79713 Improving Gender and Development Outcomes through Agency POLICY LESSONS FROM THREE PERUVIAN EXPERIENCES Elizaveta Perova and Renos Vakis Improving Gender and Development Outcomes through Agency POLICY LESSONS FROM THREE PERUVIAN EXPERIENCES Elizaveta Perova and Renos Vakis © 2013 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org Banco Mundial, Oficina Lima, Perú Av. �lvarez Calderón 185, piso 7, San Isidro - Lima 27 - Perú Telephone number: +51 1 622 2300 www.bancomundial.org.pe This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Authors: Elizaveta Perova, Renos Vakis Cover and layout design: Florencia Micheltorena First edition: July, 2013; 500 copies Printed in Peru by GALESE S.A.C Cayetano Heredia 839, Jesús María The Legal Deposit entered at the National Library of Peru N° 2013-09737 ISBN: 978-612-46470-0-0 Follow us on: /bancomundial @worldbanklac Table of contents Foreword 6 Preface 8 Acknowledgements 9 Introduction 10 What is agency? 13 How do we know? Research behind the lessons 17 Policy lessons Policy lesson 1. Starting small and going slowly helps to develop agency and reinforce other impacts of the programs 22 Policy lesson 2. In gender interventions, think of both genders! 30 Policy lesson 3. It pays to socialize! 34 Policy lesson 4: Managing your own resources creates agency 44 Policy lesson 5. Clear, accessible, and continuous information campaigns for beneficiaries can increase agency and final outcomes 50 Policy lesson 6: Involving the community at large may have an added positive impact on agency and on overall program outcomes 56 Conclusion 61 References 63 Appendix Appendix 1: Agency: what do we understand by it, and why does it matter? 67 Appendix 2: Overview of interventions analyzed in this study 69 Appendix 3: Overview of studies used as background research 73 Appendix 4: More equitable access to care - the ISIAT project 83 Appendix 5: The role of repeated peer accompaniment 84 Foreword Probably, all of us can emphasize with frustration caused by failure to do simple things with presumably high payoffs. We have all experienced procrastination, inability to stick to resolutions, lack of motivation at least once in our lifetime, and know firsthand what a strong obstacle they may be for achieving goals, acting on our desires, turning life plans into reality. We may also remember in amazement how powerful equally simple things - an encouraging word, support from friends, seeing a street sign that resonates with our aspirations - can be in overcoming these and other, more tangible obstacles. Personal experience is likely to attest to the tremendous significance of intangible drivers and snags in our lives. Would not our choices be wiser if only it were possible to boost determination with a dose of encouragement, or to curb procrastination with some magic bullet? 6 In recent years, academics and policy makers alike have been trying to understand how to unleash the power of these intangible drivers and snags behind human behavior. Neuroscientists do brain imaging to understand, for example, how stress impairs our abilities. Behavioral economists examine the role of default options or perceived community norms in decision making. Their work has been increasingly used by policy makers. The British government has made the most foray into the field, setting up in 2010 the Behavioral Insights Team, often called the ‘Nudge Unit’, to apply insights from academic research in behavioral economics and psychology to public policy and services. Improving Gender Outcomes through Agency: Lessons from Three Peruvian Experiences belongs to the same intersection of research and policy work. It uses research on agency to forge hands-on recommendations for policy practitioners. There is an overlap between behavioral drivers and snags and the concept of agency. Agency, following the definition of the World Development Report (WDR) 2012, is an ability to make choices to achieve desired outcomes. Through focusing on agency the book describes behavioral barriers that frequently stand in the way to achieving one’s dreams: lack of aspirations, self-esteem, or community support, discrimination or inequitable gender norms. Most importantly, it shares observations about how three Peruvian programs have successfully overcome some of these barriers. Written as a handbook for Peruvian policy practitioners, the book attempts to disentangle behavioral mechanisms behind the effects of three programs: Juntos, Corredor Puno-Cusco and Innovative Socio-Economic Interventions against TB (ISIAT). Was a boost in self-esteem at the beginning of the Corredor Puno-Cusco partially responsible for its success? Does nurturing the sense of being part of the community improve adherence to TB treatment in the ISIAT Pprogram? Would a different targeting scheme protect Juntos beneficiaries from discrimination within their communities and hence improve the outcomes? These are some of the questions the book attempts to answer. Most importantly, it extracts policy lessons that are applicable and can be experimented with in other settings. The authors argue that enhancing agency may be a powerful and cost-effective catalyst for improving the primary objectives of the programs: poverty reduction, lowering TB incidence, or improving financial literacy. Based on three Peruvian experiences, they offer practical and 7 easy suggestions of potential tweaks to the programs that can increase beneficiaries’ agency, and overall outcomes of the program. Hopefully, these suggestions will find their audience, and will prove useful in ‘tweaking’ programs and interventions – for greater agency of the beneficiaries and better outcomes. Susan Goldmark Country Director for Bolivia, Chile, Ecuador, Peru and Venezuela World Bank June 2013 Preface Peruvian public policy is currently focused on economic growth with social inclusion. The Ministry of Development and Social Inclusion (MIDIS) – created in October 2011 – leads the sector and promotes evidence-based public policy using three strategic guidelines – (i) Matching criteria and mechanisms for the selection of areas and target population, (ii) generation of instruments for inter-sectorial and inter-governmental result-based coordination, and (iii) activation of monitoring and evaluation procedures to measure interventions’ progress and results. These guidelines have been set out in the National Strategy of Development and Social Inclusion “Incluir para Crecer�. The Strategy is MIDIS’ main managerial tool and aims at guiding social inclusion-related interventions under a life cycle approach and integrating ongoing results-oriented sectorial initiatives, as well as those of regional and local governments. 8 To achieve its objectives, MIDIS promotes citizen-centered policies and social programs. Under this approach, the notion of agency – understood as the ability people have to make decisions to achieve the results they expect – is central. When discussing social inclusion interventions, several actors are involved in actively carrying out efforts towards life quality improvement. Among them, women are key agents. Their contribution, following Moser’s framework, is associated to the triple-role they play – reproductive, productive and community-managing work. Identifying experiences and deriving lessons about these contributions is an invaluable input for policy formulation. That is the main contribution of Improving Gender and Development Outcomes through Agency: Policy Lessons from Three Peruvian Experiences. The book highlights specific lessons about progressiveness in program implementation, community involvement, resource management, gendered approaches, information campaigns and social interaction. Bringing the notion of agency at the core of public policy formulation is both an opportunity and a challenge. In MIDIS we believe that this is an approach that needs to be promoted, institutionalized and consolidated. Putting people at the core of policy formulation is a way to guarantee the sustainability of social inclusion efforts. We are pleased to say that work has begun. We recognize the progress made to date and take responsibility for the challenges ahead. Carolina Trivelli �vila Minister of Development and Social Inclusion June 2013 Acknowledgements This study was written as a part of the Economic and Sector Work by Poverty, Gender and Equity Group in the Latin America and Caribbean Department of the World Bank. The authors of the study, Elizaveta Perova (Economist) and Renos Vakis (Lead Economist) are deeply grateful to Carlton Evans (Principal Investigator, Innovation for Health and Development) and Johanna Yankari (Coordenadora del Proyecto Capital, Instituto de Estudios Peruanos) for generously sharing their knowledge and insights on Innovative Socio-Economic Interventions against Tuberculosis (ISIAT) and Corredor Puno Cusco programs, respectively. Miriam Muller provided invaluable research support on the theoretical underpinnings of the concept of agency. The authors would also like to thank personnel and beneficiaries of the Juntos, Corredor Puno Cusco, ISIAT and Sierra Sur programs, and the PRISMA team, who have generously shared their time to participate in the interviews or to support them, thus making the qualitative component of the study possible. 