Another important concept is that of social exclusion, which prevents the poor from having access to assets and markets, and from participating (and being represented) in society. Social exclusion is associated with discrimination on the basis of age, ethnic origin, or gender, among other characteristics, and thereby with poverty. Still another influential concept is that of social capital, with its focus on the role of networks and relationships as assets. Many of these new concepts acknowledge that both the questions that we ask and how our knowledge is organized are mediated by pre-existing social structures. These structures need to be accounted for in order to have a better understanding of the research questions themselves and the processes shaping the issues. To deal with these new concepts, along with other international agencies, the World Bank has enriched its traditional quantitative analysis of poverty with qualitative and participatory research. Recent examples include the Voices of the Poor project devoted to the perception of the poor of their own situation (Narayan et al., 2000), and a report done in partnership with the International Movement ATD Fourth World on how to reach the extreme poor and make institutions more sensitive to their needs (Wodon, 2001). The Bank's evolution towards broader forms of assessment of disadvantage and the use of qualitative as well as quantitative methods results in part from the lessons of the existing evidence in the development literature. But it is also a response to the challenges faced in its lending operations. Quantitative methods have long been used for project design, for example in improving the targeting and implementation of interventions on the basis of the findings from monitoring and evaluation studies. But qualitative I The authors are with the World Bank. Comments can be sent to cclert@worldbank.org, Egacituamario@worldbank.org, and qwodon@worldbank.org. 2 Measurement and Meaning research tools have also proved essential in identifying key social issues, assessing stakeholder interests and interactions, their likely effect on proposed Bank operation and the potential consequences for individuals and groups (Clert et al, forthcoming; Gacitua-Mari6 et al. 2000). In Poverty Reduction Strategies Papers (PRSPs) as well, qualitative research and findings have been found essential to complement the traditional focus on quantitative analysis. The trend toward a richer understanding of poverty and disadvantage through the combination of quantitative and qualitative methods has particular resonance in Latin America. With democracy consolidated in most countries, the challenge of empowerment goes along with that of citizenship and the right of the poor to have their voices heard and faithfully reflected by development researchers. Persistent inequities have contributed to the use of the concepts of exclusion and inclusion in both civil society and government circles, and the region's socio- cultural diversity is an additional reason to systematize the used of mixed research techniques in applied work. In order to promote a broader use of mixed research techniques by World Bank staff and other interested parties, this report presents three case studies drawn from World Bank economic and sectoral work recently completed in Latin America. Each study takes a social exclusion approach and relies on both types of research methods to analyze the factors and processes contributing to poverty and social disadvantage The quantitative methods include statistical and regression- based analysis. The qualitative methods range from key informant interviews and semi- structured interviews to focus groups. In this introductory chapter, before presenting the three case studies, we provide some general background regarding the need for combining quantitative and qualitative methods for the analysis of poverty within a social exclusion framework. ARGUMENTS FOR COMBINING QUANTITATIVE AND QUALITATIVE METHODS The arguments for combining quantitative and qualitative research methods are well known (Bourdieu 1992, Neuman 1999) and have been recently summarized in several Bank related publications (Bamberger 2000; Baker 2000; Coudouel, Hentschel and Wodon 2001; Hentschel 1999). Neither approach is better than the other. Both have strengths and weaknesses. There are often gains in combining both approaches and the issue is to choose the most appropriate combination once the problem or research question to be examined has been properly defined (e.g., Bamberger, 2000). Methods based on statistics provide robustness to the results if they rely on appropriate samples, and regression analysis helps to control for a large number of other variables when measuring the impact of a specific variable on a given outcome. Yet quantitative data cannot fully capture causality because of their failure to provide contextual information (Hentschel, 1999). Qualitative methods such as participant observation or community surveys with key informant interviews help to shed light on the economic, socio-cultural and political context of the processes under study. In project evaluations, the combination of quantitative and qualitative research techniques is especially important because "qualitative methods allow the in-depth study of selected issues, cases, or events and can provide critical insights into beneficiaries' perspectives ..., or the reasons behind certain results observed in a quantitative analysis" (Baker, 2000:8). Said differently, qualitative assessments '"provide a better understanding of stakeholders 'perceptions and priorities." Chapter 1. Combining Qualitative and Quantitative Methods. 3 In the specific case of policy research on poverty, inequality, and vulnerability, Coudouel, Hentschel, and Wodon (2001) suggest that qualitative methods serve three purposes: i) help design appropriate household survey questionnaires; ii) assess the validity of survey results at the local level and evaluate how much policy responses should take into account the heterogeneity of local conditions; and iii) gather information that household surveys are not able to capture, or can capture only partially. Regarding this last purpose, the authors refer to assessing dimensions of poverty such as (among others): its subjective meanings; perceived barriers to escaping it; political, socio-cultural factors determining it; and intra-household dimensions. Qualitative methods also help in addressing research questions that cannot be answered easily through quantitative methods. For example, qualitative methods make a unique contribution to the understanding of processes shaped by the subjective perception of the social actors. ARGUMENTS FOR USING A SOCIAL EXCLUSION FRAMEWORK The social exclusion framework (SEF) has been proposed as a heuristic device to understand the linkages and interactions between different risk factors (economic, social, cultural, political and institutional) which generate poverty and inequality (Gacit-ua-Mari6 et al. 2000). The SEF acknowledges that the risks factors are not linked through linear causality but rather in a complex process of reciprocal causation and interactions. Beyond being "goods-centered" (traditional poverty work seeks to improve the command on goods and services of the poor), the SEF is also "people-centered" (following Sen's work on the importance of freedom and capabilities to achieve functioning) and "institutions-centered" (since exclusion is a process rather than a condition at a given point of time, the role of institutions in permitting or creating exclusion must be analyzed). The SEF also contains both an objective and subjective dimension, to the degree that it considers both the objective conditions of people's lives and their perceptions of being connected or disconnected from wider spheres of social, political and cultural life. The SEF is not a substitute for traditional poverty or vulnerability analysis. It fully recognizes the importance of the traditional dimensions of poverty, such as the inability to generate a sufficient and stable income and to have access to quality social services in order to meet basic needs. The advantage of the SEF is that it provides a framework for interrelating different levels of analysis (multidimensionality) and cumulative processes that maintain or pull social groups into social disadvantage. It also incorporates other dimensions that belong to the relational/symbolic domain: the socio-organizational cultural and the political. Social exclusion is a process through which social groups are wholly or partially excluded from full participation in the society in which they live due to the cumulative effect of risk factors. In fact, all social groups are exposed to some risk factors. The problems develop when risk factors start to build up in time and space and the affected group cannot control them. As a consequence, the affected groups are exposed to the cumulative pressure of a multiplicity of risks, one of them being the lack of income, engendering a more or less permanent state of deprivation and destitution. The distinction between social exclusion as a process and poverty as a social condition allows the policy maker to find ways and instruments to fight the risk factors before it is too late, that is, before the processes of social exclusion result in extreme poverty. Policy makers can also take coherent sets of measures taking various dimensions of poverty into account. Furthermore, the SEF leads to social agency, i.e. the process that makes it possible for those who have run into a situation of poverty to come out of it by mobilizing available resources to eliminate or control 4 Measurement and Meaning the risk factors that have engendered it. The central value added by the SEF lies the emphasis on dynamic processes where institutions and agents are involved in the analysis, for example of extreme poverty. This refers to the ways institutions, rules and perceptions interact to generate or combat processes of social disadvantage. Depending on the context or problem being examined, different types of processes may including the possible impact of economic processes on the functioning of the markets; institutional modifications on the social policy system; or cultural practices on the exclusionary or inclusive behavior of certain institutions. In this framework, Government-backed public policies intervene in the regime of social risk by providing resources to those affected. Equally relevant are those regimes that people themselves produce by getting together and getting organized. These are voluntary organizations, neighborhood associations, NGOs, both profitable and non-profit enterprises, neighborhood committees, and so on. As a matter of fact, each of these aggregations arises from the need to reduce a danger or a risk. QUANTITATIVE AND QUALITATIVE METHODS IN A SOCIAL EXCLUSION FRAMEWORK The arguments supporting integrated research methods under a social exclusion framework can be extended further by referring to what scholars call research access. While no hasty conclusions should be made about the advantages of qualitative research techniques (respondents may refuse to be interviewed while they may accept to fill in an anonymous questionnaire), qualitative methods remain best suited to address sensitive issues such as exposure to violence and the psycho-social dimensions associated with it. First, developing a relationship of trust with the "researched" is essential for breaking walls of fear or shame and improve data collection. Second, the possibility to adopt a "non-threatening language" and to adapt the language according to the type of actors under study is also essential for the discovery of knowledge (Buchanan, 1988). For municipal agents at the lower echelons of the hierarchy in the Chilean case study in chapter 3, for instance, the risk of under-reporting was considerable as discussions related to their frustrations with the system of targeting. The language of 'learning from your experience' was adopted and well received. Thirdly and lastly, accessing certain types of interviewees such as elite interviewees, high-level officials or firms may require official and technical steps as well as infornal strategies and opportunities- all of which can hardly be achieved by sending out a questionnaire. A second series of arguments in favor of integrated research methods relates to the potential of using actor-oriented perspectives in poverty research. An actor-oriented perspective entails "recognizing the 'multiple realities' and diverse social practices of various actors, and requires working out methodologically how to get to grips with these different and often incompatible social worlds (Long and Long, 1992:4) ". In poverty research, key actors include not only the poor, but also firms and professionals concerned with poverty reduction at both civil society and government levels. The experience and voice of the poor as citizens is starting to be better captured, and civil society organizations such as NGOs tend to be better included in opinion surveys. But the perspective of government professionals at the different echelons of the policy process still tends to be overlooked, or at least not systematically and rigorously researched. As argued by Clert (2000), this is true for research on social exclusion in developing countries. A contribution of the French literature on social exclusion has been the in-depth interviews with social workers. In Paugam (1993) for example, insertion or inclusion are seen as an operative concept which cannot be fully apprehended without an adequate understanding of the roles, Chapter 1. Combining Qualitative and Quantitative Methods. 5 values and practices of a wide range of actors involved in the policy process, from the politicians in the National Assembly debates to senior officials and social workers in the field. In the British literature as well, this kind of research focus has emerged recently (Barry, 1998). But in developing countries, there is still a lack of focus in the qualitative literature on the perceptions of policy makers. At the World Bank as well, while social assessment and institutional analyses do include the views and experience of public officials, consultations in poverty research and strategies seem to remain mainly concentrated on civil society representatives. This might be a missed opportunity given the interest of the Bank to support increased ownership of clients and better understanding obstacles to articulated strategies poverty reduction and social inclusion. A last series of arguments in favor of integrated research methods in a social exclusion framework relates to policy making. Qualitative data derived from interviews and focus groups is often criticized for its subjectivity. This is a legitimate concern, and it underscores the fact that qualitative research methods must be implemented rigorously by well trained researchers, with their results ideally supported by further quantitative analysis. But at the same time, policy oriented social analysis is concerned with change and agency - i.e., how program beneficiaries, the social workers in the field, and the policy makers can act outside and sometimes against a system which makes the poor vulnerable and disadvantaged. In this context, the question of the subjectivity of the actors, and how as persons they perceive their situations of deprivation and/or the limitations of their interventions is key to understanding the basis of agency. In conclusion, the use of integrated methods under a social exclusion approach has four main advantages: (i) it contributes to the empowerment of the social actors by incorporating the knowledge of the subjects and it enhances the transparency, accountability, and thereby effectiveness of policy interventions; (ii) it helps to identify premonitory signs of poverty, inequality and marginality and to assess the likelihood that new groups may fall under poverty or the conditions under which it may be reproduced in the future; (iii) it enables decision-makers to set priorities by identifying the processes (and different factors) conducive to poverty and, thus, to evaluate the relevance, and effectiveness of anti-poverty programs and; (iv) it allows to profile the most serious risk factors of social exclusion present in a given place or territory. This last point should be emphasized because many social exclusion phenomena depend on the way in which the territory is physically, economically and socially organized. Therefore, the SEF makes it easier to carry out the analysis of the degree of exposure to risk of specific social groups. PRESENTATION OF THE CASE STUDIES The next three chapters provide case studies from World Bank economic and sectoral work combining qualitative and quantitative research methods for the analysis of poverty and social exclusion in Latin America. The case studies are devoted to reproductive health in rural Argentina (chapter 2), the targeting of social programs in Chile (chapter 3), and social exclusion in urban Uruguay (chapter 4). There are variations between the case studies in terms of their objectives and context justifying the use of mixed quantitative and qualitative research methods. In chapter 2, statistical and econometric methods are combined with focus groups of men and women so as to provide a better understanding of reproductive health issues for poor families living in rural Argentina. The quantitative methods are used to describe the families and their use 6 Measurement and Meaning of family planning devises, and to analyze the impact of contraception on the probability of delivery as well as the impact of delivery on work patterns. The qualitative methods provide finer information on women's and men's attitudes on a broad range of sensitive issues related to their respective productive and reproductive roles, and the interaction between these roles. In chapter 3, the authors use mixed research methods in order to evaluate the targeting performance of government social programs in Chile, i.e. to assess the ability of social safety nets to reach the poor and vulnerable. Using data from a nationally representative survey, the quantitative methods are used to measure the targeting performance of various programs relying on the same means-testing instrument (the so-calledficha CAS). The qualitative methods elicit insights into the fairness and effectiveness of the targeting methods, as experienced and perceived by poor citizens and practitioners using those methods in their daily work. In chapter 4, the author uses both types of methods so as to capture the links between exclusion and poverty. Initially prepared for the poverty assessment for Uruguay in a context of high inequalities and rising area-based marginalization, the study complements quantitative analysis with interviews and focus-groups so as to obtain a better understanding of key vulnerable groups exposed to processes of exclusion. Quantitative methods provide the basis for the selection of the areas where the qualitative study is done and for the identification of specific groups of interest. The qualitative analysis explores the dimensions of exclusion more in-depth and collects information on existing perceptions of exclusion in various settings, including labor markets. Many of the arguments presented above as to the gains from combining alternative methods in policy research can be illustrated with the case studies. Without going too much in details, a few illustrations can be provided. In chapter 2 on reproductive health, the authors stress the importance of the cultural context and intra-household power relations. Qualitative findings guard against hasty conclusions derived from the regressions. For example, while the regression results suggest that economic status does not affect the probability of using contraception, the focus groups indicate that cost is a major constraint for access to pill and intra-uterine devices. Similar nuances between quantitative and qualitative findings are found in chapter 3 on Chile. While the results from the quantitative analysis suggest fairly good targeting overall, the qualitative analysis points to the fact that the targeting system does not handle vulnerability well. The value of qualitative methods as a tool for actor-oriented policy research is also illustrated in chapter 3. In that chapter, the actor-oriented approach allows for assessing possible deficiencies in the fairness and effectiveness of means-testing instruments, as seen by social workers. In all chapters, the qualitative work enables the authors to reach a finer understanding of the issues involved. In Argentina, the statistical analysis touches on forced sexual relations as a cause for pregnancies. But focus groups suggest that the phenomena is under-reported, and they shed light on role of incest in early pregnancies. Similarly, the women's perception of the quality of the health service is partly assessed by the statistical survey through a ranking question. Results from focus groups provide additional information on mistreatment. This shows the need to respect women's opportunity to express in their own words how they feel about the services. Chapter 1. Combining Qualitative and Quantitative Methods. 