9 The study has greatly benefitted from numerous revisions, comments and support from Sue Goldmark (Director, Andean Countries, World Bank), Louise Cord (Sector Manager, Poverty, Gender and Equity Group, World Bank) and Carlos Silva (Sector Leader, Poverty Reduction and Economic Management, Andean Countries, World Bank). The authors would also like to express gratitude to Pablo Fajnzylber (World Bank), Gilette Hall (Georgetown University), Johannes Hoogeveen (World Bank), Maria Beatriz Orlando (World Bank), and Nistha Sinha (World Bank) for their insightful comments. MIDIS staff and participants of the Social Inclusion conference at the Catolica University provided invaluable early feedback and suggestions. The study was translated by Ivonne Treneman. Samantha Lach and Alan Jose have provided excellent editorial support in English and Spanish, respectively. Florencia Micheltorena designed the cover and did layout of the book. Cristina Cifuentes, Karem Edwards, Tatianna Schlottfeld and Claudia Gonzalez Guillen provided excellent administrative support. Publishing the document would have been impossible without Sandra Arzubiaga, Maria Elena Guillen and financial support from Gender and Development anchor unit of the World Bank. Introduction T his study is about the incredible and frequently underestimated role of agency - the ability to make choices to achieve desired outcomes1 - in economic development. The authors share the view that agency has inherent value for development: it is an attribute and manifestation of development, or using Sen’s words, it is constituent to development2. It is also one of the principal dimensions of gender equality3, which is a core development objective in itself. This study however, focuses on the instrumental role of agency for more tangible manifestations of development, such as, poverty reduction and economic growth. It attempts to show that expanding individual agency is a powerful catalyst for improving welfare, as measured by these concrete and widely used metrics of policy success. Moreover, it argues 10 that in many cases, improving development outcomes through agency is highly cost-effective. This study centers on several policy initiatives in Peru, which as will be subsequently shown, have improved the agency of their beneficiaries (although the effects of these interventions were by no means limited to greater agency). Yet, before zooming in on modern-day Peru, the attention will turn to the classical literature for an example that brilliantly captures the main thesis of this study – the story of Jean Valjean, the main character in Victor Hugo’s Les Miserables. On his first day as a free man, after 19 years of imprisonment for stealing bread to feed the children of his poor sister, Jean Valjean is forced to sleep on the streets. Although he has paid his dues to the law, none of the inn-keepers of Digne wants to shelter an ex-convict. Valjean is picked up from the streets by the bishop of Digne, who invites Valjean to spend the night in his house. Disillusioned and embittered by the treatment he has received earlier, Valjean returns the bishop’s kindness by stealing his silverware during the night. After Valjean is captured, and dragged back to the house of the bishop, the bishop announces to the police that the silverware was given to the ex-convict as a gift. He chastises Valjean on leaving early, and forgetting in haste the other part of the gift – even more valuable silver candlesticks. 1  World Development Report, 2012. 2  Sen, 2000. 3  World Development Report, 2012. Valjean is released by the police with the bishop’s generous bequest, stunned by the man’s ability to see a deserving human being in what appeared to be a detestable ex-prisoner to everyone else. From this encounter, Valjean embarks on an odyssey to become a prosperous entrepreneur; a caring employer who sets up fair salaries and generous benefits; a progressive mayor of Montreuil-sur-Mer; and a munificent benefactor of the poor. Such are the positive effects of his entrepreneurship and charity that Valjean is offered the award of the highest distinctions in 19th century France – the Legion of Honor (which he declines). Would the life of Valjean take the same trajectory, had not it been for the fateful encounter with the bishop of Digne? The ripples of this encounter were multiple and far-reaching, bringing up positive change in the lives of thousands: from the workers of Valjean’s factories in Montreuil-sur-Mer to numerous individuals he comes across in his arduous journey through Les Miserables. Admittedly abstracting from many layers of complexity in Hugo’s novel, let us carry out a thought experiment. Imagine a contemporary policy analyst, fascinated by the effect of the bishop’s “intervention�. It would not take long to conclude that the “intervention� has hugely improved social welfare while being highly cost-effective. The more difficult question is to define the nature of the intervention (beyond bequeathed candlesticks and silverware). By seeing more in Valjean than an ex-convict destined to a prompt return to prison, the 11 bishop profoundly changed the man’s aspirations and attitude to life. By bequeathing him silverware and candlesticks, the bishop gave Valjean the resources to act on newly found aspirations for a more purposeful life. The encounter between the bishop and Jean Valjean is probably one of the most striking literary examples of agency. The term agency belongs to the jargon of several fields and habitually enters everyday speech. Hence, some clarification on the definition used in this study is due. Following the World Development Report (WDR) 2012, agency is defined here as the ability to make effective choices and to transform those choices into desired outcomes. The process involves both a psychological and a material component. First, the individual needs to believe in the existence of choices; and second, the choices are only valid if he or she has the means to act upon them. By recognizing Valjean’s dignity and believing him capable of living a crime-free life, the bishop tremendously expanded the man’s horizon of choices. From that encounter, Valjean’s vision of his future included ‘living an honest and purposeful life’ as an option. But let us not downplay the importance of the candlesticks: would Valjean be able to act on his resolution without the starting capital that the candlesticks afforded him? This study traces remarkable transformations of life trajectories in Peru, similar to Valjean’s, made possible through social policies and interventions. These initiatives have something in common with the bishop’s intervention: they affect the beneficiaries’ vision of themselves; and give them practical means to act upon the newly founded vision. The first – psychological – component of agency is purposefully incorporated into the design of the policies and programs analyzed only in some cases. Generally, while many social programs address material obstacles standing in the way to a better future, such as lack of capital or assets, very few aim explicitly at also tackling emotional obstacles, such as lack of self-esteem or the ability to aspire. Interventions aimed at enhancing the psychological component of agency – for instance, raising aspirations - rarely enter policy debate despite their tremendous potential and cost-effectiveness. Lack of information may be one of the reasons. Indeed, rigorous evidence showing that in order to reduce poverty it makes economic sense to increase self-esteem and aspirations is just beginning to emerge4. The purpose of this study is twofold. First, it aims at bridging this information gap, providing a review of evidence that shows how the psychological components of agency, such as aspirations and self-esteem, can effectively contribute to more traditional development objectives—ranging from higher investments in human capital to increased income. Second, the study reviews and synthesizes research on several policy interventions in Peru, which have empowered their beneficiaries. In this way, the study aims to derive practical 12 recommendations on how to incorporate psychological elements of agency into policy interventions in order to achieve better development outcomes. The study is structured as follows: the next section discusses the concept of agency, providing examples of its broad role in achieving development objectives. The following section reviews the quantitative and qualitative research that served as the basis for this study and elaborates on the methodologies used to derive the conclusions presented in the ensuing section. The last section synthesizes the conclusions of the review of different interventions in Peru into six practical “policy lessons�. 4  See, for example, Macours and Vakis (2009), Appadurai (2004) or Raj (2006). What is agency? T he World Development Report 2012 defines agency as “the ability to make effective choices and transform these choices into desired outcomes�. More specifically, it can be understood as “the process through which women and men use their endowments and take advantage of economic opportunities to achieve desired outcomes� (WDR, 2012)5. This very definition implies two dimensions: (i) the internal motivation to make a choice, the willingness to act upon one’s desires (internal aspect of agency); and (ii) the absence of unsurpassable exogenous constraints (external aspect). Consider an example of two women who work in subsistence farming, and who do not diversify their production into cash crops for two distinct reasons. One woman lives in a village where the growing of cash crops is considered too fancy; a general disbelief in the possibility of success of such an undertaking 13 prevails; and information on techniques for growing cash crops is difficult to access. Discouraged by the skepticism of her neighbors, she never takes advantage of a nearby microfinance NGO that offers credits for buying seeds. The other farmer has been introduced to the advantages of cash crops by an NGO. She received inspiring information about similar farmers improving their livelihoods through the diversification of their production, as well as practical information on planting and growing techniques. The farmer is willing and eager to try, but she has neither the resources to buy expensive seeds nor access to credit, and continues planting the traditional variety of crops. In the first case, the absence of the internal component of agency prevents the farmer from acting on her desires; while in the second case, the missing external component is the culprit. In the example of Jean Valjean from the introduction, the gesture of the priest set in motion the internal component of agency. The money Valjean received from selling candlesticks and silverware, on the other hand, removed external (budget) constraints to starting his enterprise, thus enhancing the external aspect of agency. It is important to keep in mind that in order to successfully improve agency, both of its components - internal and external – should be addressed. This may happen through a single intervention aimed at both components or through a combination of interventions and/or enhanced external factors that complement each other. 5  Appendix 1 provides more information on the literature on agency, its relation to concepts such as empowerment, aspirations, and gender equality. Consider the story of Heidi. Heidi is an indigenous woman in her early twenties living in a community about an hour south of Cusco. Heidi comes from a poor family, and she was hoping to change the life of her family for the better by starting her own business – a corner store that sells groceries. In the hope that this dream would someday be realized, she was converting her meager savings into the raising of cuys6 - a highly unreliable investment strategy, unlikely to generate sufficient resources to change the status quo. In 2008, the Cusco branch of Credinka Bank launched a product aimed at poor rural women – a savings account that did not require a large opening deposit. Opening a savings account could grant Heidi the access to the formal financial system, raising the possibility of receiving a credit in the future. As such, this might constitute a much-needed jumpboard that could enable her to act upon her dream. Nevertheless, Heidi did not pursue the opportunity to join the formal banking system. To her, as to so many of her fellow residents of Pomacanchi, the very idea of entering a bank seemed surreal, something from a different world. Everything changed when a financial literacy program, Corredor Puno Cusco, arrived to Pomacanchi with a series of charlas de autoestima (self-esteem talks) and informational events. Corredor Puno Cusco organized a series of events where women of Pomacanchi could learn more about their rights, including the right to access the financial system. The highlight of these events for Heidi, however, was seeing other rural women from neighboring 14 centros poblados (communities) talk about their experience using savings accounts. These women were very similar to Heidi – they were all indigenous, rural, some not even literate – but they had been regularly making deposits and receiving interest on their savings. They were not intimidated by the formal banks and their saving strategies had yielded a multitude of positive results, such as paying for children’s education or investments in productive activities, which ranged from sewing machines to elaborate souvenirs for tourists, to brick making machines to improve existing business. The charlas de autoestima helped Heidi consider the opening of a savings account as a possibility completely open to her. Although Credinka had been offering savings accounts for rural women for a while, hearing somebody tell her that it was in the realm of her options, as well as seeing other rural women doing it, helped Heidi to actually take advantage of the opportunity. Three years after opening her first savings account, by the time she was interviewed for this project, Heidi owned three corner stores that sell groceries. She had also saved enough to buy a motorcycle, which she rents out as a moto-taxi. Her next task is learning to drive it. Her next aspiration is to expand her three grocery stores into a local chain, and she is absolutely confident in her ability to do so. Heidi’s story illustrates the two aspects of agency. Credinka’s introduction of a new product, specifically targeted at poor rural women, addressed the external component of agency. 6  Cuys or guinea pigs are a local delicacy and are frequently raised as livestock by rural families in the region as a source of supplementary income or savings. Indeed, without an actual possibility to open a savings account, Heidi would not have been able to join the formal banking system, no matter how much resolve or ambition she had. Nevertheless, the mere possibility to join the banking system did not have much effect, until she attended the charlas organized by Corredor. The financial literacy talks triggered a change in Heidi’s perception of her compatibility with the banking system, shifted her view of her rights as a Peruvian citizen, and imbibed her with self-esteem and ambition. In other words, the mere lifting of external constraints did not have much effect, until the internal component of agency was addressed by Corredor’s inspirational interventions. It is the dual nature of agency - and the subsequent importance of simultaneously lifting external and internal constraints - that makes it a powerful vehicle, allowing people to take advantage of economic opportunities and assets towards better outcomes. There is another important element in Heidi’s story. Credinka’s innovation proved to be highly successful. By targeting poor rural women the bank opened a new market segment, a highly profitable one. Despite having insufficient resources for large deposits, rural indigenous women proved highly responsible in managing their accounts, with very low rates of default. Credinka expanded its presence in rural areas, and started a credit line also targeted at poor rural women – crediwarmi. Moreover, other banks followed suit, and by the time Heidi opened her third grocery store, she and other residents of Pomacanchi could choose between several banks to open accounts all competing for their savings. Competition lowers interest rates, while the penetration of the banking sector into rural areas gives a boost to local economic 15 activity. Thus, the enhancement of the individual agency of women like Heidi can have a multiplier effect, highly beneficial for the local economy as a whole. Heidi’s story is a perfect miniature illustration of the important thesis of the WDR 2012: agency is smart economics. Of course, Heidi’s story is presented here without much rigor: an inquisitive reader may question the assumption that the charlas de autoestima were indeed a prerequisite in order for Heidi’s dreams to materialize. Maybe simply addressing what this study refers to as the external component of agency - within the realm of traditional policies aimed at lifting material constraints - would suffice? Fortunately, over the