7 CONCLUSION Social and behavioral sciences have used many different experimental and quasi-experimental methods to approach their research questions, leading to what Ritzer (1975) calls a multi- paradigmatic discipline. The problem with most methods taken individually, including formal methods in mainstream economics, is that they take for granted the context and relationships that constitutes the phenomena under study. According to Bourdieu (1992), the objectivist point of view reifies the structures it constructs by treating them as autonomous entities and slipping from model to reality. At the other end, the subjectivist or constructivist point of view asserts that social reality is an ongoing process that social actors continually reconstruct, failing to see the existing regularities. A key challenge for policy is to analyze the objective conditions of reality while identifying how perceptions influence reality. For this, a flaw of most approaches to knowledge generation is methodological monism and the opposition between theory and research, between quantitative and qualitative methods. A methodological approach involves a theory on how a research question should be analyzed. A research method is a procedure for collecting, organizing and analyzing data. There are important connections between methodologies and research methods. Research methods mediate how our knowledge of the world is organized. "It is a truism that methods per se mean little unless they are integrated within a theoretical context and are applied to data obtained in an appropriately design study" (Pedhazur, 1982: 3). The issue then is what are the research questions that we ask and why we ask those and not others. Only after having clearly defined the issues under study, can we look at the array of methods available and choose those that fit the problem, taking into consideration that both quantitative and qualitative methods may be complementary or even necessary. While the World Bank used to rely mainly on quantitative methods in the past, it is now agreed that the analysis of socio-cultural and institutional issues requires an assessment of processes and values as well as the perceptions individuals have of them, all of which cannot be adequately addressed by quantitative methods alone. Qualitative research helps to identify likely sources of support or opposition to poverty reduction interventions, and it contributes to the assessment of the operational feasibility of the proposed interventions, their sequencing, and the selection of appropriate social and institutional arrangements to promote social inclusion and empowerment. Especially when it takes into consideration the rules, values, and perceptions of the groups involved, qualitative research ensures that operations are responsive to the needs of intended beneficiaries in all their social and cultural diversity. In very practical terms, mixed methods improve our understanding of poverty and social exclusion. To take again the example of chapter 2 on reproductive health in Argentina, access to, or exclusion from, health care can be captured quantitatively using household surveys. But, the quality of the service provided and the perceived relationship between patient and health workers are better captured by a combination of techniques. Mixed methods also yield insights on access to, or exclusion from, less tangible assets such as social relations and social capital, or the relationship of the poor to of the justice system. Disadvantage is a complex and cumulative process whereby mechanisms of exclusion from tangible and intangible assets interact. One could argue that the very nature of poverty and social disadvantage and their interrelated causes require the use of mixed research methods. If exposure to domestic violence and abuse is in part 8 Measurement and Meaning driven by overcrowding housing conditions, which may also affect the probability of early pregnancies, this has policy implications, in terms of the need for housing policy and programs to be in tune with health policy. Comprehensive analyses of poverty and social disadvantage lead to better assessments, which in turn give rise to different or complementary policy approaches. Of course, while qualitative methods can help to enrich areas which have traditionally been dominated by quantitative research, the reverse is also true: quantitative methods can enrich these areas which have been dominated by qualitative research. Indeed, the absence or difficulty of quantification has been a factor in the still relatively slow take up of concepts such as those of vulnerability, social capital, and social exclusion. Chapter 1. Combining Qualitative and Quantitative Methods. 9 REFERENCES Baker, Judy. 2000. Evaluating the Impact of Development Projects on Poverty. Washington, DC: World Bank. Barry. 1998. "Social Exclusion and Social Work: An Introduction." In Barry and Hallet (eds) Social Exclusion and Social Work. Issues of Theory, Policy and Practice, Lyme Regis: Russel House Publishing. Bamberger, M., editor. 2000. Integrating Quantitative and Qualitative Research in Development Projects, Washington, DC: The World Bank. Bourdieu, P., and L. J. D. Wacquant. 1992. An Invitation to Reflexive Sociology. Chicago: University of Chicago Press. Buchanan, D., D. Boddy, and J. McCalman. 1988. "Getting In, Getting On, Getting Out, and Getting Back." In A Bryman ed., Doing Research in Organizations. London: Routledge. Chambers. 1998. "Vulnerability: How the Poor Cope." IDS Bulletin Vol. 20, N°2: 1-7. Clert, Carine. 2000. "Policy Implications of a Social Exclusion Perspective in Chile: Priorities, Discourse and Methods in Question." Ph.D. thesis, London School of Economics, London. Clert, Carine, A. Dani, Estanislao Gacitfia-Mari6, B. Harris, and D. Marsden. Forthcoming. "Assessing the operational contribution of social assessment: Lessons from task team perspectives." Social Development Papers. World Bank, Washington, D.C. Coudouel, A., J. Hentschel, and Q. Wodon. 2001. "Poverty Measurement and Analysis." In J. Klugman, editor, Poverty Reduction Strategies Source Book, Washington, DC: World Bank. Gacidia-Mari6, Estanislao, and Carlos Sojo. 2000. Exclusi6n Social y Reducci6n de la Pobreza en America Latina y el Caribe. San Jose, Costa Rica: Banco Mundial/FLACSO. Hentschel, J. 1999. "Contextuality and Data Collection methods: A framework and application to health service utilization." Journal of Development Studies 35: 64-94. Long, A., and N. Long, editors. 1992. Battlefields of Knowledge: The Interlocking of Theory and Practice in Social Research and Development. London: Routledge. Narayan, D., with R. Patel, K. Schafft, A. Rademacher, and S. Koch-Schulte. 2000. Voices of the Poor: Can Any-one Hear Us? New York: Oxford University Press. Neuman, W. L. 1999. Social Research Methods: Qualitative and Quantitative Approaches. Chicago: Allyn & Bacon. Paugam, Serge. 1991. La disqualification sociale: essai sur la nouvelle pauvrete. Paris: Presses universitaires de France. 10 Measurement and Meaning Paugam, Serge. 1993. La societe fran,aise et ses pauvres: I'experience du revenu minimum d'insertion. Paris: Presses universitaires de France. Pedhazur, E. J. 1982. Multiple regression in Behavioral Sciences: Explanation and Prediction. Second edition. Forth Worth, Texas: Holt, Rinehart and Winston Inc. Ritzer, G. 1975. Sociology: A Multiple Paradigm Science. Boston: Allyn and Bacon. Wodon, Quentin. 2001. Attacking Extreme Poverty. Learning from the Experience of the International Movement ATD Fourth World. World Bank Technical Paper No 502, Washington, DC: The World Bank. World Bank. 2001. Attacking Poverty. World Development Report 2000/2001. Washington, DC: World Bank. CHAPTER 2 REPRODUCTIVE HEALTH IN ARGENTINA'S POOR RURAL AREAS Estanislao Gacittia-Mari6, Corinne Siaens, and Quentin Wodon2 INTRODUCTION Reproductive health is fundamental for improving welfare, reducing poverty and decreasing gender discrimination (Rodriguez-Garcia and Akther, 2000). Lack of control over the reproductive process increases women dependency in economic, social and political terms. The impact of reproductive health issues on poverty and quality of life is framed by the relations within the household, women's reproductive and productive work, and the income generation strategies adopted by household members (Acosta-Belen and Bose 1990; Fraad, Resnick & Wolff 1989; Kirwood 1986; Leacock 1983; Nash, 1990; Weston and Rofel 1985). To reduce reproductive health issues to a pure physical-health dimension would ignore the fact that gender roles correspond to a certain economic, social and cultural order that helps to ensure the maintenance of social relations and the reproduction of social units. Household labor is both productive (production of goods and services for the market) and reproductive (production of goods and services for household consumption). Productive and reproductive work are two sides from the same coin as all productive work involves and requires a process of reproduction. Work can broadly be understood as labor that produces something of value for other people, which includes household or domestic work required for the reproduction of the workforce and wage-work (Gerstel and Gross, 1987; Hartman, 1987; Mann, 1990). In all societies gender is the primary means of determining the division of labor and responsibilities for their members. Being male or female is a biological fact, but becoming a man or a woman is a cultural and social process (Correia, 1999). Gender analysis is thus crucial because it looks at the social scripts people are exposed to, the type of choices they see as viable and legitimate, and the cost and benefits associated with those choices (Staudt, 1994). Women's domestic work - particularly that related to biological reproduction - increases disproportionately in relation to other adults in the household as the number of children grows. Mothers of large families spend more time pregnant, breast-feeding, and caring for children. Besides the health risks of high fertility for the mother3, empirical evidence suggests that This paper was prepared for the report "Argentina: Rural Reproductive Health" at the World Bank (2001). The collaboration of the Universidad Nacional de Santiago del Estero and the George Washington University Center for International Health are gratefully acknowledged. Comments can be sent to Egacituamariogworldbank.org, csiaens@worldbank.org, and qwodon@worldbank.org. A 1996 report by the Ministry of Health and Social Action describes four characteristics of maternal mortality in Argentina: (i) maternal mortality is reducible in the majority of cases through technology and knowledge already available; (ii) the vulnerability of females during the reproductive period is greater due to factors associated with 12 Measurement and Meaning children with more closely spaced siblings are at greater risk of long-term malnutrition. Frequent, unwanted, or ill-timed pregnancies can also cause emotional hardship to women and their families, and poor maternal health "drains women of their productive energy, jeopardizes their income-earning capacity, and contributes to their poverty" (World Bank, 1999). In addition to the benefits for the current generation, reproductive health improves the health and productivity of the next generation. In contributing to sustainable development through improving equity, quality of life, and economic potential, reproductive health thus confers benefits to the whole society (World Bank, 2000; World Bank, 1999; World Bank, 1998. This paper combines quantitative and qualitative data from primary and secondary sources in order to provide a better understanding of reproductive health issues in rural Argentina. From a methodological point of view, this study analyzes the impact of reproductive health issues on quality of life as the result of economic, socio-cultural and institutional processes that affect women's capacity to control their reproductive life and participation in society. The paper aims at: (a) identifying the knowledge, attitudes, and practices related to reproductive health of poor families in rural areas of three Northern provinces of Argentina; (b) exploring the impact of reproductive health issues on women's quality of life, with special attention to their ability to control their fertility and increase their income generating potential; and (c) providing information and recommendations that would facilitate the discussion of policies and programs aimed at advancing poor rural women in Argentina, their productive capacity and reproductive health status. The second section of the paper provides basic reproductive health statistics obtained from the a survey conducted especially for this study (see box 2.1). The third section, which is devoted to quantitative regression-based techniques, analyzes the impact of family planning on the probability of having a delivery, and the impact of a delivery on the probability of working and the number of hours worked. It is found that lack of family planning increases the probability of a delivery by up to 15 percent (and thereby the probability of being poor since the household resources have to be share among a larger number of household members). However, the evidence of an impact of a delivery on work patterns is weaker. The fourth section provides the results of the qualitative analysis on a wide range of topics related to reproductive health. A conclusion follows. pregnancy and childbirth; (iii) the risk of death during the reproductive period increases with each pregnancy, indicating the importance of family planning and the reduction of fertility rates; and; (iv) maternal mortality is greatest among women of low socio-economic status living in areas with limited access to health services. Chapter2. Reproductive Health in Argentina's Poor Rural Areas 13 REPRODUCTIVE HEALTH IN RURAL ARGENTINA: BASIC STATISTICS Statistics for the survey sample Since early childhood the women are socialized in their expected reproductive and nurturing roles, first by taking care of their siblings, and later of their own family. The women in our sample established their first union early in their lives (90 percent of the women leave their parents households by age 17) and start their reproductive life soon after, much of the times without having a good knowledge of their reproductive system or how to prevent a pregnancy, if desired. Most women suggest that their first pregnancy was not intentional. However, women see pregnancy as a way out of the parental family. In focus groups, the women indicated that the best age for having the first pregnancy is between 18 and 20 years old. However, they recognize that such assertion is based on their current experience, after having had their first pregnancy younger than what they now indicate is the best age. Women with teenage daughters mentioned they have told their daughters to wait. However, social pressure, economic reasons, and an acceptance of the role model provided by the elder generation (of their mothers) are difficult to surmount for most girls. Regarding the use of family planning, while 51 percent of the women declare using some type of contraceptive method, only 25 percent currently use modem family planning methods. Use of family planning is higher among women that participate in social organizations (55.4 percent), and those aged 31 to 36 years old. The family planning methods most commonly used are natural family planning, which includes periodic abstinence, temperature, the rhythm method (50 percent) and the birth control pill (45 percent). Half of the women who declare using some family planning method are using "natural" methods. This reflects a resistance that women face from their partners to use modem methods. Also, it shows that access to modem methods is limited. When women use "natural" methods, they are exposed to the permanent risk of undesired pregnancies because they do not have control over sexual relations. Only nine percent of the women declared that their partners use condoms despite the fact that knowledge of condoms is high, at respectively 92 percent, 63 percent and 55 percent in the provinces of Misiones, Salta, and Santiago del Estero respectively. The difference between Misiones and the other two provinces is probably a result of the AIDS educational program that is being implemented in Misiones. The women state that the men accept certain family planning methods more than others, and they believe that the men prefer the women to use family planning rather than themselves. The men declare that it is important to be relaxed and not to worry about contraception when having sex. Because of these attitudes, both the men and the women feel that contraception is mainly a female responsibility. Chapter2. Reproductive Health in Argentina's Poor Rural Areas 14 Box 2.1: QUANTITATIVE ANALYSIS: SURVEY, BASIC STATISTICS, AND ECONOMETRIC MODELS This paper is based on a combined quantitative and qualitative analysis of a household survey, focus groups and in-depth interviews. For the quantitative analysis, a multi-stage cluster sample4 of 300 rural households (1,973 individuals, 52 percent women, 48 percent men) from districts that have at least 50 percent of households with unsatisfied basic needs (NBI) in three provinces of Northern Argentina (Missiones, Salta, and Santiago del Estero)5 was selected. This box presents selected socio-economic characteristics of the households interviewed in the survey. In this section, we focus on a description of reproductive health statistics obtained from the survey, and a comparison of these basic statistics with similar data available for urban Argentina and the rural areas of other Latin American countries. Basic Statistics for the survey sample Poverty. The survey data do not include good income indicators, but other indicators can be used as proxies. According to an index of unsatisfied basic needs taking into account the physical infrastructure of the household, crowding, and access to services, many of the households are poor. Another useful indicator is mean food expenditure. At US$ 1,454 per year per household (US$ 241 per capita), it is comparable with the food expenditures for the lowest quintile of Argentina's rural population in Salta and Misiones (World Bank 1997). If food expenditures are adjusted to reflect own production and consumption on the farn (the adjustment factors are based on Amadasi and Neiman 1998), they increase by a third, but still remain low. Education. On average, the women have completed 6 years of schooling (7 years for men). While they report an ability to read and write, only one fifth have gone beyond 6Gh grade (primary education) and 4 percent have completed secondary education. Importantly, the women with 3 or less years of education have 4.12 children on average, versus 2.55 children for those with completed secondary education. Work. Women are involved in household work (primary activity for 70.7 percent of the women) and subsistence farrning (secondary activity for 54.1 percent of the women), and to a lesser extent in wage work and other micro-entrepreneurial work. This indicates that while household work remains the main activity declared by the women, their productive role is broader. The degree to which women participate in wage-earning or income-generating activities is indicative of the socioeconomic position of households. Among better off households, women's participation in wage work is lower than that among low income households. Among the 51 women with wage work as primary or secondary activity, 80 percent are from households in the two lowest quintiles of per capita food consumption. Health insurance. Only 25 percent of the households have health insurance coverage, and two thirds of those with health insurance do not use it for reasons such as inability to make the co-payment and lack of and/or inability to pay for transportation. By contrast, in Argentina's urban centers 52.2 percent of pregnant women were covered by private insurance or an Obra Social in 1997 (SIEMPRO/INDEC, 1997) 4The sample is representative only of poor rural households for the three selected provinces. Key parameters (food expenditures and education level by sex, among others) estimated for the sample are consistent with the figures estimated for other large samples of poor rural population in Salta and Misiones. 