last few years new experimental work has provided some evidence on the value of agency. Box 1 summarizes the experience from a program in Nicaragua where the improved aspirations led to higher program impacts. Beneficiaries received productive grants, and while incomes increased as a result of the program, income increases among beneficiaries whose aspirations were affected positively via interactions with local leaders were much higher. What is more, the cost of addressing the internal component of agency was very low (it was not even planned!). In this way, similar to the example of Heidi, lifting internal constraints on agency proved to be highly cost-effective7. These simple examples trigger an important question: why do most programs fail to take seriously the removing of internal constraints when the potential is huge and the cost is low? 7  Macours and Vakis, 2009. Box 1. The positive impact of lifting internal constraints on agency In a recent experiment in rural Nicaragua (Macours and Vakis, 2009), a group of beneficiaries who participated in a productive grants program was found to have interacted more with nearby inspiring leaders who were also participating in the program. These interactions helped lift internal constraints on agency by increasing aspirations. The researchers were able to separate the overall program impacts from the effects for the group of beneficiaries that interacted more with leaders and found something surprising. 9 months after the program started, while the overall incomes of beneficiaries receiving the productive grant increased relative to those in the control group, the incomes of beneficiaries who interacted with leaders increased by almost twice as much! In other words, the increase in aspirations helped to double the program impact. Moreover, these interactions with leaders were not even planned in the original program design! 16 As mentioned earlier, one of the reasons why traditional policies address only the external component of agency - neglecting the benefits of simultaneously setting in motion its internal component - is the lack of information. Evidence on the value of expanding agency, which requires addressing both its components in conjunction, is just starting to emerge. Furthermore, even if a policy maker is willing to champion “applied agency�, and introduce interventions aimed at lifting internal constraints on agency, she may lack resources on how to do so. How do we know? Research behind the lessons T he objective of this study, as stated earlier, is to bridge knowledge gaps regarding the potential and cost-effectiveness of agency-enhancing interventions. By agency-enhancing interventions the study refers to the policies and programs that simultaneously address both the external and internal components of agency. Not only do these interventions lift external constraints to achieve desired outcomes, such as lack of education, skills, capital, access to services or resources; they also address internal constraints, for instance low aspirations and self-esteem. Given that policies have traditionally focused on removing external constraints, the study places greater emphasis on addressing the internal constraints, such as increasing aspirations through social interactions or improving self-esteem and self-efficacy through psychological support. 17 To bridge informational gaps, the study relies on the quantitative and qualitative research of three interventions that were launched in Peru over the last decade. The initiatives refer to the Corredor Puno-Cusco financial literacy program; the Innovative Socioeconomic Interventions against Tuberculosis (ISIAT) project – a conjunction of socioeconomic interventions in communities affected by tuberculosis (TB) and the national Juntos Conditional Cash Transfer (CCT) program. As briefly described in Box 2, these programs were not only successful at reaching their objectives, they also contributed to enhance the agency of their beneficiaries. As the Corredor Puno-Cusco program included more elements directly addressing psychological component of agency, the narrative is centered on this program, while referring to the other two programs to provide supporting examples. To demonstrate the changes in agency and understand the mechanisms behind them, this study takes advantage of the existing quantitative research, supplementing them with rich qualitative work. Such a combination of qualitative and quantitative methods is useful to really dig into the essence of the changes in agency triggered by these programs. As illustrated in Box 3, the measurement of agency is not without challenges, and a combined approach is usually the best. Box 2. Juntos, Corredor Puno-Cusco and Innovative Socio-economic Interventions against Tuberculosis (ISIAT) programs The present study draws from the evaluations of three programs implemented in Peru during the last decade: the Juntos CCT program, the Corredor Puno-Cusco financial literacy program, and the Innovative Socio-economic Interventions against Tuberculosis (ISIAT) project. The existing research suggests that these programs – briefly described in the Appendix – have achieved their objectives. Juntos has increased welfare outcomes of beneficiary families, as well as their utilization of medical services (Perova and Vakis, 2009). The Corredor Puno- Cusco program improved the financial inclusion of rural indigenous women (Trivelli and Yankari, 2008); while ISIAT interventions have shown positive impacts on a number of TB related outcomes (Rocha et al., 2011). While these programs were successful in achieving their goals, the evidence suggests that they also contributed to enhancing the agency of their beneficiaries. In some cases, the programs jointly lifted external and internal constraints. In other cases, the programs only affected one component of agency – external or internal – and the beneficiaries’ agency 18 expanded due to a joint effect of the program and some additional factor, which addressed the other component. Box 3. Measuring Agency – a combined quantitative and qualitative approach Agency is notoriously difficult to measure. Most of the studies on agency rely on proxies or manifestations of agency (or its absence) in the analysis of the phenomenon. Such proxies include, for example, domestic violence, teenage pregnancy, and control over resources. There is extensive theoretical literature dedicated to the measurement of agency (for example, Alkire, 2008) but very few of these measures are included in the large scale national surveys, which generally provide data for quantitative work. The limitations of data on agency imply relying on a combination of methods. Employing the more crude manifestations of agency or of its absence allows access to the extensive survey data and subsequent use of econometric techniques to establish causal effects. Domestic violence is one of the most salient examples of such manifestations of the lack of agency: it is difficult to argue that a person in a situation of domestic violence has much freedom to execute choices to achieve desired outcomes. Qualitative work, on the other hand, makes it possible to refine the measures of agency, to probe beyond its most striking manifestations. Nevertheless, the smaller sample sizes and shorter timeframes characteristic of qualitative 19 research preclude drawing conclusions regarding causal effects. Understanding the impact on agency from the evaluations of the three Peruvian programs: The present study relies on a combined quantitative/qualitative approach to draw policy- relevant conclusions on agency. It extracts information from several quantitative evaluations of Juntos, which establish the effect of the program on domestic violence and the program’s overall impacts. Administrative records and surveys carried out under the ISIAT project represent another source of quantitative information. To complement the quantitative research – which relies on very crude indicators of agency – the study also takes advantage of qualitative surveys from the beneficiaries of the financial literacy program Corredor Puno-Cusco. Finally, a series of interviews have been carried out with the beneficiaries of Juntos and Corredor Puno-Cusco specifically for this study. These interviews served as an important source of insights on agency and its potential to change people’s lives. The Appendix provides a brief description of the studies used, which include the specific questions addressed, methods and a summary of the results. Box 4: Economics, agency and historical institutions It is well established that the present day economic outcomes are strongly impacted by history (Nunn, 2008; Banerjee and Iyer, 2005; Glaeser and Shleifer, 2002). This is particularly true in the case of Latin America, where extractive historical institutions negatively influenced long-run prosperity (Acemoglu et al, 2002). Arguably, extractive institutions in Peru have had a much more negative impact on the indigenous populations, who are the target population of some of the programs reviewed (Corredor), or form a significant fraction of beneficiaries (Juntos). The example of mita, a forced mining labor system, illustrates this point: beginning 1573, indigenous villages located in contiguous regions to Potosi and Huancavelica mines, were required to provide one seventh of their adult male populations as rotating laborers to these mines. The mitayos were receiving below subsistence wages for extremely difficult and dangerous work, where the risks of death from mercury poisoning may have been nearly as high as the risk of death from brutal labor. The effects of the mita system proved to persist for centuries: rigorous econometric evidence 20 suggests that up to date the heritage of mita lowers household consumption by nearly a quarter, and increases the prevalence of stunted growth in children by around six percentage points (Dell, 2011). Although there is no evidence on the effects of historical institutions like mita on agency of contemporary Peruvians, it is highly plausible that they played an important role. For example, the eventual collapse of the mita system in 1812 was followed by extensive confiscation of peasant lands, peasant rebellions in response to it, banditry and livestock rustling that spanned the end of the 18th and the 20th century. Such violent events are likely to have long-lasting impacts on perceptions of self-efficacy, or to trigger feelings of stress and despair which may have long-lasting impacts beyond one generation. The mita system also had effects on the provision of public goods, such as road networks, which may affect relations within and between communities. Male-female ratio in mita districts was 22 lower than in districts not subject to forced conscription – an important demographic imbalance that could plausibly affect gender stereotypes and roles. It is beyond the scope of this study to disentangle the role of historical institutions on agency of contemporary Peruvians, especially of indigenous heritage… However, it is important to recognize the presence of long-lasting ripples of Pizarro’s landing while trying to understand agency in present day Peru. The study is thus supported on a combination of statistical methods applied to nationally representative data together with in-depth qualitative interviews, as a magnifying glass to understand the mechanisms behind the changes in agency. The analysis of these mechanisms forms the basis for concrete policy lessons, which may be applicable to other programs and have potential not only to increase agency of their beneficiaries, but also to improve final outcomes. It is important to note that the very richness and multidimensionality of the concept of agency preclude this study from providing an exhaustive description of factors that contribute to its various manifestations. For example, Peru’s historical institutions that systematically disfavored indigenous populations may continue to affect agency of contemporary Peruvians (Box 4 provides an example of one such institution8). Rather than futilely attempting a complete description of all contributing factors, the following sections focus on mechanisms that can be affected by concrete policy actions. 21 8  Box 4 is largely based on Melissa Dell’s 2011 paper. POLICY LESSON 1 STARTING SMALL AND GOING SLOWLY HELPS TO DEVELOP AGENCY AND REINFORCE OTHER IMPACTS OF THE PROGRAMS Former beneficiaries of the financial literacy program Corredor Puno Cusco in Pomacanchi are undoubtedly experienced users of the financial system. They confidently talk about shopping around for credit with different banks and comparing the interest rates. They discuss the purchases they made using credits – mostly productive inputs, such as sewing machines, motorcycle taxis or equipment for brick production businesses. Given that most of these credits have already been paid off, they make plans about their next productive investments. They have strong opinions about different lenders and products, available in the Cusco area, and give well-grounded arguments for sticking with a specific one or dropping another.9 “We get loans, we know whether they give us a good rate, we defend ourselves. In the past we did not know and were losing opportunities, but not anymore, now it is easier. (…) I know more, I know how to save, I know how to get a loan, I know many things about the bank and the money.10� Fausta, the Corredor program These women exude confidence. It is almost hard to believe that only four years ago their knowledge of the financial system was nearly non-existent and that the prospect of entering a commercial bank seemed as intimidating as surrealist. 23 Four years ago, when the promotoras of the Corredor Puno Cusco came to Pomacanchi for the first time, they were met with profound skepticism. The previous encounter of Pomacanchi residents with the financial system ended up in a debacle: credits from Programa del Banco Agrario, promoted among the rural poor in the early 2000s did nothing except leaving residents heavily indebted. This was not surprising: credit products were complex to understand for the residents of Pomacanchi, who had never had any previous interaction with the financial system. The raising of cuys served as their principal savings medium, while basic financial literacy terms, such as interest rate and principal, probably sounded like gibberish. Indeed, the diagnostic survey before the implementation of the program showed that 47 percent of women in the target group kept their savings at home, 31 percent used small animals as savings, 16 percent did not save at all, and 6 percent saved by lending out informally (Figure 1). In the end, the program was retracted, leaving negative impacts – greater indebtedness and a deeper distrust in the financial system and government programs. 9  This snapshot from the interviews with selected beneficiaries is fully consistent with the survey evidence: a follow-up survey with the beneficiaries administered in 2010 suggests that 40 percent of them have used credit services after the termination of the program. Of these, 28 percent have obtained more than one credit. Credits are not the only financial product that former beneficiaries feel comfortable with. The survey showed that 23 percent of them have used insurance or micro-insurance products, mostly to protect themselves in the case of accidents or death of a family member. Additionally, the survey confirms a sound knowledge of other financial entities (for more detailed results of the survey, see the corresponding background paper). 10  Original quote: Sacamos préstamos, sabemos si nos dan más o menos intereses, nos defendemos. Antes no sabíamos y perdíamos oportunidades, ahora no, ahora es más fácil. (…) Sé más, sé ahorrar, sé sacar créditos, sé muchas cosas del banco y de la plata. What made the experience with Corredor different? Indeed, the evidence shows that Corredor succeeded in encouraging the rural residents of Pomacanchi to turn into successful users of credit. This was a particularly impressive achievement, especially given that, unlike the case of Programa del Banco Agrario, promoting the use of credits was not one of the intended objectives of the program. Probably the very fact that the Corredor program started out with much less grandiose goals, compared to Programa del Banco Agrario, constitutes part of the answer. Corredor’s initial objective was much simpler – introducing beneficiaries to the financial system through the most basic financial product: savings accounts. Indeed, it started “small�. Additionally, the program provided continued accompaniment to its beneficiaries. For example, the charlas took place on a monthly basis for over a year. The program also organized and facilitated “savings groups�, where the owners of individual accounts could share their experiences, socialize and support each other. There were monetary incentives to save that were linked to specific periods of time: for example, beneficiaries received an award for keeping the balance of their accounts positive for three months. This also generated continuous engagement. Thus, the program spent a significant amount of time and effort on ensuring that the beneficiaries were “internalizing� knowledge and skills – or “going slowly�. Starting small and going slowly – in other words, starting with small objectives and ensuring 24 their internalization by the beneficiaries - contributed to the success of the Corredor Puno- Cusco program by expanding the agency of their beneficiaries. Amartya Sen emphasizes that once poor people become “agents� of their own destiny, they