5In these three provinces, the share of rural population is well above the national average. Two of the provinces concentrate more than half of the total number of small farmers in the country (estimated at 180,000). Farm structures in Argentina are highly unequal, and it is also the case for these provinces. According to the Agricultural Census of 1988, farmers below 50 ha (50 percent) owned less than 2 percent of the land, while large farms, over 5,000ha (1.7 percent) owned almost 50 percent of the land. The differences are even more striking for the three provinces in the study where on average farmers below 50 ha (60 percent) posses less than I percent of the land. Chapter2. Reproductive Health in Argentina's Poor Rural Areas 15 Box 2.1 (Continued) While not all women in our sample were pregnant at the time of the survey, this suggests large disparities in coverage between poor rural women in the sample and the national average. Since most of the women interviewed (and their male partners) were self-employed, they did not have a compulsory social health insurance (Obra Social Obligatoria). Neither do they have voluntary insurance. This is frequent in rural Argentina among small farmers and minifundistas (see Amadasi and Neiman, 1998), particularly among the poorest groups (such as in this case), which cannot afford health insurance. Housing. While most households own their house (72.3 percent), dwellings are made of adobe and brick (64 percent), with compacted soil floors (70 percent) and cardboard, straw, wood or tin sheets ceilings (85 percent). Many households have latrines, but only 10 percent have a toilette inside the home, and 14 percent running water. Less than half of the households (46 percent) are connected to the electric grid. One in two dwellings has either one or two rooms and 80 percent have no more than three rooms. Crowding is severe, with on average three persons per room, and often, one single bedroom is shared by all household members with little or no privacy. Econometric models Impact of contraception on delivery. The determinants of contraception and delivery are analyzed using a bivariate probit model. Using bivariate probits generates efficiency gains in the estimation because the correlation between the error terms of the contraception and delivery regressions is taken into account. It also enables us to compute the probability of having a delivery conditional on using contraception or not. Denoting by D* and C* the latent and unobserved continuous delivery and contraception variables, by D and C their categorical observed counterparts, and by X the vector of independent exogenous variables, the bivariate probit model can be expressed as: D* =/8DX+e£ D =J ifD * > O, D = otherwise C* = fiCX + gc C =I if C * > O, C =O otherwise Ele D I = ElEC = 0 Var[ £ D ] = Var- £ C = I Cov[ £ D,eD£C = P The impact of using contraception on the probability of delivery is computed as the difference in the two conditional probabilities of delivery: AP = P(D=1 I C=0, X) - P(D=1 I C=1, X) Impact of delivery on work patterns. To estimate the impact of family size on the women's income generating potential, we could in principle estimate the loss in earnings per capita or in wages due to delivery, but this is not feasible here due to the low quality of the wage data and the small sample. As an alternative, we estimate the impact of a recent delivery on the probability of working through a standard probit regression, as on the number of hours of work through a standard tobit regression. Source: Authors. 16 Measurement and Meaning The age for women interviewed ranges from 16 to 65 year, with a mean of 32. For their male partners the range is from 21 to 67 years and the mean is 38. About 85 percent of the women are in reproductive age. One fifth are 24 year old or younger, and only 5 percent are above 50 year. While the average age for the first pregnancy is 18 years old (the mode is 17 years old), about 54 percent of the women become pregnant before 18 and 16 percent under 15 years old. Adolescent pregnancies are not an exception but a common fact and, in a way, an expected passage, a way out of parental control. However, incest was mentioned as a cause of very early pregnancies in focus groups as well as in the in-depth interviews. Housing conditions may contribute to incest, because crowding is severe, with on average three persons per room, and often, one single bedroom is shared by all household members with little or no privacy. This issue was identified in focus groups as a key factor in the sexual initiation and the early expulsion of children and youth from the house (similar findings are found in Barone, 2000 and Fogel et al., 1993 Fogel and Pantelides, 1994)6. The oldest woman having a first pregnancy among the group studied was 37 years old. The average age for the last pregnancy was 29 years old. About 40 percent of the women had children after age 30 and 5 percent over age 40 (some up to age 46). The average total number of pregnancies is 4.86 per interviewee. There is a significant negative correlation between educational attainment of women and age of last pregnancy. An important factor for late pregnancies is the establishment of new relationships. Most women feel that after establishing a relationship with a new partner, they have to "give" him a child. The range of pregnancies goes from a minimum of I to a maximum of 14 pregnancies per women. It is likely that through breastfeeding the women manage to prevent a higher number of pregnancies. Also, if they devote an average of two years for child rearing for each offspring, this may again limit the number of children. The women have from 1 to 11 children. As a result of the high fertility rates, the share of population under 14 years old (53 percent) is higher than the national average (30 percent) despite the fact that the share of the population aged 20 to 59 is lower (35.1 percent) than the national average (43.6 percent). This may also suggests out-migration of the economically active population, starting at age 15 to 19. Most women are married or cohabiting (83 percent). However, only 16 percent of the households are female headed. Female headed households have more children on average (4.7) than male- headed households (3.7). The dependency ratio (i.e., number of members aged 14 or less or 65 and older divided by number of adults aged 14 to 64) is 1.41 on average, 1.71 for female-headed households, and 1.35 for male-headed households. One in four women declares that her children are all from the same father. From those with children from more than one man, 89 percent mention two fathers and 10 percent more than two fathers. Among women with children from more than one man, just 8 receive alimony or support from their former partner(s). Most of the interviewees (76 percent) are Catholic, followed by Evangelists (17 percent). Half of the women (53 percent) indicate that the decision to use family planning is made by the couple, and one out of four says that it is the woman who decides. In focus groups, there was 6Early pregnancy may set in motion an intergenerational cycle of ill health and growth failure (World Bank, 1994). Studies conducted in Argentina by Pantelides and Cerrutti (1992) and Pantelides (1995) found higher rates of adolescent pregnancy among girls with less formal education. Other risk factors for early pregnancy included: (i) having had a parent who had his/her first child before the age of 20; (ii) having a sibling who had become an adolescent mother; (iii) having had a parent, particularly the mother, absent from home (see also Correia, 1999). Chapter2. Reproductive Health in Argentina's Poor Rural Areas 17 consensus that women are generally responsible for choosing family planning methods because of the implications pregnancy has on their bodies and lives in general. Men, in general feel that they were not responsible, while at the same time they are not in agreement with their women taking the decision to use forms of family planning other than "natural methods". In spite of the discourse regarding joint decision making, woman must make the decision to use contraceptives because men won't. This is evident in the low utilization of barrier methods such as the condom (10 percent) in which men take responsibility. Women with less than 6 years of education indicate more frequently (34 percent vs.26 percent for women with more than 6 years of education)) that there is no conscious decision to have children. Among Coya (indigenous) women in the province of Salta, regardless of their educational level, there is a significantly larger proportion of women (63 percent) that respond that pregnancy is a decision of the couple (only 18 percent indicate that children just come). The results about the reproductive health behavior of Coya women should however be interpreted with caution because of the small size of the sub-set of Coya women in the sample and to high degree of assimilation of Coya communities in Salta. The prevalence of forced sexual relations among the women interviewed is reported to be about 21 percent, though the actual percentage may be higher as these situations tend to be under- reported. Of the women who reported at least one episode of forced sexual relations, 80 percent had told her partner "no." In cases where women had verbally declined to have sex, only one of five male partners expressed understanding of the situation. Forty-five percent of men proceeded with having sex against the woman's will and 31 percent of men reacted with violence. One in three women declares having been subjected to violence due to the refuisal to have sex. In addition, another 15 percent acknowledge the occurrence of domestic violence. More than half of these women (59 percent) indicate that alcohol is one of the most important factors in domestic violence and abuse. Twenty-two percent also identify men's feelings of anger or jealously as reasons for domestic violence, which may also be associated with alcohol abuse in some instances. When asked whether women should accept abuse from men, 92 percent of the women responded no. Yet the women indicate that frustrations due to lack of economic resources and/or jealously if the man believes that the woman is seeing another man justify a violent reaction. Some of the women even indicate that some types of physical aggression are indication of care and love from the man. One out of four (24.4 percent) women has had a natural or induced abortion7 at least once. The women attributed the occurrence of abortion to multiple causes that range from clinically defined 7 Abortion is a sensitive issue in Argentina due to its legal and moral implications. The law penalizes both health personnel and women with incarceration of up to 4 years for performing an induced abortion. The law stipulates some exceptions, such as life threatening situations endangering the life of the mother and in cases involving rape of a mentally disabled woman. In spite of the legal sanctions against induced abortion there is evidence that women, particularly those from lower socioeconomic echelons, resort to clandestine abortion in dealing with unwanted pregnancies (Felder and Oszlak. 1998; Ramos and Romero, 1998, Ramos et al. 1997). Different estimates (Ramos and Romero, 1998) suggest that between 335,000 and 500,000 abortions are performed yearly in Argentina, or about one every 16 to 24 women in reproductive age. According to recent research on the subject of abortion (Ramos and Romero, 1998) in Argentina, illegal abortion services are diversified and segmented by economic status. Poor women have only access to clandestine abortions performed in poor conditions by various means. Abortion (or their complications) are responsible at least for about 32.4 percent of maternal deaths and is the second cause of maternal mortality. 18 Measurement and Meaning conditions such as ectopic pregnancy (9.7 percent) and complicated delivery (8.1 percent) to poorly defined (2.7 percent) and unknown (10.8 percent) causes. Only 2.7 percent of women openly admitted that they did not want the pregnancy. However, other causes of abortion such as excess of physical activity (6.8 percent), use of contraceptives (4.1 percent), trauma (27.0 percent), and uterine bleeding (13.5 percent) may have an inferred association with induced abortion (Ramos and Romero 1998). Thus up to 60 percent of the causes for losing a pregnancy may have been the result of an induced abortion. This represents 16 percent of the overall sample, which is high considering that estimates for Argentina indicate that no more than 10 percent of the women in reproductive age go through an induced abortion. Half of the women do not receive postnatal care after delivery. One fourth receive one check-up, and another fourth attend more than one postnatal care visit. Almost all women breastfeed their children, yet the probability of breastfeeding for more than six months is lower among women who declare wage work either as their primary or secondary activity, regardless of their educational level. Women who do not declare to wage work tend to breastfeed their children until the baby is entirely weaned or the woman cannot continue breastfeeding. As women age, the length of breastfeeding decreases only slightly. Most women breastfeed on-demand as the child is with them while they perform daily tasks. For many women, breastfeeding operates as a family planning method. Nevertheless, the long breastfeeding period in which women engage in represents a severe toll on their time as well as on their bodies. Women indicated that breastfeeding results in their seeing the first changes in their bodies after pregnancy, which in some extreme cases could end up in clear physical weakening, including in some cases loosing teeth. Women's health status was assessed by asking for the occurrence of illness during the four weeks prior to the interview. Of those women who answered positively (36.2 percent), only 77 percent had sought health care. Women tend to minimize their health problems. They see their poor health status as the norm. Women's perception of their health in general is also associated to the fulfillment of their productive and reproductive tasks. If the illness does not severely impede her from performing in the short term her responsibilities she does not consider herself to be "sick"; rather it is just "not feeling well". Of those women that were ill, about 50 percent continued carrying on with their daily activities as usual, 20 percent indicated that were unable to do all their daily activities for two or three days and, only 25 percent of the ill women declared that they were impaired for more than three days. Only one in four women that reported to be ill during the last four weeks received some type of treatment. Of those that sought care, one out of two went to the public sector for services, while 20.7 percent were treated at home. In the survey, the women's assessment of the quality of health services they received is good for 56 percent and very good for 26 percent of respondents. However, women indicated in focus groups that they often feel mistreated at public hospitals, particularly when the reason for the consultation refers to reproductive health. This dichotomy between the assessment of the quality of service registered using a close scale (survey) and the open ended format of the focus groups suggests that women's perception of their health and the services they get cannot be assessed only by asking them to rank the quality of service in a point scale. Rather, it is required to give women the opportunity to express in their own words how they felt about the services. Besides this methodological problem, the results indicate that the Chapter2. Reproductive Health in Argentina's Poor Rural Areas 19 women have low expectation levels regarding quality of services. Women indicated that they cannot afford a better service because they are poor and, thus, had no other option. Comparing the survey sample to other data sets According to a 1996 report by the Argentinean Ministry of Health and Social Action, the reproductive health profile of Argentina is worrisome and precarious due to specific economic, social, cultural, and institutional factors (Gogna et al., 1998). These include the low social status of women; the sexual division of labor and lower pay for women; lack of reproductive health services, including family planning for both men and women; lack of insurance coverage, day care, and preventive health services, particularly cancer and Sexually Transmitted Diseases screening; and lack of social services dealing with domestic violence and other situations of abuse and neglect. As compared to this diagnosis, which is valid for the nation as a whole, the situation in poor rural areas is even more serious. In rural areas, the lack of support, poor access to health services, and domestic violence results in multiple reproductive health-related conditions that impair rural women's well being and contribute to their social exclusion. In this section, we provide data to help situate the basic statistics provided in the previous section in the Argentine and Latin America context. There are clear disparities between urban and rural areas in reproductive health statistics, such as the percentage of births attended by a health professional, which is 96 percent for the nation, yet 75 percent for the rural areas nation-wide. The maternal mortality rate reaches up to 65 per 100,000 live births in rural areas, compared to 48 for the whole nation. The women in the northern provinces of Argentina have a general mortality rate approximately 4 times higher than the men. While these provinces have only 20 percent of live births, 25 percent of the total infant mortality and 38 percent of maternal mortality are concentrated there (OPS/OMS 1999.) A study by Bortman et al. (1999) examining the relationship between gender, premature mortality, and low income found in 1999 that rates of premature mortality are highest in the poorest provinces of the nation. Premature mortality in these regions was found to be much greater among female population than among male population. A comparison of selected reproductive health indicators from the survey sample and nation-wide data for poor urban women (table 2.1) suggests that there are significant differences between poor urban and rural women regarding reproductive health behavior which cannot only be attributed or explained by their economic status relative to their location. That is, poor rural women have worst indicators than poor urban women. Table 2.1: Reproductive Health Indicators for Poor Rural and Urban Women in Argentina Poor rural women in Poor urban women survey sample in country statistics Average age of first pregnancy 18.2 19.5 Pregnancies in women age 18 and under (percent) 54.2 % 43.1 % Use of modem family planning methods (percent) 25.3 % 47.8 % Average family size 6.6 5.5 Number of children 3.8 3.1 Single female headed households (percent) 16.0 % 19.4 % Health insurance coverage 24.7 % 42.0 % Pre-natal care (percent of women with one or more control) 92 % 96.5 % Source: Siempro (1997) Social Development Survey; INDEC (1994) Childhood and Motherhood Study. Adolescents, individuals from the lowest socio-economic levels, and rural populations are among those with the most limited access to family planning services. In the province of Santiago del 20 Measurement and Meaning Estero, it has been reported that family planning services are not available at rural health posts. Although one of the most popular forms of contraception among Argentinean women, the birth control pill costs approximately US$ 5 per month, making it inaccessible to the majority of rural, poor women. A study conducted in Argentina by Pantelides et al. (1995) showed that more than 40 percent of all adolescents surveyed used no form of birth control, with 25 percent of adolescent females abstaining from the use of contraception due to a wish to become pregnant. More generally, table 2.2 compares reproductive health indicators between our sampled rural population and rural poor households in other Latin American countries. The data presented provided sample statistics for the bottom two quintiles of the income or wealth distribution in each country. The results suggest that rural households in our sample are doing better in most cases than rural households in the first quintiles of other countries for several indicators (including prenatal care, place of delivery, knowledge of AIDS). However, this does not hold for family planning since the value for Argentina in our sample is only marginally above the mean value for the other countries. That is, table 2.2 suggests that by international standards access to family planning services remains irregular in Argentina (Stout and Dello Buono 1996, Subsecretaria de Programas de Salud 1997). We focus on family planning and contraception in what follows. Table 2.2: Rural reproductive health indicators for bottom two quintiles, selected countries Argentina Haiti Nicaragua Bolivia Guatemala Qi Q2 Q1 Q2 Qi Q2 Q1 Q2 Q1 Q2 Prenatal Care Doctor/nurse 63.4 84.7 44.3 58.5 65.3 80.2 38.0 54.1 33.4 39.2 Doctor 41.7 72.7 18.5 23.1 27.4 38.0 25.2 43.0 23.7 28.5 Nurse 21.7 12.0 25.8 35.4 37.9 42.2 12.9 11.1 9.6 10.6 Two or more 66.0 65.7 43.3 53.4 60.1 76.0 30.7 47.7 75.6 82.0 Place of delivery Public 35.7 74.3 1.9 3.7 29.8 49.1 13.0 27.9 8.4 13.5 Private 3.5 14.0 0.0 0.4 0.2 0.3 1.5 6.2 0.1 0.6 Home 59.6 11.7 97.5 95.1 68.5 48.6 85.0 64.6 90.8 85.2 Use of (contraception) 32.9 37.2 5.0 7.0 39.6 52.8 6.2 14.4 5.1 9.5 HIV knowledge 56.0 69.7 Na na 56.4 66.5 40.0 49.5 30.9 32.9 Dominican Republic Paraguay Colombia Peru Brazil Prenatal Care Doctor/nurse 95.9 98.3 69.5 79.5 61.6 81.2 36.4 60.5 59.2 86.0 Doctor 94.3 97.9 13.3 20.0 57.3 78.0 8.6 18.9 50.7 83.5 Nurse 1.6 0.3 56.2 59.5 4.3 3.1 27.8 41.7 8.4 2.5 Two or more 92.9 97.0 77.6 81.8 57.4 80.7 32.8 57.8 55.8 85.0 Place of delivery Public 82.5 79.8 19.8 23.0 40.8 70.2 6.8 21.2 69.9 81.5 Private 4.4 16.5 4.7 6.3 2.0 4.1 0.4 2.0 1.7 10.9 Home 12.2 2.9 74.5 68.8 56.4 24.4 91.3 75.4 26.2 6.1 Use of (contraception) 50.4 63.1 20.3 24.9 42.4 61.4 23.0 31.3 61.8 74.6 HIV knowledge 81.5 90.2 Na Na 60.0 82.6 28.9 43.1 54.5 67.3 Source: Own survey for Argentina; Gwatkin et al. (2000) for other countries. All figures are percentages. The quintiles are defined by wealth rather than by income. Chapter2. Reproductive Health in Argentina's Poor Rural Areas 21 QUANTITATIVE ANALYSIS: CONTRACEPTION, DELIVERIES, AND WORK PATTERNS Impact of contraception on the probahility of a delivery Poor households need to pull multiple income flows together to secure their livelihood. Thus, women from poor households are more likely to have wage employment or other types of work. At the same time, the poorest households are larger and women have to take care of more children, which decreases their ability to seek full time employment, as they are forced to combine productive and reproductive tasks overwhelming their physical and mental capacity to cope. In this section, we use regression analysis to analyze the impact of contraception on the probability of having a child in the last three years, and the subsequent impact of a delivery on the probability to work and the number of hours worked. The methodology is presented in box 2.2 and the regression results are presented in table 2.3. The main findings are as follows: * Education: Women who have more than seven years of schooling have a higher probability of using modem contraception and a lesser probability of delivery. This result is interesting because the average number of years of education in the sample is 6.3 years and other studies of rural areas suggest that a significant number of women do not go beyond primary school (6th grade). The education of the spouse does not have a similar impact on the use of contraception and deliveries. * Religion and ethnicity: perhaps surprisingly, there is a higher probability of using contraception among Catholics. However, the data do not indicate that religious affiliation has an impact on the probability of delivery. There is a higher probability of using modem contraceptive methods among indigenous Coya women. * Poverty (as measured by food consumption) and health insurance: When food consumption per capita is higher (that is, when the household is richer), the probability of having children decreases. This result is compatible with the general idea that poor people have more children. However, economic status does not affect the probability of using contraception. The data also indicate that there is a higher probability of delivery over the last three years among those women that have access to health insurance. However, as expected, health insurance has no impact on the use of contraceptive methods. * Regional effects: Living in Misiones increases the probability of using family planning and lessens the probability of having a delivery over the past three years. This may reflect the fact that Misiones is the only of the three provinces in which there has been a reproductive health program in place. * Age and previous pregnancies: As expected, having older children reduces the probability of delivery, while having babies increases it. Also as expected, younger women are more likely to deliver. 22 Measurement and Meaning Table 2.3: Impact of Family Planning on Delivery in Last 3 Years, Rural Argentina, 2000 Probability of delivery in Probability of using modern the last 3 years contraception Coefficient Std.error Coefficient Std.error Misiones -1.03** 0.49 1.09** 0.38 Salta 0.60 0.49 -0.75** 0.37 Number of babies above three 0.63* 0.37 0.65* 0.33 Number of babies squared 0.02 0.14 -0.19 0.13 Number of children (from 6 to 14) -0.69** 0.32 -0.12 0.24 Number of children squared 0.09 0.06 -0.01 0.05 Number of adults -0.15 0.37 0.20 0.34 Number of adults squared -0.02 0.04 -0.02 0.04 No spouse -1.36 0.99 -0.13 0.85 Married -0.81 0.95 0.35 0.78 Cohabitating -0.84 0.94 0.47 0.77 Literate -1.25* 0.68 0.03 0.43 Years of school of the woman: 4 to 6 0.20 0.47 0.02 0.37 Years of school of the woman:7 -0.67 0.45 0.52 0.36 Years of school of the woman:> 7 -1.48** 0.59 0.89** 0.44 Years of school of the spouse: 4 to 6 -0.13 0.49 -0.06 0.37 Years of school of the spouse:7 -0.30 0.45 0.00 0.35 Years of school of the spouse:> 7 -0.90 0.56 0.32 0.44 Between 21 and 25 years old -0.49 0.72 -0.17 0.45 Between 26 and 30 years old -1.10 0.78 0.17 0.50 Between 31 and 35 years old -2.00** 0.83 0.68 0.54 Between 36 and 40 years old -2.24** 0.86 0.49 0.55 Between 41 and 45 years old -3.39** 1.03 0.36 0.64 More than 45 -4.67** 1.07 -0.11 0.60 Wage of the spouse 0.54 0.39 0.13 0.29 Birth migration (more than 100 Km) 0.00 0.00 0.00 0.00 Catholic 0.22 0.39 0.68** 0.31 Indigenous -0.21 0.31 -0.17 0.25 Colla -0.89 0.65 1.03* 0.55 Criolla 0.56 0.39 0.03 0.27 Other ethnic identity -0.95 0.65 0.22 0.55 Is the one who manages money -0.31 0.31 0.20 0.25 Visits family at least every two weeks -0.58* 0.30 0.31 0.23 Visits friends at least every two weeks -0.15 0.33 0.26 0.27 Goes to town at least every two weeks 0.13 0.27 0.18 0.22 Thinks she has a good health 0.00 0.29 0.10 0.23 Knows about HIV transmission -0.32 0.34 0.08 0.26 Is member of a religious organization 0.08 0.33 -0.17 0.27 Has a health insurance 0.85** 0.39 0.33 0.28 NBI index 0.07 0.06 0.07 0.05 Food consumption per capita -0.03** 0.01 0.01 0.01 Closest distance from telephone -0.01 ** 0.00 0.00* 0.00 Constant 7.02** 1.98 -3.91** 1.42 Source: Authors' estimation. 243 observations, Wald Chi2 of 113.16. Coefficients marked with ** are significant at the 5 percent level and coefficients marked with * are significant at the 10 percent level. Province of Santiago del Estero is omitted. Chapter2. Reproductive Health in Argentina's Poor Rural Areas 23 As explained in box 2.2, the regression estimates presented in table 2.3 and the correlation structure between the error terms of the two regressions can be used to estimate the impact of contraception on delivery, controlling for the other characteristics identified in the regressions. Table 2.4 suggests that contraception reduces the probability of a delivery by about 11 percentage points, and this result is robust to changes in the specification of the bivariate probit regressions. Table 2.4: Probability of Delivery Conditional to the Use of Modern Contraception Baseline Sensitivity tests Delivery in the Delivery in the last Delivery last Delivery in Regression last three years 3 years, without year the last two only for demographic years women variables in below 46 regression Probability of delivery 59.69% 62.50% 36.82% 50.88% 61.98% with contraception Probability of delivery 71.04% 70.57% 47.58% 65.47% 73.35% without contraception Difference in probability 11.35% 8.07% 10.77% 14.59% 11.37% Source: Authors' estimation. Impact of a recent delivery on work Rural fertility is closely linked to a women's socio-economic status, but the causality goes both ways. Women in rural areas and those with little education are unlikely to delay childbearing until the age of 18. A large share of adolescents give birth outside of a stable relationship, adding emotional and financial strains to motherhood. These young women have to cope without partner and/or family support at the same time that they see curtailed their ability to get the skills demanded for good jobs, thus limiting their ability to attain economic self-sufficiency (Alan Guttmacher Institute 1998). Upon the arrival of children, women often withdraw from the labor market in need of caring for the home and family. Women are forced to return to the labor market after the arrival of children if they are single mothers or if their partner is unemployed. But the decision of women or girls living in poverty to work outside of the home is also determined by the needs of the domestic unit and the life cycle phase of the family. Development opportunities for women are thus likely to reap more benefits if they take into consideration reproductive health issues. Family planning allows women not only to decide whether and when to have children (the reproductive and private dimensions of their life), but also to have better control over their participation in the labor market, other productive income generating activities and public/civic life. Intuitively, a key factor affecting the income generating capacity of women is the number of children they have to take care of, as the reproductive activities associated to child rearing diminish the time and energy the women may have for developing new skills and embarking in income generating activities. While intuitive, this hypothesis must be tested empirically. This is done in table 2.5 which provides an analysis of the determinants of the probability to work and the number of hours worked (for details on the methodology, see box 2.2; we did not analyze the determinants of wage earnings because the data on wages did not appear to be of good quality; the data on hours worked is of better quality). 24 Measurement and Meaning According to table 2.5, better educated women are more likely to work, and they also tend to work more hours per week. Similarly, healthy women (i.e. who perceive their health as good) work more hours. Migration (i.e., being a woman living at more than 100 kilometers from her birth place) is also associated with a higher probability of working and working more hours. As expected, not having a partner or spouse increases the probability of having wage work. There is a negative correlation between being indigenous (Coya) woman and having wage work.. Older women are more likely to work, and to work for longer hours. This suggests that young women encounter more difficulties in entering the work force in rural areas, especially the poorest, less educated, less skilled youth. This finding is consistent with other studies of youth and women participation in the labor market in other countries from the region that indicate that young women, particularly poor, experience higher rates of unemployment and underemployment. More than half (53.5 percent) of the women that have wage work are 31 to 50 years old, while only 8 percent are 15 to 20 years old. This situation is worrisome since it has been demonstrated that overcoming poverty is conditioned in great measure by the possibility of gaining a second household income. One factor which may influence the women's entrance into the labor market is the high rate of adolescent pregnancies among women of lower income, but this is difficult to capture in the regressions because of the small sample size. Another element that affects young women's participation in the labor market in rural areas is the socio-cultural context that discriminates against women. Table 2.5 suggests that women do not stop working because of a delivery, nor do they reduce the number of hours they work. The coefficient for hours worked is negative but significant only at a 30 percent level. This low level of significance could be due to sample size as well as the fact that only paid work and paid hours are considered in the analysis and only a few women had paid work. The small number of observations is a constraint to the analysis, but nevertheless the direction and magnitude of the coefficient of delivery in the hours work equation was found to be robust to different specifications. One interpretation of this (admittedly fairly weak) evidence on the impact of recent delivery on work patterns could be that poor women with large households are forced to work on more insecure labor arrangements as they cannot have full time employment. This would increase the vulnerability of women and their households as these women are more prone to take on seasonal and part-time jobs that pay less and have no benefits. If we are willing to argue that women's reproductive work (non-wage work) reduces the probability of undertaking wage work, or reduces the number of hours worked, this may have other implications as well since women's participation in the labor market has a significant impact on the amount of money they dispose. On average, women in the survey that have either as primary or secondary activity wage work manage about 35 percent (US$ 133 compared to US$ 97) more money than those who do not have salaried work8. 8Managing more money does not automatically translate into an increase in women's status because the economic mobility of women remains linked to the income generation capacity of their partners. Also, women's identity need not change as they hold wage work because it is still attached to the fulfillment of their cultural roles as nurturers. Chapter2. Reproductive Health in Argentina's Poor Rural Areas 25 Table 2.5: Impact of Delivery on Probability of Working and Number of Hours Worked Probability of Working Number of Hours Worked (Probit model) (Tobit model) Coefficient Std.error Coefficient Std.error Misiones 0.06 0.33 17.03 12.88 Salta 0.50 0.36 33.43** 12.71 Delivery within the last three years 0.06 0.28 -10.95 10.32 Number of babies above three -0.02 0.28 -0.71 9.17 Number of babies squared 0.03 0.09 2.21 2.61 Number of children (from 6 to 14) 0.37 0.24 5.46 8.26 Number of children squared -0.05 0.04 -0.73 1.58 Number of adults -0.33 0.28 -17.79* 9.73 Number of adults squared 0.03 0.03 1.92* 1.05 No spouse 2.20** 0.92 63.99* 32.78 Married 0.11 0.86 13.89 30.53 Cohabitating 0.37 0.84 18.39 29.20 Literate -0.65* 0.37 -7.94 16.86 Years of school 0.06 0.04 5.36** 1.79 Years of school squared 0.20 0.24 19.10** 8.43 Years of school of the spouse 0.1 1** 0.05 1.48 1.84 Years of school of the spouse squared -0.26 0.26 -8.14 9.55 Age 0.19* 0.11 4.90 3.95 Age squared 0.00 0.00 -0.04 0.06 Birth migration (more than 100 Km) 0.50* 0.29 18.92 10.41 Catholic -0.34 0.27 -1.67 11.02 Indigenous -0.30 0.26 -8.26 10.27 Colla -0.89** 0.44 -26.68* 15.32 Criolla -0.39 0.27 -13.10 9.97 Other ethnic identity -0.91 0.56 -25.47 20.47 Thinks she has a good health 0.41 * 0.23 9.63 8.33 Per capita food consumption 0.02** 0.01 0.43 0.33 Constant -5.84** 2.16 -204.13** 79.78 # observations 262 259 Pseudo R squared 0.26 0.12 Source: Authors' estimation. Coefficients marked with ** are significant at the 5 percent level and coefficients marked with * are significant at the 10 percent level. Province of Santiago del Estero is omitted. The number of hours worked includes only paid work as a principal activity (due to limitations in the survey questionnaire). 