generally do a good job of getting out of poverty themselves. Corredor was successful in making the beneficiaries Figure 1: How did you save money before joining the savings program? Did not save 16% Saved by lending to third parties 6% Saved at home 47% Saved in small animals 31% Source: survey of the bene�ciaries of the Corredor project, 2010; authors’ calculation. internalize the main objective of the program – financial inclusion – as a desired outcome, and equipped them with the necessary skills to become partners in reaching this objective. What was the contribution of “starting small and going slowly� to this process? Interviews with beneficiaries suggest the following four channels. Mastering basic skills helped beneficiaries feel more prepared to undertake more complex tasks Unlike previous attempts to introduce rural beneficiaries to financial programs in the region, Corredor made this introduction a truly gradual one. Additionally, it did not limit the initiation into the world of financial services purely to teaching financial skills. The Corredor program defined the skills necessary for a successful integration into the financial system in a much broader sense, including self-esteem, civil consciousness, liquidity and risk management, dealing with formal documents and administrative processes, among others (see Box 5). Such broad coverage of multiple topics successfully achieved an important objective. It broke down a vast unknown into a sequence of (more) manageable tasks. Successfully handling a specific task, as related by the beneficiaries, prepared them for the next step, helping to limit their frustration and subsequent belief that they were simply not meant to partake in the 25 official financial system. Importantly, the snowball effect of mastering small tasks did not end once the objective of the Corredor program (ensuring that the beneficiaries could successfully manage a savings Box 5. Broad coverage of topics in Corredor’s workshops The Corredor program started with a series of workshops, which focused on 8 themes : 1. Development of stronger self-esteem and organizational skills 2. Civic consciousness and consumer culture 3. Discussions of the advantages of savings accounts and savings groups 4. Risk management, liquidity management and management of personal financial resources 5. Introduction to the financial system 6. Basics of personal investment 7. Culture of living up to financial and moral commitments 8. Dealing with administrative documents and processes account) was reached. Follow-up studies indicate that, after the completion of the program, a significant percentage of beneficiaries started using other financial services, such as credits, in other financial institutions. The process of mastering a skill fosters a feeling of greater empowerment The gradual introduction of beneficiaries to the financial system through a series of manageable tasks had another important effect. Accomplishing mastery in a skill is likely to improve self-esteem and to render a feeling of greater empowerment. This is exactly what the beneficiaries described happened to them. In the interviews they confided that every new achievement filled them with pride, changed their perception of self, and most importantly enhanced their belief in the possibility of changing their lives for the better. To some extent, independently from imparting important knowledge, the workshops and their content of manageable tasks ended up serving as a training ground in self-esteem, self- efficacy, empowerment, and the ability of becoming an agent of one’s own destiny. Would this have been possible with a more ambitious program or a more demanding curriculum? It is quite likely that an insurmountable challenge could have scared the participants away, even strengthening their conviction that the formal financial system was simply not for them. The Corredor program managed to find a sweet spot between imparting 26 the knowledge needed to move forward, and doing it in a way that encouraged further striving and learning. “When somebody tells me that I did something well, I want to learn much more.11� Maria-Luisa, the Corredor program Feeling empowered encourages beneficiaries to search for additional resources on their own Even a few years after having graduated from Corredor, the program’s former beneficiaries have not lost the connection with social programs. On the contrary, they report to actively seek out opportunities for further growth and development. During the interviews, the women shared experiences of successfully requesting technical assistance and training programs from their municipalities and NGOs, and organizing Women’s Associations to facilitate these processes. This multiplier effect, driven by the reported newly found confidence and self-reliance of the former beneficiaries, is probably one of the most striking changes unintentionally brought about by the Corredor program. 11  Original quote: “Cuando me dicen que lo hice bien, tengo ganas de aprender mucho mas.� “Now there are 10 members in our group Nuevo Despertar because some members did not want to give their contribution for RRPP12. The good thing about being a group is that it helps to get support from the municipality.13� Simeona – 41 years old, 6th grade completed, two children Continued accompaniment played an important role in helping internalize the objective, mastering the tasks and sustaining the feeling of empowerment It is important to emphasize the crucial role of continued accompaniment in making the “starting small, going slowly� strategy of the Corredor program so successful. First, internalizing the objective of the program – no matter how basic - does not happen overnight. This is particularly true in the presence of prior negative experiences, as was the case in the Corredor program (Programa del Banco Agrario). For example, women agree that the monthly charlas “les quitaron las dudas� (took away the doubts) that they originally had on the prudence of enrolling in the program, disclosing that they needed more than one talk to overcome these doubts. Second, the role of repetition and continued practice for mastering tasks is well recognized 27 by the educational literature. The interviews with the beneficiaries are fully consistent with this. For example, the participants emphasized how important it was for them to have the opportunity to practice speaking in front of the other ahorristas, in order to start feeling more comfortable speaking in public in community assemblies, defending their opinions and standing up for their rights. Finally, continued accompaniment helped the beneficiaries sustain their feeling of empowerment, and rendered credibility to the “yes, you can� messages of the original workshops. The fact that the program staff was there for the beneficiaries in moments of doubt - or conflict with family members opposed to their participation in Corredor – was a tangible, real and credible manifestation of the staff’s belief that the beneficiaries had the ability and determination to change their lives for the better. The experience of the ISIAT project also exemplifies the importance that starting small and going slowly (or its absence) can have in achieving the project’s outcomes. Box 6 illustrates how the project adjusted its microcredits and vocational training schemes towards starting small and going slowly, with positive results in achieving the project’s objectives. 12  Public registry. 13  Original quote: “Nuestro grupo Nuevo Despertar tiene ahora 10 personas, porque algunas no querían poner sus cuotas para RRPP. Lo bueno de estar en el grupo es que sirve para pedir apoyo del municipio.