26 Measurement and Meaning QUALITATIVE ANALYSIS: OBSTACLES TO CONTRACEPTION AND REPRODUCTIVE HEALTH Conceptual framework Among small farmers, woman's productive and reproductive roles are essential for the maintenance of the household unit (Babb 1990; Deere, 1982; Friedmann 1986; Gonzalez and Salles 1995; Mann 1990). The household is the basic unit where sexual division of labor takes place. As a result, a hierarchical household structure emerges, where even sexuality becomes an asset subject to control. The way in which the production of goods and services needed for exchange is organized shapes gender relationships. Men achieve upward mobility through their work, while women's social mobility is linked to the fulfillment of their culturally defined role as nurturers (Acker, 1990; Kelly-Gadol, 1987). While women have both reproductive and productive roles, women often do not have the same possibilities than men to participate in the public sphere of social and political life (Acosta-Belen and Bose 1990; Deere 1982; Gonzalez and Salles 1995; Jelin, 1990). Although women are increasingly left alone to care for the family, the children, and the farm, the pervasiveness of patriarchal social constructs in most rural communities difficult women's participation and capacity to control their own lives. Even single female-headed households tend to reproduce similar practices and are bound by social relations that maintain women's subordination. Gender roles among rural, poor Argentinean families do not depart much from the above and may be characterized as fairly rigid, with the female partner being responsible only for some productive and social reproductive activities (Barone, 2000; Schiavoni, 1996;). Gender roles are so imbedded in rural culture in Argentina that even when the husband is away for some considerable time, women tend to identify the husband as the head of the household (Schiavoni, 1996). When rural women establish domestic partnerships (either through marriage or common agreement), they generally become responsible for the daily domestic/reproductive activities of the family, and for activities in subsistence agriculture, such as caring for a home garden, raising animals, baking, and milling of grains (Barone 2000; L. Schiavoni 1996). Rural women may spend an average of 16 hours per day in productive and reproductive activities, and in crunch days of harvest periods, a woman's workday is increased by at least two hours. Gender identity is defined by multiple factors that contribute to the construction of a world-view and a representation of the self in relation to others, of a gendered reality. As such, it reflects particular social, economic, cultural and institutional arrangements. Men and women play different roles in society based on their gender roles and identity. The identity of a group defines its capacity for accessing and mobilizing resources. As such, gender identity should be analyzed as a specific manifestation self-representation strategies that refers to the power relations between men and women. However, to reduce gender identity to a pure relation between man and woman would be to ignore the fact that cultural stereotypes correspond to a certain economic and social order that consolidates a certain specialization, a division of functions and attitudes which can be mapped along two main dimensions: (i) the productive-reproductive, and (ii) the public-private continuum that split activities and responsibilities of the household members. This division shapes interpersonal relationships as well as how men and women relate to the social, economic and political spheres. Chapter2. Reproductive Health in Argentina's Poor Rural Areas 27 Figure 2.1 depicts the relationship and interactions between the public-private domains and the reproductive-productive roles. The figure illustrates women's activities in the context of their social roles of production and reproduction and allows to analyze how these interactions shape gender identity and reproductive behavior. The activities in the different spheres are not mutually exclusive: women play many roles which often overlap. The survey used for this study asked multiple questions to elicit women's perception on issues related to their productive/reproductive and public/private lives. Although this has not been discussed in the previous section, the answers to these questions were later on classified and a score was assigned to each answer based on the place they occupied on both axes. Then, the scores were summarized in an index for each woman in each dimension and plotted. The results of the analysis suggest that most of the women represent themselves in the sphere of the "reproductive-private". Little variation was observed in the answers either across provinces, economic status, educational level or age of women. Figure 2.1 Dimensions of Women's Activities and Roles Reproductive Social Production Sexual Reproduction Caregiving Household work Family networking Public -Z2 Self Identity Private Wage work Subsistance production Accumulation Social networking Productive Source: Authors 28 Measurement and Meaning Box 2.2: QUALITATIVE ANALYSIS - METHODOLOGY Qualitative data was gathered through focus groups and in-depth interviews with key informants. A total of 13 focus groups were organized (6 only women, 4 only men and 3 men and women together). Female participants were all females of reproductive age with and without children. Male participants included both partners of the women and single males with and without children. Focus groups sought to identify perceptions about reproductive health including definitional issues, control over reproductive health, decision making processes, access and management of assets, access to services, and economic and social activities. A detailed guide was designed as a protocol for conducting and recording focus group sessions in order to ensure consistency in the different provinces. In addition to the focus groups, 18 in depth-depth interviews (7 in Misiones; 4 in Salta and 7 in Santiago del Estero), with key informants were conducted to obtain context infonnation, check consistency of the findings and register the views of key stakeholders regarding reproductive health issues. Informants included staff from government programs (extension agents) working on rural development; staff from government reproductive health services; rural women leaders; policy makers; school teachers, and religious leaders from different denominations at the community level. In an attempt to comprehend how the women in the sample viewed their own identity, this study asked multiple questions to elicit women's perception on issues related to their productive/reproductive and public/private lives. The answers to these questions were latter on classified and a score assigned to each answer based on the place they occupied two axes: (i) productive-reproductive and; (ii) public-private. A modified Delphi technique was used to code and assign scores to the women's responses to the open- ended questions about self-identity and gender roles. Four persons codified the answers independently (3 research team members and a specialist not involved with the research team, who acted as a "control"). After the individual scores were assigned, in the case of divergence a consensus was reached regarding the coding and the related score for the answers in each axis. The scores were summarized in an index for each women in each dimension and plotted. Finally, two consultation workshops with national level government officials, civil society representatives and academics working on reproductive health issues were organized in Buenos Aires. In these workshops the study design and, later on, the preliminary results were presented and discussed with the participants to ensure the views of local experts and policy makers were properly considered in the analysis. All data collection was completed in calendar year 2000. Source: Authors. What constitutes the domestic and public spheres vary from culture to culture. In the case of poor rural communities in Misiones, Salta and Santiago del Estero, the activities related to the market or to social participation are considered to belong to the public sphere and they correspond primarily to men. At the same time, the relations within the household, while they may be affected by the market, still are not defined by the market, as long as they remain within the preview of women. Despite most women are linked to the market through simple commodity production and/or through the sale of their labor force, women have little access and control of productive resources and are basically excluded from political power. The result of this situation is a common identity developed from an inherent subordinated position within the social structures. "Fulfillment" for her is only possible through her submission to reproductive and private roles. In relation to the above, the convergence of multiple mechanisms of subordination becomes crucial in the explanation of women's position both at the household and the market. The relationship between women's work and household strategies reduces subsistence costs and facilitates the Chapter2. Reproductive Health in Argentina 's Poor Rural Areas 29 production of services and goods that, otherwise would require more wage labor. Women's subordinated position is the expression of the interaction between these dimensions. Women see their sexuality and biological reproductive process divorced from the public sphere. However, this view reflects the roles assigned to women to assure the reproduction and maintenance of their families. By excluding women from certain areas of activity, the biological reproduction is left as the only sphere of self-realization and, at the same time weakens women's control over their own sexuality. Sexuality, procreation and emotions all intrude upon and disrupt the reproduction of the established institutions. Pregnancy is viewed as a fundamental event in the lives of women because it opens the way towards motherhood and their identity as women. Maternity is valued over sexuality as pleasure, because reproduction gives a woman a place in society and is the social reference allowing her to construct her identity. Therefore, pregnancy is accepted as a positive event and is socially valued despite the recognition that one of its consequences is that it increases the woman's responsibilities and work within the home. Women define themselves through motherhood. Becoming a mother is what makes a woman. At the same time, despite the great variety and importance of women's domestic and productive activities in sustaining the household, they tend to see those activities primarily as reproductive (i.e. related to the caring and maintenance of the household; that is to social reproduction). Further, even if a woman has a salaried job, that activity is subsumed under the category of the social reproduction of the household. As has already been mentioned, a woman's social reproductive work principally consists of daily activities tied to the domestic group (cleaning the house, washing and ironing clothing, cooking, and caring for children), as well as subsistence activities (caring for a home garden and raising small animals). At the same time, women participate in commercial/market oriented activities (such as working on the family farm, selling of goods, and making handcrafts). When women have wage-work they are responsible for the household, carrying a double burden which is not usually recognized. To carry out her responsibilities, the female partner is generally assisted by other members of the household (sons, daughters, her mother). The only acceptable reasons for a woman to excuse herself from her daily responsibilities are when she is about to give birth, having generally worked until the onset of labor, or when suffering from serious health problems. The gender identity reported by women in this study has a direct impact on reproductive health. First, women are raised to be mothers. This implies that controlling the (biological) reproductive process is important in as much as it ensures the fulfillment of the expected role. As a result, women start their reproductive lives very early because that is the main avenue for asserting their position in society. Third, the control of the reproductive process and of women's sexuality becomes privatized. Women's reproductive functions are seen as a subset of the household and, thus, subject to the dictates of men. Finally, the reproductive behavior of these women results in: (i) More children and increased time devoted to care giving and subsistence activities, (ii) a lack of preventive health behavior which leads to larger family sizes and increased health risks, and (iii) diminished capacity to control assets and generate income. In summary, the social position of these women in their communities is defined by the division of labor along gender lines (productive vs. reproductive), the construction of symbols and images that explain, express and reinforce those divisions (motherhood); and a set of social and economic norms and processes that make difficult for a new generation of women to build a different gender identity. 30 Measurement and Meaning Reproductive health awareness In order to be enabled to make decisions about one's own sexuality and reproduction, minimal conditions are necessary in terms of education, access to economic assets, and health. Reproductive health is a state of physical, mental, and social well being in all matters related to the reproductive system, its functions, and processes. Reproductive health involves awareness of reproductive and sexual life, controlling disease, promoting safe motherhood and providing family planning services and fostering a better quality of life. For both men and women in this study, the notion of reproductive health is extremely limited. The concept, in their minds, was related to the notion of health in general, revolving around the presence or absence of bodily illness requiring a visit to the physician, excluding any action or idea of prevention. When asked to attempt to verbally conceptualize the term reproductive health it was defined as "something pertaining to a woman 's health in her fertile years". Their idea of reproductive health is organized around the knowledge of menstruation and its function in procreation, and includes both the use of contraception and the importance of prenatal check-ups in the event of becoming pregnant. Therefore, it is seen as the exclusive responsibility of women as it is related to motherhood. There was no clear sense that motherhood is linked with fatherhood. Rather it seems that men and women considered "parenthood" primarily as a female "biological" event and social construction. Both women and men associated problems with reproductive health as mainly sexually transmitted diseases (STDs); problems during pregnancy, principally regarding abortion; and tumors related to the use of modem contraceptive methods, which many believe can cause ill health. Despite the fact that women did not possess a wide comprehension of the term, it was clear that women did have practical knowledge on certain aspects of reproductive health, particularly those regarding birth control and handling some health risks and illnesses. In a few cases, the idea was linked with improving oneself in order to have a better future. Nevertheless, both women and men demonstrate an interest in receiving information about these topics. In the focus groups, women indicated acceptance of their difficulties in exercising their sexuality and described attempts to resolve these issues, but often encountered socioeconomic barriers in doing so. To a large degree, most men and women consider that reproductive health and family planning are things that they themselves cannot control. The nature of reproduction is considered to be something "given", which is a belief that contributes to poor family planning and the maintenance of a vicious circle of dependency and poverty. In Salta, the women stated that, "if a child comes, it comes" and it happens "without thinking about it." Women also stated that, "there is nothing you can do about it". At the same time women placed high value on being capable of fulfilling their reproductive role as that was the cornerstone of their identity and place in society. Family planning Several elements influence the decision of whether or not to use family planning. In the case of the pills, women reported that even if they had wanted to use them they would not do it because of cost. In the case of intra-uterine devices (IUD), cost and access to services were mentioned as restrictions. In the majority of cases, free supplies were not available in the local hospital, and since these women do not generally have the money to purchase supplies on a regular basis, family planning methods become inaccessible. The use of family planning is also affected by privacy issues. Women stated that the place where to get contraceptives is very important. If the C'hapter2. Reproductive Health in Argentina's Poor Rural Areas 31 location is not private, women feel embarrassment and fear of being the subject of rumors. This is especially important in small communities. Using pills or other modem methods can be seen by men (and society in general) as an indicator of unfaithfulness. That is, men see the use of modem contraceptive methods by women as mechanisms that would allow women to "cheat" on them. In this cultural context, family planning becomes a tool of control of women by men who do not feel any responsibility vis-a-vis "motherhood." To effectively promote family planning would require to explicitly challenge accepted cultural controls such as this. Among men, one could argue that there is a double discourse on the issue of using modem family planning methods. On the one hand, there is a general acceptance that family planning is necessary for household economic reasons. On the other hand, they see it as a responsibility of the women that should be carried out using "natural" methods. Men differentiated between two main reasons for using family planning: (i) to prevent pregnancies; and (ii) to terminate pregnancies. Among the methods to prevent pregnancies, the men distinguish two sub-types: (i) natural methods, that they prefer to use with their spouses/partners; and (ii) condom, which they associate to single men and/or their relationships with other women (besides their wives or partners). Similarly, terminating a pregnancy is seen as a last resort, mostly for single women (not their wives or daughters). While some men expressed the use of a condom to be unnatural, some women saw it as an expression of respect. Other considerations in the decision to use family planning relate to their perceived health implications. Some women perceived modem methods as abnormal for the body because they alter the body's natural functioning and a woman's physical appearance. One of the focus group participants stated that, "it is not good to use modern methods (the pill) because your menstruation is not like always, and that means that your body is not healthy. " Women make the comparison with their mothers who did not use modem methods and lived for many years. Existing beliefs are firmly routed in the women's traditional knowledge regarding reproductive issues which have been passed through generations by mothers to daughters (see Barone, 2000 and Fogel, et al 1993). Yet, family planning is a topic in which the mothers themselves generally lack precise knowledge. While women do have some basic tools to manage their reproductive life, at the same time their lack of knowledge limits the possibility of incorporating new practices. The women lack basic understanding of their bodies and reproductive systems, and as a result are misusing some of the natural family planning methods. The focus groups results also suggest that health personnel do not generally provide information on family planning methods or attempt to create a consciousness on the importance of family planning. As for the women, they fear seeking family planning consultation from health personnel because they perceive a power differential between themselves and health personnel, and they are afraid that a moral judgment will be made against them by health personnel. It can be inferred that poor rural women need more information and access to resources that would help them to redefine their gender identity and acquire more control over their bodies and reproductive processes. Prenatal care and abortions The male partners do not take part in the prenatal control or the delivery. Women tend to go with other females (kin or friends). While women see this behavior as "normal", they would like their partners to have a more active involvement as that would give them more security. In the opinion of women, the reasons for this are the lack of interest on the part of men as well as the need for 32 Measurement and Meaning men to keep working. More generally, prenatal and postnatal care are seen as something needed only if there is a problem. Both, women and men see medical attention due to maternity as the loss of a day of work, plus the expenses associated with the visit to the care center (co-payment), and additional expenses such as transportation and meals. Pregnancy does not typically alter the domestic routine. Women work until the day of delivery. When the woman is away from the home for delivery, men often take charge of some of the household duties with the help of their daughters (if any) or other female relatives. For legal as well as cultural reasons, induced abortions are often hidden and described as natural occurrences. In the focus groups, both men and women express disapproval of abortive practices. Nevertheless, women do accept that it is a frequent practice and that it is understandable, particularly in cases where the woman is alone, very young, or has too many children. Both men and women agree that abortion should be prevented by the use of family planning methods. However, if unavoidable, it is the responsibility of the women to decide to have an abortion. When interviewees were asked to consider themselves making a decision about whether or not to have an induced abortion, four out of five expressed being against abortion. Only 7.3 percent asserted that they would have an abortion, while 9.7 percent said that they would perhaps consider abortion. Males in general were more open to speak about abortion, but this may be primarily because they do not take responsibility for making the decisions. Despite these findings, female focus group participants agree that abortions occur frequently in their communities and, as indicated before, about 16 percent of the women interviewed are likely to have had an induced abortion in their lives (the rate may be much higher in the sample). Despite that both abortion and women electing to have abortions are viewed negatively, there also exists an unspoken social sanction against young, single women having children. One woman characterized this by saying, "They (referring to the community) point their fingers at you." Thus women fear double social standards. It is not seemingly to use contraception, or have abortions or have children while young and single. Yet, there are instances in which women believe abortion is justified. These include cases of sexual abuse and instances when the pregnancy is the product of incest or relations with a married man, the women are unmarried or, more frequently, in families with many children. Sexuality, sexual abuse, and domestic violence Women view themselves primarily in the context of motherhood and social reproduction. Sexual relations are not seen to be a dimension of self-fulfillment, even though a large number of women (73 percent) indicate that it is important to enjoy sexual relations. For women the concept of sexuality is limited to procreation and menstruation. Puberty marks the beginning of sexual activity and fertility, and acts as a rite of passage from girlhood to womanhood/motherhood throughout adolescence. Men, on the other hand, are considered to be highly sexual. Women state that sexuality is important for both men and women, though more so for men due to their highly sexual nature. Sexual education, then, is seen to be especially important for young men. Women's perception is that men experience unlimited sexual desire, which women must satisfy whenever the man demands it of them. Women, in general, conceptualize sexual relations not as a choice, but as a consequence of couplehood in which rejection of sexual advances by the man is not possible. Most women indicate that their main source of information on reproductive health is their mother. However, the topic of sexuality is not addressed within the family unit, Chapter2. Reproductive Health in Argentina's Poor Rural Areas 33 even between mothers and daughters. Data from the focus groups show that sexual information is passed between brothers and sisters, neighbors, friends, and partners. Information on sexual health is also disseminated through mass media. In some cases (in Santiago del Estero), the women indicate that discussions on sexual health are conducted in secondary schools with males and females students together. In the province of Salta, community health agents conduct discussions on sexual health with groups of women. Discussions on sexuality are also organized by the church. Women indicate in the focus groups that they are often forced to have sexual relations with their husbands. These forced relations are perceived by women to be violations and bring about feelings of strong anger and depression. One woman characterizes it as follows, "They (men) believe they have rights (to sex). The man doesn't ask the woman how she feels." Perceptions related to the ability or inability to control and respond to such violence are linked to inequalities between men and women. These cultural constructions subordinate women to men because women are economically dependent on their husbands. These constructs are so strong that in the mixed focus groups in all three provinces, women tended to remain silent on the topic of domestic violence. While men and women agree that domestic violence exists, men tended to minimize the problem of violence and sexual abuse. They argue that it is something "from the past". However, the women emphasize that they have to obey to the wishes of their husbands/partners. Shame, fear and lack of economic support are some of the factors that women mentioned prevented them from going to the doctor for treatment or from reporting incidents of violence to the police. Further, when reporting to the police, the women indicate that, in most cases, there is no sanction or even worst, the men reacts violently, aggravating the situation9. Social Net workling Women's participation in social organizations is low. The majority of the women do not participate in mother's groups, parent's groups, neighborhood organizations, farmers organizations, or political parties. Participation in religious organizations seems high, however, as was indicated in the focus groups, it refers more to attending service than to be involved in specific groups or activities within the church organization. Although religious groups are important sources of information and cultural norms, participation in these groups is not associated with participation in other type of social organizations. These results are consistent with the findings of a recent World Bank (2001) study on social capital in Argentina which indicates that less than 20 percent of the population participates in any form of organization, church participation being the most frequent. Women that participate in social organizations with the exception of church groups tend to be more educated and younger. Women with more than 12 years of education are more likely to participate in neighborhood organizations. However, women that participate only on religious groups tend to be older and less educated. Participation in neighborhood organizations and political parties is higher among single women. Women with 9A qualitative study on gender stereotypes, power relations, and risks for STDs conducted among both men and women in a low-income neighborhood in Greater Buenos Aires, Argentina, has suggested links between health/illness, sexuality, and gender (Gogna and Ramos 1998). The study found that beliefs regarding STDs and the risk of infection were strongly tinted by deeply rooted ideas regarding gender identities, gender relations, and sexual matters; and that STDs have a very special social and cultural meaning that greatly affects prevention and treatment behavior. For some women, the perceived eventual risk of being beaten or abandoned, or of losing a source of emotional or financial support, far exceeds the perceived health risk of a STD. 34 Measurement and Meaning no children show little participation. Women in the focus groups indicated that socialization and participation in social organizations for young (never married) single women is more difficult because of social norms. On the other hand, women also indicated that the number of children represents a constraint to participation. Although the women interviewed have interest in social participation, they lack opportunities to engage in such activities. First, the data show that women in the early years of motherhood (typically with two or three children) tend to participate less than women with older children or women without children, as they typically lack childcare options or are unable to bring their children with them. Second, community organizations do not facilitate the building of social networks that are pertinent to women. Unlike men, who typically have numerous opportunities for recreation and diverse social participation related to productive, sport, political, or other activities, women are often limited to religious organizations. This is significant for two reasons. First, the data show that women's participation in social organizations depends to a great extent on their control of the reproductive process. Second, the results indicate that women's participation in social networks is important for increasing awareness of economic opportunities for income generation. It may be inferred that the responsibility and time involved in raising children preclude women with children from engaging in social participation activities, as women are compelled to develop their social and productive lives around the household. Health care and coverage In the focus groups, both men and women agree that their biggest problem is the lack of medical coverage due to shortcomings in the health system. They state that doctor visits to their communities are inadequate; health facilities are not well equipped, there are few opportunities to make appointments, and few days of consultation. Some women indicated that regional hospitals did not have enough beds to take in al patients. Other participants talked about mistreatment at the hands of providers. Problems with access seem to be primarily one of cost: 1) opportunity cost by losing income generation time; 2) cost of transportation; 3) availability and actual payment for services and drugs; and 4) co-payment for the few who have health insurance. The efforts to overcome these hindrances are too great given the quality of and variety of services available. This is problematic because according to national data from Argentina, early detection of pregnancy and prenatal check-ups play an important role in the prevention of maternal mortality, and health coverage is closely related to the early detection of pregnancy. When health coverage is private or is provided by an Obra Social, 87 percent of pregnancies are detected in the first trimester, and less than one percent are never detected. However, when the only health coverage available is through a public hospital, the situation changes dramatically with only 66 percent of pregnancies detected during the fist trimester and 5 percent never detected (SubSecretaria de la Mujer, 2000). A similar relationship is seen regarding prenatal visits and delivery. Thus, the lack of health coverage, as found among the women interviewed, is likely to result in negative health impacts. Women report in the survey that they tend to use regional hospitals instead of district hospitals or local health posts closer to home. This is often due to the lack of gynecologic services, as local Chapter2. Reproductive Health in Argentina's Poor Rural Areas 35 health posts are usually staffed with generalists onlyl. Women have to travel long distances to reach appropriate medical attention, facing obstacles such as lack of economic resources, lack of transportation, and limited hours of operation, among others. Some women indicate that they had felt mistreated by staff in regional hospitals, who did not treat them humanely. One woman indicated that "the health personnel say anything and have things done their way, while we just have to listen quietly". Still, overall, when services are available, the main reason for not utilizing them is their limited accessibility rather than the low quality of treatment. For the rural population, traveling large distances to public hospitals is difficult and costly, and represents an opportunity cost to the family. Some male focus group participants complained about the long distances to health facilities. Both women and men denied the importance of the curandero and traditional medicine. However, with some probing, both men and women discussed situations in which they consulted traditional practitioners/healers. The population often limits the utilization of health services to instances of diagnosed illness (i.e. every day aches and pains are self- medicated and treated either with pharmaceuticals or traditional medicines), prenatal care in the days immediately before delivery, and for the birth itself. The study population felt that the health coverage available is insufficient to meet their basic needs. The overall perception, then, is a feeling of being unprotected, without equal access to health services and good health. CONCLUSION There is no clear population and reproductive health policy in Argentina today. Current efforts on responsible parenthood are moving in the right direction but slowly. Without political will at the national and provincial levels to implement comprehensive reproductive health programs, the welfare of rural poor populations, particularly women, will remain precarious. A sector analysis is important for providing information for reform based on the identified needs of specific populations as well as on weaknesses on the supply side. Existing statistics are scant and the available information tends to be out-dated, based primarily on hospital mortality statistics, and managed at the provincial level with no consistent standards across provinces. There is a need to develop and/or improve Argentina's socio-demographic-health information and surveillance system beyond the medical/biological dimension. An improved information system should include indicators not only on reproductive health (the epidemiology of and trends in high-risk fertility behavior, the performance and quality of reproductive health services, and the health, population and nutrition outcomes relevant to reproductive health), but also on productive activities (income generating opportunities for women) and the availability of social networks/support groups for women. Awareness of reproductive health, access to services, and use of family planning methods are low in poor rural areas. The adolescent and youth pregnancy rate and the abortion rate are high. Prenatal care is inadequate and there is no evidence that post-abortion care and guidance is available or that alcoholism or domestic violence are addressed by the health system. A constant in the majority of problems affecting the population under study is the lack of reproductive health education, which keeps them in the dark as to services they could demand. More generally, this study has documented the importance that basic education has on the reproductive I0In Misiones many prefer to use public hospitals despite the fact that public hospitals are generally farther away from their homes, because there they may receive free drugs. In this way, availability and cost of medications play important roles in the decision of where to go for health services, even more so for those without health insurance. 36 Measurement and Meaning behavior or poor rural women. Completion of the 6th grade has a significant impact on the use of family planning, family size, as well as other important behaviors that improve the family quality of live and decrease infant mortality. A key recommendation would thus be the development of adult education programs for poor rural women age 15 to 35 that have not completed their primary education. Besides improving reproductive health indicators, completing primary education would allow these women to have better chances of finding employment or developing income generating activities. A second recommendation would be to implement targeted population education activities to promote reproductive health and preventive behaviors, while recognizing the populations' social, economic, cultural and political context. Health promotion efforts could be implemented in partnership with private sector and civil society organizations that have experience in addressing these issues and are used to work with rural populations. Additionally, until now most family planning education efforts have been directed to women alone. The focus should be broadened to include the women's partners. The involvement of men and youth as much as women is essential for social progress when addressing issues such as gender roles, high risk behaviors, family size or domestic violence. The health services now available are primarily pregnancy related. There is a need to revisit the current health/maternity services package so that they can meet the needs of rural poor populations for more comprehensive reproductive health services. Primary health care, as offered in both community health posts and hospitals, needs to better respond to the population's reproductive health care needs. Our study suggests a few priorities: (i) early and complete prenatal control, including nutritional status; (ii) postpartum controls, including port-abortion care; (iii) routine breast and cervical cancer check-ups (although this has not been discussed explicitly in this paper, it is supported by the data); and (iv) family planning for women, men and youth. To increase the utilization of such health services, it will be necessary to: 1) make available the appropriate provider (e.g. physician, midwifes) and medications; 2) have providers on the premises at all times, or at least increase their presence; 3) give prompt appointments; 4) improve education and counseling of clients and how they are treated; 5) utilize extension health workers for outreach; and 6) facilitate transportation to health facilities. Decentralized and mobile services staffed by specially trained women health specialists could be an option to facilitate access. In addition, promoting health preventive behaviors (including those to address unwanted pregnancies) as well as the prevention and treatment of alcohol abuse and domestic violence is recommended. Particular attention should be given to adolescents since they are at higher risk and in greater need of reproductive health services. Teenage pregnancy increases the risk of health problems and it lowers the future socioeconomic status of women since pregnant girls often leave school. Reproductive health cannot be divorced from broader socioeconomic and political issues that shape and are shaped by reproductive behavior. The distribution of household tasks contributes to the definition of gender identity. It reflects a complex arrangement of responsibilities that results in the creation of different types of "workers" who need each other to ensure that their needs would be met. The data reviewed in this study indicate the splitting of household labor into two components, the productive (linked to the market) and the reproductive (as an input) only reinforces the social exclusion of poor rural women. Whether women are household heads or not, they organize their behavior first and foremost in order to ensure the subsistence needs of their families. Focus groups suggest that through this process, women become ideologically circumscribed to the reproductive sphere. As a result, there is a privatization of the social life of Chapter2. Reproductive Health in Argentina's Poor Rural Areas 3 7 women, with the public sphere emerging as a space primarily for males. Social institutions contribute to a situation in which women usually have little or no bargaining power for having more control over their reproductive and productive roles or their sexuality. This is compounded by women's low socioeconomic status which is affected by the amount of hours spent in housework activities, which increases with the addition of every child. Improving rural