� Box 6. Starting small and going slowly – the ISIAT experience The contribution of starting small and going slowly is well illustrated by the experience of the ISIAT project. Among its objectives, ISIAT seeks to facilitate income-generating opportunities for impoverished TB-affected shantytown residents. These poverty reduction activities include microcredit loans, vocational training and microenterprise activities, which were popularly taken up by more than three quarters of the households. While microcredit loans were requested by over a third of households, only nine percent of loans were repaid on time, and over a third of borrowers defaulted. Although there were some exciting success stories where loans did lead to profitable businesses, in many cases the microcredit loans increased the debts of families. Globally, there is controversy on whether microcredit actually reduces poverty or whether it fails to work for the poorer people (see, for example review of literature in Zhuang et al, 2009). The ISIAT experience supports both of these perspectives. Initially, microcredits relied on the village banking system, with group-shared responsibility. However, this was unpopular amongst TB-affected households, and qualitative research indicated that stigma and financial pressures associated with TB discouraged the uptake of group loans. Even when personal microcredits were 28 promoted, less than a quarter of households initiated microcredits. The principal barrier was the reluctance of microcredit organizations to offer high-risk loans to the very poor, a barrier that was only partially addressed by establishing an independent microcredit guarantee fund that assumed all financial risk for unpaid loans. While microcredit was a valuable opportunity – that encouraged ISIAT project participation, synergized community mobilization and health promotion activities – the formal loans did not optimally serve the poorest TB-affected households. On the other hand, vocational training – the opportunity for formal training for work –attracted almost as much enthusiasm as microcredit loans. Nevertheless, only 9 percent of the TB-affected households actually enrolled, and this training lead to only imperceptible increases in income. These ambitious microcredit and vocational training schemes tried to help people start new businesses and learn new trades. However, the ISIAT project concluded that these activities were too ambitious a start for most of the impoverished members of TB-affected households. Gradually, the program evolved to focus on ‘starting small’ with household-based and small group-based microenterprises. The project replaced formal microcredit loans with small grants, start-up materials and equipment loans. Formal vocational training was replaced with subsidized small-group practical training by peers. The microenterprise activities supported by the ISIAT project included raising animals (rabbits and chickens) and home- based manufacturing (e.g. foodstuffs, recycling, greeting cards, knitting, weaving, jewelry, toy and handicrafts). Products were consumed in the household or sold locally. The transition from ambitious business goals to ‘starting small and going slow’ was associated with greatly increased success. The small-group, peer-to-peer nature of the microenterprise training and activities integrated with and strengthened the other objectives of the project, namely, promoting health and forming a mutually supportive civil society of people living with TB. Policy lesson 1 in a nutshell Starting small and going slowly helps to develop agency and reinforce other impacts of programs Why does this make sense? Programs that break down their goals into small, realistic objectives and tackle them sequentially are more likely to achieve them. How does it work? • Mastering basic skills helps beneficiaries feel more prepared to undertake more complex 29 tasks • The process of mastering a skill fosters a feeling of greater empowerment • Feeling empowered encourages beneficiaries to search for additional resources on their own • Continued accompaniment plays an important role in helping internalize the objective, mastering the tasks and sustaining the feeling of empowerment Try this… 1. If a program’s objective refers to a behavioral change, break it down into small steps, and introduce them sequentially 2. If possible, reward completion of small steps 3. Structure interactions between beneficiaries and the program in such a way that there is a continuous engagement POLICY LESSON 2 IN GENDER INTERVENTIONS, THINK OF BOTH GENDERS! In contexts where machismo is prevalent, reaching out to men may be a prerequisite for the success of programs targeted at women. Indeed, the beneficiaries of the Corredor Puno- Cusco program as well as the program’s staff agree that many women who decided to join the financial literacy programs had to overcome significant resistance from their spouses and other family members. Interviews with the beneficiaries and with the program’s staff suggest that such resistance constitutes an important issue. Unfortunately, the issue was overlooked at the design stage, and failed to be incorporated into the follow-up surveys to the program. Thus, the data from the interviews only suggest the pertinence of the problem, rather than making it possible to assess its magnitude. The interviewed women agreed that, based on their own experience and their knowledge of the stories of friends and neighbors, only a small fraction of spouses supported the participation in the program from the very beginning. Among the husbands who were opposed, the least recalcitrant gave up in about two weeks. Yet, individual examples suggest that in many cases it was much harder to overcome the resistance. Many ahorristas joined the program secretly, without telling their husbands, and revealed the truth about their participation only several months after. In a few cases, the husbands died without knowing that their wives were participating in the savings program. Some beneficiaries mentioned that the opportunity of showing the real benefits – the interest earned on their accounts – helped them to convince their husbands. One of the beneficiaries 31 shared her ingenious scheme to persuade her husband to the benefits of the savings program. She secretly took 100 soles from him to open the account. Seven months later, she invited him to come with her to the bank, withdrew 100 soles with interest, and returned the capital to him - keeping the interest to herself. Over time, the manifest benefits of the program helped many women become able to reconcile their husbands with their participation in the savings program; in some cases they have even been able to turn husbands into supporters of the program. Nevertheless, even in the small group of the women interviewed there are a few who still struggle with the spouses’ passive, and in some cases active, resistance. The number of stories regarding spouses’ resistance, reported both by the beneficiaries and by the program’s staff, suggests that it is an important hurdle for the successful implementation of financial literacy interventions. The program’s staff agrees that the workshops and the continued engagement of the program with beneficiaries were instrumental in helping women persevere, despite such resistance. However, the question remains on whether interventions aimed at spouses and family members could potentially overcome this problem in a more efficient and effective way, sparing the beneficiaries from potential family conflicts. “The program emphasized the issues of self-esteem in its training workshops in order to strengthen the decision of women to start saving. Many of them felt discouraged because they did not have the support of their husbands or their family; therefore these workshops were an important part of financial literacy trainings.14� Former project officers, the Corredor program The ISIAT project further illustrates the complexity of gender issues. Box 7 describes some of the challenges related to gender inequality in the diagnosis of TB affected patients, and the program’s mechanisms to overcome them. Box 7. The complexity of gender issues – the ISIAT experience Health centers in the shantytown communities of Lima diagnose TB significantly less often for women than for men. By visiting TB-affected households, questioning each resident in their home and actively requesting diagnostic samples from anyone with possible symptoms, the ISIAT project found rates of TB that were slightly higher in women than in men. In this 32 way, the apparent lack of TB in women represents a barrier of women’s access to care, rather than a true gender difference in TB disease rates (Onifade et al., 2010). Qualitative studies that investigated the nature of this barrier, however, found a strikingly equitable health care system, staffed by and mainly used for by women (Onifade et al., 2010). Rather than a health system issue, the barrier to care for women seems to be the almost universal belief - amongst TB-affected households and the community in general – that male health is intrinsically more important than female health. This is attributed to the fact that, in these settings, males generally earn all the income for the entire household. Men’s health being considered more important than women’s is not only clearly unjust, putting women’s health a risk; it also endangers the entire household and community, as delays or failure to diagnose TB prolongs the dissemination of the contagious disease. The ISIAT project attempted to overcome this gender inequality through different channels. Community workshops focusing on gender equality were pragmatically complemented with the facilitation of income-generating activities for both sexes; and with the promotion of TB screening for women as well as for men during all community and household events. 