women's reproductive health and quality of life thus requires an integral approach. Education to both men and women as well as access to quality reproductive health services is just one component of this strategy. Equally important are the issues of access to assets and income generation opportunities. For example, programs such as nurseries and day care centers could be developed to facilitate women's insertion in the labor force. Finally, poor rural women have little participation in social organizations (other than religious). There are multiple factors that explain this lack of participation, from the burden of child raising to the lack of outreach strategies aimed at increasing women's civic engagement. Social organizations in rural areas tend to be male dominated or gender segregated. In the former, women's interests are not being addressed, while in the latter, the focus is on tasks or issues that do not necessarily advance women's overall well-being. To define family planning as the only or main strategy for improving the quality of life of rural poor women without proposing changes in life conditions would place women in a situation without the proper tools at their disposal. The possibility of constructing alternative or complementary identities to motherhood should go hand by hand with more equitable socio-economic changes that would promote equity and empowerment. At the societal level, the prevalent discourse regarding reproductive health issues has been constructed and accepted as legitimate. These truths have defined which are the accepted (sexual and reproductive) behaviors and the roles that each of the household members is responsible for maintaining. While male sexuality is socially permitted, female sexuality is split in a negative side related to eroticism which women is not allowed to express, and an overt (positive) side springing from its reproductive potential. Comprehensive reproductive health services and other social programs would need to address (in their design and/or implementation) issues such as this, that have a direct impact on the health of both men and women. This is doubly important given not only the lack of knowledge about reproductive health, but also the prevalent high-risk behaviors of men and women. Women's social exclusion is the expression of the interaction between ideological and economic dimensions. A central factor contributing to this exclusion is control over productive and reproductive processes, particularly those related to the biological reproduction of the household. The division of labor at the household level represents the interaction of specific material condition and ideological constructs that cannot be transformned without changing the conditions that support them. A key lesson from this study is that the success of a reproductive health program cannot be separated from the transformation of the socioeconomic conditions that engender women's social exclusion. That is, the transformation of reproductive behavior requires the transformation of the social, economic and cultural arrangements that shape that particular behavior. The foundations for this change are to be found in the daily processes in which women are involved taking advantage of the multiplicity of productive and reproductive roles women display in public and private spheres. A process of communication and sharing of meanings is essential to reformulate the prevalent gender identity of poor rural women. 38 Measurement and Meaning REFERENCES Acosta-Belen, Edna and Christine E. Bose. 1990. "From Structural Subordination to Empowerment: Women and Development in Third World Contexts." Gender and Society 4:299-320. Acker, Joan. 1990. "Hierarchies, Jobs, Bodies: A theory of Gendered Organizations." Gender and Society 4:139-158. Alan Guttmacher Institute. 1998. Into a New World, Young Women's Sexual and Reproductive Lives. Washington DC: William H. Gates Foundation. Amadasi, Enrique y Guillermo Neiman. 1998. Niveles de Vida y Pobreza Rural en las Provincias de Misiones y Salta. Secretaria de Programaci6n Econ6mica y regional. Direccion de Programaci6n y Gasto Social, Argentina. Babb, Florence E. 1990. "Women and Work in Latin America. (A review Essay)." Latin American Research Review 25(2):236-247. Barone, Myriam E. 2000. Dos Historias de Nunca Acabar. Posadas Misiones: Editorial universitaria de Misiones. Bortman, Marcelo, Guadalupe Verdejo and Juam Manuel Sotelo. 1999. Situaci6n de la Equidad en Salud. Argentina, 1999. Buenos Aires: Organizaci6n Panamericana de la Salud Argentina (OPS). Correia, Maria. 1999. Las Relaciones de Genero en la Argentina. Buenos Aires: Banco Mundial. Deere, Carmen Diana. 1982. "La Mujer Rural y la Producci6n de Subsistencia en la Periferia Capitalista." Pp 11-22 en Las Trabajadoras del Agro, Vol II. Debate Sobre la Mujer en America Latina y el Caribe: Discusi6n acerca de la Unidad Producci6n - Reproducci6n, Ed. por Magdalena Le6n. Bogota, Colombia: Asociacion Colombiana para el Estudio de la Poblaci6n, ACEP. Felder, Ruth and Oscar Oszlak. 1998. "La Capacidad de Regulaci6n Estatal en la Argentina." Pp. 165-204 en La Argentina Que Viene, Analisis y Propuestas para una Sociedad en Transici6n. Isuani, A., and D. Filmus, eds. Buenos Aires: UNICEF, Grupo Editorial Nacional. Fogel, Ramon and Pantelides, Alejandra. 1994. Determinantes Principales de la Fecundidad en areas rurales de Paraguay; El caso de Itapua. Asunci6n, Paraguay: CERI. Fogel, Ramon; Hiekel, Maria Victoria and Olarzar, Cristina. 1993. Mujeres Campesinas y Conducta Reproductiva. Asunci6n, Paraguay: CERI-CEEP. Fraad, Resnick and Wolf. 1989. "For Every Knight in Shining Armor, There's a Castle Waiting to Be Cleaned: A Marxist-Feminist Analysis of the Household." Rethinking Marxism 2(4): 10-69. Chapter2. Reproductive Health in 4rgentina's Poor Rural Areas 39 Friedmann, Harriet. 1986. "Patriarchy and Property. A reply to Goodman and Redclift." Sociologia Ruralis 26(2):187-193. Gerstel, Naomi and Harriet Gross. 1987. Families and Work (Women in the Political Economy). Philadelphia: Temple University Press. Gogna, Monica, and Silvina Ramos. 2000. "Gender Stereotypes and Power Relations, Unacknowledged Risks for STDs in Argentina." In Parker, Barbosa, and Aggleton, eds., Framing the Sexual Subject, The Politics of Gender, Sexuality and Power. San Francisco CA: University of California Press. Gogna, M., J. J. Llovet, S. Ramos and M. Romero. 1998. "Los Retos de la Salud Reproductiva: Derechos Humanos y Equidad Social." La Argentina Que Viene, Analisis y Propuestas para una Sociedad en Transici6n. Isuani, A., and D. Filmus, eds. Buenos Aires: UNICEF, Grupo Editorial Nacional. Gonzalez, Soledad y Vania Salles. 1995. "Mujeres que se quedan, mujeres que se van... Continuidad y cambio en las relaciones sociales en contextos de aceleradas mudanzas rurales." Pp.15-53 en Gonzalez y Salles (eds.) Relaciones de Genero y Transformaciones Agrarias. Mexico, DF: El Colegio de Mexico. Gwatkin, Davidson R., Shea Rustein, Kiersten Johnson, Rohini P. Pande, and Adam Wagstaff. 2000. Socio-Economic Defferences in Health, Nutrition, and Population. Washington DC: World Bank. Hartman, Heidi. 1987. "The Family as The Locus of Gender, Class, and Political Struggle. The example of Housework." In S. Harding, ed., Feminism and Methodology. Pp.109-134. Bloomington, IN: Indiana University Press & Open University Press. -NDEC. 1994. Estudios Materno Infantil. Buenos Aires: Ministerio de Economia y Obras y Servicios Publicos. Jelin, Elizabeth, ed. 1990. Women and Social Change in Latin America. London: Zed books Ltd and United Nations Research Institute for Social Development. Kelly-Gadol, Joan. 1987. "The Social relation of the Sexes. Methodological Implications of Women's History." In S. Harding, ed., Feminism and Methodology. Pp.15-28. Bloomington, IN: Indiana University Press & Open University Press. Kirwood, Julieta. 1986. Ser Politica en Chile. Las Feministas y los Partidos. Facultad Latino Americana de Ciencias Sociales. Santiago, Chile: FLACSO. Leacock, Eleonor. 1983. Myths of Male Dominance. Collected Articles on Women Cross- culturally. New York: Monthly Review Press. Mann, Susan Archer. 1990. "The Civilization of Nature." Pp.128-140 in Agrarian Capitalism in Theory and Practice. Chapel Hill, NC: The University of North Carolina Press. 40 Measurement and Meaning Nash, June. 1990. "Latin American Women in the World Capitalist Crisis." Gender and Society Vol 4(3):338-353. OPS/OMS (Organizaci6n Panamericana de la Salud and Ministerio de Salud). 1999. Indicadores Bdsicos. Argentina 1999. Buenos Aires: OPS/ Ministerio de Salud. Pantelides, Edith, Rosa Geldstein and Graciela Infesta Dominguez. 1995. Imagenes de Genero y Conducta Reproductiva en la Adolescencia. Buenos Aires: Centro de Estudios de Poblaci6n. Pantelides, Edith and Marcela Cerrutti. 1992. Conducta Reproductiva y Embarazo en la Adolescencia. Buenos Aires: Centro de Estudios de la Poblaci6n. Ramos, Silvina, and Romero, Mariana. 1998. "El aborto en la Argentina." Ginecologia y Reproduccion 6(3):108-121. Ramos, Silvina, Gogna, Monica and Romero, Mariana. 1997. "La Salud reproductiva en la Argentina: dimensiones epidemiol6gicas y socio-demograficas." In Foro por los Derechos Reproductivos, ed., Nuestros Cuerpos, Nuestras vidas: Propuestas para la promoci6n de los derechos sexuales y reproductivos. Pp.15-28. Rodriguez-Garcia, R., M. Akther. 2000. Human Rights: The Foundation of Public Health. Washington DC: George Washington Center for International Health. Schiavoni, Gabriela. 1996. Colonos y Ocupantes: Parentesco, Reciprocidad y Diferenciaci6n Social en la Frontera de Misiones. Posadas, Argentina: Editorial Universitaria de Misiones. Schiavoni, Lidia. 1996. Pesadas Cargas, Fragiles Pasos. Las Comerciantes Fronterizas dePosadas-Encarnaci6n, Misiones. Posadas, Argentina: Editorial Universitaria de Misiones. SIEMPRO. 1997. Encuesta de Desarrollo Social. Buenos Aires. Argentina: SIEMPRO/INDEC. Staudt, Kathleen. 1994. Technical Assistance and Women: From Mainstreaming Toward Institutional Accountability. UN Technical Paper, Division for the Advancement of Women. Stout, K and R. A. Dello Buono. 1996. "Birth Control and Development in Three Latin American Countries." In P. Dubeck and K. Borman, eds., Women and Work. London: Garland Publishing, Inc. Subsecretaria de la Mujer. 2000. Mujeres en Argentina. Buenos Aires: Ministerio de Relaciones Exteriores, Comercio Intemacional y Culto. Subsecretaria de Programas de Salud. 1997. Vigilancia y Control de la Mortalidad Materna. Buenos Aires: Ministerio de Salud y Acci6n Social. Weston, Kathy and Lisa Rofel. 1985. "Sexuality, Class and Conflict in a Lesbian Workplace." In Gelpi, Johnson, and Weston, eds., The Lesbian Issue p.199-222. Chicago: University of Chicago. Chapter2. Reproductive Health in Argentina 's Poor Rural Areas 41 World Bank. 2001. Together we Stand, Divided we Fall: Levels and Determinants of Social Capital In Argentina. Washington DC: World Bank. World Bank. 2000. Engendering Development: Enhancing Development through Attention to Gender. Washington DC: World Bank. World Bank. 1999. "Safe Motherhood and The World Bank, Lessons from 10 Years of Experience. " Washington DC: World Bank. World Bank. 1998. Improving Reproductive Health. Washington DC: World Bank. World Bank. 1997. An Analysis of Rural Poverty in Argentina. Washington DC: World Bank. World Bank. 1994. A New Agenda for Women's Health and Nutrition. Washington DC: World Bank. CHAPTER 3 THE TARGETING OF GOVERNMENT PROGRAMS IN CHILE Carine Clert and Quentin Wodon"l INTRODUCTION The government of Chile has been using for many years a system for the targeting of many of its income transfers and other social programs. The system is based on the ficha CAS, a two page form that households must fill if they wish to apply for benefits. Each household is attributed a score on the basis of the ficha CAS, and this score is used to determine eligibility not only for income transfers (e.g., pension assistance and family allowances), but also for water subsidies, access to social housing, and childcare centers. At the local level, municipalities also use the form for the targeting of their own programs and safety nets. Almost a third of all Chilean households have been filling the form. Taken as whole, the programs which are targeted using theficha CAS play a major role not only in the alleviation of poverty, but also in its prevention by enabling vulnerable households to receive or not state and municipal support. This papers provides an assessment of the ficha CAS system using both quantitative and qualitative methods of investigation. After describing theficha CAS system and the main income transfers and other programs which are targeted using the system, the paper uses data from the nationally representative 1998 CASEN survey to provide quantitative measures of performance for each program. Following Wodon and Yitzhaki (2000), the quantitative performance measures are based on a decomposition of the Gini income elasticity of the various programs into a targeting component which is based on who benefits from the programs and who does not, and an allocation component which captures the impact of the variability in program benefits among participants. Overall, the programs appear to be well targeted. The good quantitative performance of the program does not mean that theficha CAS is without any limitations. In order to look in some detail at these limitations, the paper relies on a study of the experience, perceptions and recommendations of poor citizens on the one hand and practitioners using the ficha CAS at the local level on the other hand (Clert, 2000a, 2000b). Evidence derives from a stratified survey of 88 randomly sampled households in the municipality of Huechuraba, a comparatively poor area in the Greater Santiago area and from qualitative interviews with a sub-sample of households. Evidence also derives from focus-group I I Both authors are with the World Bank. Comments can be sent to cclert@worldbank.org and qwodon@worldbank.org. The paper was funded by the World Bank under the Chile Poverty Assessment and the Regional Study on Extreme Poverty and Social Exclusion in Latin America. Assistance from Rodrigo Castro- Fernandez and Corinne Siaens is gratefully acknowledged. The authors are also grateful for the Chilean government's comments which improved this paper. The views expressed in the paper are those of the authors and need not represent the views of the World Bank, its Executive Directors, or the countries they represent. 44 Measurement and Meaning discussions and semi-structured interviews with professionals located in that municipality and from interviews with central government officials. The fieldwork was carried out between December 1997 and June 1998. The triangulation of household-level interviews and focus group discussions with municipal staff in Huechuraba revealed that poor households often lack information about the government programs and how to apply for their benefits. The qualitative work also revealed potential deficiencies and biases in the eligibility criteria and associated targeting methods based on the ficha CAS. While the targeting system as a whole is sound, recommendations can be made for improving its effectiveness and its fairness. The paper is divided in four sections. Section One introduces the paper by presenting some background information on theficha CAS and the targeted safety nets and other social programs reviewed in this paper. Section Two provides evidence for a quantitative assessment of the targeting of some of these social entitlements using the CASEN household survey data while Sections Three sheds light on the a more qualitative assessment of targeting methods, based on the experience and views of the poor themselves and of social practitioners. Conclusions and policy implications are provided in the last section. BACKGROUND This section first sheds light on the official means-testing instrument used by the Chilean government, the ficha CAS or CAS form. It then puts the forthcoming findings of the paper in context by presenting the key safety nets and programs reviewed in this paper and their importance in the social protection of the poor. The Ficha CAS Introduced during the military regime (1973-1989) and modified by the post-1990 democratic governments, theficha CAS' is a two page form which is used for determining the eligibility of households to a number of Government programs including not only monetary transfers (Subsidios Monetarios), but also access to low income housing and childcare centres.'3 A reproduction of the form is provided in Appendix Two. The form provides detailed information on housing conditions of the dwelling unit (e.g., material used for the construction of the housing unit, number and type of rooms, access to water, latrine and sanitary services, access to electricity, etc.); on members of the dwelling unit (their occupation, educational level, date of birth, and income. Additional information is provided on material assets held by the household (housing status, television, heating equipment, and refrigerator). Points are allocated to households on the basis of the information provided, with the number of points fluctuating between 380 and 770 points. Households with a total of inferior to 500 points are considered as extremely poor and those with a total number of between 500 and 540 points are considered as poor. The Ministry of Planning MIDEPLAN is responsible for the design of the ficha CAS. The recruitment of the employees administrating the form is done at the discretion of the municipality, but training must be provided by the Ministry. Municipalities usually separate the activities of data collection and data entry from those of needs assessment. Data collection and 12 Ficha de Estratificaci6n Social. 13 At present the official name of the form is theficha CAS-JI. Chapter3. The Targeting of Government Programs in Chile. 45 entry tend to be done by a department of social information within the municipality, while the control of the needs assessment is usually done by social workers and tecnico-sociales (welfare assistants). The national income transfer programs which are targeted using the CAS scoring system apply the formula in a strict manner in order for determining eligibility. The score obtained by a household automatically and exclusively prevails, so that eligibility depends only on the number of points obtained. The ficha is also used for targeting locally financed safety nets, but in this case social workers and other professionals can often give some weight to other eligibility criteria such as the presence of a chronic illness, the civil status of household members, and their actual financial resources at the time of request (the ficha is completed every three years, and there may be differences between the status of households when they apply for benefits, as compared to their status when they filled the form). For housing programs as well, differences can be observed in the use of the ficha at various levels of government. Professionals dealing with central government programs (viviendas basicas and vivienda progresiva) must follow the method of calculation defined by the Ministry of Housing (SERVIU), while professionals involved in municipal initiatives have some discretionary power. One of the advantages of using the ficha for many different programs is that this reduces the cost of means-testing. The cost of a CAS interview is about US$8.65 per household. The Ministry of Planning estimates that 30 percent of Chilean households undergo interviews, which seems reasonable given that the target group for the subsidy programs is the poorest 20 percent. The CAS system is used as a targeting instrument for utility subsidies, income transfers, social housing subsidy, and pension subsidies among other programs. Because the fixed administrative costs of targeting are spread across several programs, the CAS is very cost-effective. In 1996, administrative costs represented a mere 1.2 percent of the benefits distributed using the CAS system. If the administrative costs of the CAS system were to be borne by the water subsidies alone, for example, they would represent 17.8 percent of the value of the subsidies. The Targeted Programs: Their Role in Fostering Securityand Alleviating Poverty Many national and local Government programs rely on the CAS system for their targeting. Locally, Comunas generate from their own budgets other safety net programs which vary in their amount and eligibility criteria, but these cannot be evaluated with the CASEN. The national programs implemented with the ficha CAS and reviewed in this paper include means-tested pensions, family allowances, water subsidies, social housing, and child care. As developed below, most of these social entitlements play a major potential role in decreasing vulnerability and alleviating poverty, which makes the issue of targeting a crucial one. Full descriptions of the programs is provided in Appendix One. Pensions PASIS (Pensi6n de Asistencia): Means-tested state pensions are provided to elderly and/or disabled individuals through PASIS. To be eligible, an elderly individual needed to have a total income below half of the minimum pension allowance, which was CP$ 23,415 per month in 1998.14 The eligibility threshold for the disabled is the amount of the minimum pension allowance. While the income transfers provided through PASIS are low in comparison with the 14 In 2001, PASIS was worth 35104 Chilean Pesos. 46 Measurement and Meaning minimum wage, household interviews in the Comuna of Huechuraba suggest that the transfers can be significant in the eyes of those who do not have any other source of income or family support to rely upon. In addition, those who receive PASIS pensions are automatically eligible for free access to public health services through the health gratuity card. By securing an entitlement to health, PASIS thus provides low income elderly and invalid or disabled people with a key mechanism of long term social protection. Family allowances SUF (Subsidio Unico Familiar): Family allowances are important because they help in coping with the extra expenses due to the birth of children, as well as with the possibility of a reduction in earnings due to the fact that pregnant women and women who have delivered may have to stop working for a while. The loss in earnings is particularly likely for women involved in precarious jobs which do not provide them with maternal and other family benefits. The amount per child above three years of age was CP$2,500 per month in 1998, at the time the CASEN was implemented. The amount per child below three years of age was CP$2,800. Maternal benefits were also $2,800 per month for a period of ten months, with eligibility as of the fifth month of pregnancy."5 Water subsidies SAP (Subsidio Agua Potable): The water subsidy provides an allowance for the cost of consumption of up to 15 cubic meters per month. As noted by Gomez-Lobo and Contreras (2000; see also Estache, Foster, and Wodon, forthcoming, for a review), the subsidy was introduced in 1990 to reduce the impact of rising prices after the reform of the water sector. The initial take-up of the program (i.e., the percentage of eligible households participating in the program) was low in the first year, at 5 percent only, because the eligibility threshold and the value of the subsidy were too low to make it worthwhile for households to participate. To increase take-up, water companies were given the opportunity to propose customers as potential subsidy recipients, which was in their interest in order to increase payment rates. The subsidy covers from 20 percent to 85 percent of the bill for the first 15 cubic meters of monthly consumption. MIDEPLAN uses regional data on water consumption and tariffs, as well as socioeconomic conditions, to determine the funds made available to each region. Within each region, subsidies are allocated to municipalities who then determine household eligibility using thejficha CAS. As for other programs targeted with theficha CAS, household eligibility is re- assessed every three years, and the subsidy can be withdrawn by utilities if the household has more than three months of arrears in paying its share of the bill. Social housing: The programs of vivienda baisica and vivienda progresiva (etapas I -I) provide subsidies for the construction of new housing units, or the improvement of existing units. The amount of the subsidy is determined in UFs, which are monetary units.16 Apart from eligibility criteria according to theficha CAS, the households must contribute to the construction costs and thereby provide evidence of savings when applying. The program is important not only to satisfy basic needs, but also because the lack of safe and secure shelter has been shown to reduce the ability of women to work because of their reluctance to leave their children at home (risks of 15 These figures refer to the period during which the 1998 CASEN survey and the micro-study in Huechuraba were implemented. The amounts have changed since. For instance, in July 2001, the SUF was worth 3452 Chilean Pesos. 16 In 1998, one UF was worth approximately CP$460. ChapterJ. T he largeting oJ (iovernment Programs in Chile. 41 accidents in sub-standard housing, such as electrical shocks). Good housing conditions are also essential for individuals involved in home-based wage employment or micro-enterprise, and for reducing crowding and the associated social risks of domestic violence. Child care: The childcare programs of the JUJNJI and Fundaci6n INTEGRA are also targeted using theficha CAS. The programs provide care for children whose mothers are working. Since vulnerable low-income women are more likely to be affected than men by exclusion from the labor market and by poor quality of employment, this type of program for affordable or even free childcare is important from a gender point of view and for building work experience and incentives among poor women. QUANTITATIVE EVALUATION This section provides a quantitative assessment of the targeting performance of the social programs which are implemented nationally and for which we have information in the nationally representative 1998 CASEN survey (Caracterizaci6n Socioecon6mica Nacional) implemented by the Ministry of Planning MIDEPLAN. There are various ways to evaluate quantitatively the targeting performance of the programs whose eligibility is based on the ficha CAS. The most common measures of targeting performance used in the literature are based on the so-called errors of inclusion and exclusion. An error of inclusion is observed when a household which is not part of the program's target population receives the program's benefits. An error of exclusion is observed when a household which is part of the program's target population does not receive the program's benefits. This approach for measuring targeting performance has been used among others by Gomez-Lobo and Contreras (2000) for Chile's water subsidies. In this paper however, we use an alternative (and arguably better) indicator of performance which takes into account not only who benefits from social programs and who does not, but also to what extent various households benefits (the program benefits may vary from one household to another). The method is explained in box 3. 1, and it relies on three key parameters for understanding the impact of various programs on social welfare: * Gini income elasticity (GIE): The overall impact on social welfare of changing at the margin the budget allocated by the government to a given program is a function of the program's Gini income elasticity (GIE hereafter). If the GIE is equal to one, a marginal increase in the benefits will not affect the Gini coefficient in after-tax after-benefit per capita income, and thereby the effect on social welfare can be considered as neutral (no change). If the GIE is less (greater) than one, then an increase in program benefits will decrease (increase) the Gini of income, and thereby increase (decrease) social welfare. The smaller the GIE, the larger the redistributive impact of the program and the gains in social welfare. Importantly, since the GIE is estimated for a dollar spent on the program, we can compare programs which are of different scale in terms of outlays. The GIE can be decomposed into the product of a targeting elasticity and an allocation elasticity. * Targeting elasticity: The targeting elasticity measures what would be the impact of a program on social welfare if all those who benefit from the program were receiving exactly the same benefit. In other words, the targeting elasticity provides the impact of pure targeting (who gets the program and who does not) on social welfare. Lower and upper bounds can be 48 Measurement and Meaning provided for the targeting elasticity, and these bounds depend on the share of the population which participates in the program. The higher the share of participants in the population, the closer the bounds. The intuition beyond this result is that it is easier to target a program to the very poor when the share of the participants among the population is low. The practical relevance of the bounds is that they enable an analyst to compare the targeting performance of programs of different sizes. * Allocation elasticity: The allocation elasticity measures the impact of social welfare of the differences in the benefits received by various program participants. Lower and upper bounds can also be provided for the allocation elasticity. The combination of the information provided by the targeting and allocation elasticities enables the analyst to assess whether the good (bad) performance of a given program is due to good (bad) targeting or to a good (bad) allocation of benefits among participants. In the 1998 CASEN, it is feasible to estimate both the targeting and allocation elasticities for the income transfers provided by PASIS, SUF, and the water subsidies. Additionally and for comparison purposes, we also compute the targeting and allocation elasticities for another type of means-tested family allowances which is different from SUF and does not rely on theficha CAS (according to the CASEN questionnaire, these allowances provide CP$3,025 for households with gross income below CP$91,800, CP$2,943 for households with gross income between CP$91,800 and CP$186,747, and CP$1,000 for households with gross income between CP$186,747 and CP$365,399). For the housing and child care programs, the information available in the CASEN enables us to compute the targeting elasticity only because we do not have the amounts allocated (or the cash value of the in-kind benefits), but the targeting elasticity should be fairly close to the overall GIE because there are relatively few differences in benefits allocation between households in these programs (the amounts for the housing allocations are fixed, and the child care benefits only depend on the number of young children that a working mother may have). The results of the estimation are provided in table 3.1. To understand how table 3.1 works, let's consider the case of the pension assistance provided under PASIS. The table indicates that the GIE for PASIS is -0.58, which is fairly low and hence highly redistributive (any elasticity below one indicates that the corresponding program is redistributive; a negative elasticity implies a large redistributive impact). The GIE for PASIS is equal to the product of the targeting elasticity (-0.56) and the allocation elasticity (1.05). The fact that the allocation elasticity is close to one suggests that there are few differences in pension benefits among PASIS participants. In other words, the redistributive impact of the program comes from its good targeting based on theficha CAS. As for the participation rate in the program of 6.1 percent, it determines (together with the value of the overall Gini for per capita income of about 0.57) the lower and upper bounds for the targeting and allocation elasticities. For comparison purposes, other sources of pension income have been included in table 3.1 even though these are not targeted through the ficha CAS and are provided in many cases by private operators. Clearly, and as expected, the pension assistance provided through PASIS is much more redisributive than other pensions. More generally, two main conclusions can be drawn from table 3.1: High overall redistributive impact, but differences between the various programs: All the programs targeted according to the ficha CAS have a large redistributive impact per peso spent. This is evidenced by the low values of the GIE elasticities for the income transfers and Chapter3. The Targeting of Government Programs in Chile. 49 water subsidies, and by the low values of the targeting elasticities for the housing and child care programs (for these programs, we cannot compute an allocation elasticity, so that the GIE remains unknown). Yet, some programs are better targeted than others. Among income and other transfers, the SUF family allowances have the best performance, while water subsidies have a somewhat lower performance. Among the other social programs, the child care programs tend to be slightly better targeted than the housing programs, perhaps because of the savings requirements required for participation in the later. * Good targeting. with few differences in allocation: The redistributive impact of the programs is due to their good targeting, which is based on the ficha CAS. The fact that the GIE tends to be close to the targeting elasticity (because the allocation elasticities are close to one) suggests few differences in the amount of benefits received from the various programs by different households. Only in the case of water do we have an allocation elasticity well below one, probably because those who consume more water and thereby receive more subsidies tend to be higher up in the distribution of income. Table 3.1: Gini Income Elasticity of Social Pro grams Targeted According to the Ficha CAS Income transfer programs and water subsidies Non-PASIS Pension Family Water pensions assistance allowances subsidies (not targeted) PASIS SUF Gini income elasticity (GIE) 0.91 -0.58 -1.03 -0.35 Program participation rate p 15.7% 6.1% 11.5% 6.4% Mean allocation received 7634.04 503.16 155.68 47.61 Overall Gini for per capita income G, 0.57 0.57 0.57 0.57 Targeting elasticity Lower bound -1.49 -1.66 -1.56 -1.65 Actual value 0.47 -0.56 -0.95 -0.43 Upper bound 1.49 1.66 1.56 1.65 Allocation elasticity Lower bound -1.19 -1.06 -1.13 -1.07 Actual value 1.91 1.05 1.09 0.80 Upper bound 1.19 1.06 1.13 1.07 Other targeted programs Housing Housing Housing Child care Child care Viv. Basica Viv. Prog I Viv. Prog ii JUNJI INTEGRA Gini income elasticity NA NA NA NA NA Program participation rate p 5.8% 1.1% 0.2% 1.7% 1.3% Overall Gini for per capita income G, 0.57 0.57 0.57 0.57 0.57 Targeting elasticity Lower bound -1.66 -1.74 -1.76 -1.73 -1.74 Actual value at individual (per capita) level -0.41 -0.68 -0.59 -0.50 -0.71 Actual value at household level -0.32 -0.54 -0.48 -0.44 -0.65 Upper bound 1.66 1.74 1.76 1.73 1.74 Source: Authors' estimation using 1998 CASEN survey. 50 Measurement and Meaning Box 3.1: METHODOLOGY FOR THE QUANTITATIVE EVALUATION To assess the impact on welfare of government programs per dollar spent in each program, we Wodon and Yitzhaki (2000). Denote by y the mean income in the population and by G the Gini index of income inequality. A common welfare function used in the literature is W= yV (1-G). The higher the mean income, the higher the level of social welfare, but the higher the inequality, the lower the aggregate level of welfare. This welfare function takes into account not only absolute, but also relative deprivation (people assess their own level of welfare in part by comparing themselves with others). Using the implicit distributional weights embodied in this welfare function, we can derive the marginal gains from additional investments in government programs. If x denotes the mean benefit of a social program x across the whole population, and if il is the Gini income elasticity of that program (defined below), increasing at the margin the funds allocated to the program by multiplying the outlays by I + A for all program participants, with A small, will result in a marginal social welfare gain equal to: dW = ( xA)(1 -d G) (1) Equation (1) makes it clear that considerations related to both growth (as represented by the mean marginal benefit x A) and distribution (as represented by the Gini income elasticity 11 times the Gini index G) must be taken into account in program evaluations. The Gini income elasticity q (hereafter GIE) measures the impact of an increase of one dollar, distributed as a constant percentage change in the benefits of the program, on income or consumption inequality. Denoting by x the household (per capita) benefit from the program, by y income, by F(y) the cumulative distribution of income, and by x- the mean benefits of the program over the entire population, the GIE is: cov(x, F(y)) (2) cov(y, F(y)) x If the elasticity equals one, a marginal increase in benefits will not affect the Gini coefficient in after-tax after-benefit income. If the elasticity is less (greater) than one, then an increase in benefits will decrease (increase) the Gini of income. The smaller the elasticity, the larger the redistributive impact of the program and the gains in welfare. Since the GIE is estimated for a dollar spent on the program, we can compare programs which are of different scale in terms of outlays. A decomposition of the GIE can be used to differentiate between two properties of a program that can affect its impact on welfare: targeting and the allocation mechanism among participants (internal progressivity). The decomposition enables the analyst to assess whether the (lack of) performance of social programs and policies is due to either the selection mechanism for participants or the allocation of benefits among program participants. To differentiate between targeting and internal progressivity, define z as the targeting instrument: (xip if heP (3) That is, z is equal to the mean benefit among participants for households who participate in the program and it is zero for households who do not participate (one could substitute the average benefit by an indicator which is equal to one without affecting the results.) The variable z is an indicator of targeting because it is only concerned with whom is affected by the program, rather than with the actual benefit received. Using this definition of z, we can rewrite the GIE as a product of two elasticities as follows: (cov(z, F(y)) y~ cov(x, F(y)) z = 74A cov(y, F(y)) -z ) cov(z, F(y)) x ( Chapter3. The Targeting of Government Programs in Chile. 51 Box 3.1 (Continued) The first term is the progressivity among participants (allocation effect). The second term is related to the targeting of the program (targeting effect). The distributional impact of a program depends on the product of its targeting and allocation elasticities. Good targeting, for example, can be offset by a bad allocation mechanism among program beneficiaries. Equation (6) is useful to assess whether the (lack of) performance of a program is due to its targeting or to the allocation of benefits among beneficiaries. But one can go further by establishing bounds for the values of the targeting and allocation elasticities. Specifically, the minimum and maximum values of the targeting elasticity depend on the share the population participating in the program and the overall Gini. Denoting by p the share of the population participating in the program, and by Gy the overall Gini, it is shown in the appendix that: - ( P) S 77T BS' where between-area variance is increasing ii) WS2