14  Original quote: Como parte de los talleres de capacitación se hizo énfasis en temas de autoestima para fortalecer la decisión de las mujeres de iniciar el ahorro. Muchas de ellas se desanimaban porque no tenían el respaldo de sus esposos o familias, por eso estos talleres fueron parte importante en las capacitaciones financieras. Policy lesson 2 in a nutshell In gender interventions, think of both genders! Why does this make sense? In contexts where machismo is prevalent, reaching out to men may be a prerequisite for the success of programs targeted at women. How does it work? • Intra-household interactions are an important part of every decision 33 • In contexts with strong traditional norms, a program that aims to empower women may undermine deep cultural beliefs, and a backlash may become inevitable Try this… 1. Even when women are the main target group, introduce informational workshops for men to better understand the objectives of the program 2. At the design stage of the program, take into account its potential effects on intra- household interactions, and the effects of intra-household interactions on the program outcomes 3. Emphasize the benefits of women’s empowerment for the entire household (greater impacts, more possibilities for men) 4. Package the initiatives as programs for men and women through women’s empowerment (to avoid them being perceived as “programs against men�) POLICY LESSON 3 IT PAYS TO SOCIALIZE! The claim that there are benefits to socializing seems intuitive, even bordering on common sense. Nearly any individual experience will attest to the fact that social interactions – sharing a laugh with a neighbor, asking a friend for advice, talking to an inspiring colleague – make one’s life more enjoyable. Social interactions also help in more practical terms: for example, a larger social network may help by providing access to job referrals, or relevant information regarding the labor market. However, are these benefits high enough to justify an intervention that would encourage socialization in a social program? Although rigorously measuring the effects of social interactions presents a formidable challenge, recent years have seen the emergence of research that succeeds in quantifying the effects of social interactions. For example, in their experimental study, Macours and Vakis (2009) showed that interactions with inspired leaders can double the effects of a productive grants program towards increasing income! Box 8 illustrates additional evidence on how social capital, which is generally acquired – and reinforced – through social interactions, can have a positive impact on diverse aspects of people’s lives. The three programs reviewed in this study present diverse experiences of using the power of socialization (or completely underutilizing it). The comparison of these experiences provides useful insights to explore the mechanisms behind the catalytic power of socialization. 35 “Chain reaction� in the enrollment to programs As narrated, the financial literacy program Corredor was initially met with much distrust and suspicion. Indeed, among the women who eventually became ahorristas, a high fraction confided during the follow-up survey that they were afraid to open up a bank account. The reasons varied from distrust in the program itself to distrust in the financial system at large. Remembering those early days, several years after the completion of the program, former beneficiaries – now experienced users of the financial system – mention the importance of talking to compañeras who had opened an account with the program earlier, as a crucial step in making up their minds to join the program. Similarly, the program’s staff emphasizes the importance of early adopters in the community. They highlight that word of mouth is at times more effective than the most comprehensive and engaging information campaigns carried out by outsiders. It is difficult to pinpoint exactly what makes finding a member of one’s social circle already in the program such a strong factor in the decision to enroll. It may be the greater trust in a neighbor compared to an outsider, the opportunity to discuss lingering doubts beyond a presentation or something else. However, both the program’s staff and beneficiaries concur on the importance of the social element in the decision to enroll. Box 8. The positive impact of social capital Different studies suggest that social capital, in its various forms, can have tangible and measurable impacts on income equality, health, happiness and life satisfaction (for example, see Phipps, 2003; and Helliwell, 2006). Other studies show the positive impact of social capital on lowering the transaction costs of economic activities and improving the performance of public institutions (Helliwell and Putnam, 1995; Knack and Keefer, 1997; and Stanley and Smeltzer, 2003). Turning to Peru, Maritz et al. (2011) find that, in TB patients, low social capital is a strong predictor of treatment abandonment and even death during treatment. Nevertheless, evidence on how or through which mechanisms various types of socialization work remains scant. This knowledge would be particularly important in order to use social interactions as a catalyst of other outcomes, such as income generation. 36 Socializing with other beneficiaries improves program sustainability through: mutual support, increased motivation and better internalization of the program’s objectives The financial literacy program Corredor had a number of features that facilitated communication between beneficiaries. The beneficiaries were encouraged to form self-help groups in order to earn monetary rewards. The rewards sought to create incentives to save, and amounted to a 25 percent increase in the account balance, with a ceiling of 25 soles a month. If beneficiaries formed a group, the reward increased to 33 percent, with a ceiling of 15 soles a month. When applying for rewards as a group, every member needed to meet a self-imposed savings target. If any member of the group failed to meet her target, all members had to switch to individual savings schemes. In addition to a higher reward, forming a savings group offered a variety of non-tangible benefits: the opportunity to discuss savings plans, seek advice, share experiences and support each other. Not surprisingly, the majority of ahorristas chose to save as a group. Another distinctive feature of the Corredor program was the organization of visits for the group of beneficiaries to other communities participating in the program. The visits looked to facilitate the exchange of experiences and to reinforce the determination to succeed in meeting the objectives of the program (learning how to save, improving financial skills, meeting self-imposed savings targets). The interviews with the beneficiaries suggest that the visits were very helpful. In addition to providing useful information to the participants, they inspired a feeling of pride and achievement. It has been established in the medical literature that positive emotions, such as feeling competence, autonomy, and relatedness, can increase perseverance and motivation (Ryan and Deci, 2000). Probably, the positive emotions resulting from the visits played a non-trivial role in ensuring the success of the Corredor program. “We enjoyed the topics of the awareness trainings, the classes of the workshop, and the interchange of experiences with women from other locations during the contests; that was the most interesting thing. For instance, we went to Puno, and women there told us that they were saving in dollars, and that gave you a feeling of moving forward (…) There were many workshops on financial education where we were taught about the fact that the Government protects the money that the bank was giving to us, they were making us aware of everything, and teaching us a lot, so we were learning.15� Fausta, the Corredor program It is worth noting that the beneficiaries of the Corredor program generally show a thorough understanding of the objectives of the program. Their level of comprehension of the objectives of the program exceeds by far, for instance, that of participants in the national Juntos CCT Program as to why they receive money from the government. For example, a group of beneficiaries from Juntos, interviewed for this study, was unable to summarize the 37 co-responsibilidades of their program appropriately. Of course, a big difference between programs is that participation in the financial literacy program is voluntary. Nevertheless, the interviews with the beneficiaries reveal striking differences in the pattern of socialization – and its likely impact on the internalization of objectives – between the two programs. Indeed, the qualitative evidence suggests that there is much less socialization among Juntos participants. The stigma they encounter as beneficiaries of the program discourages spending time together as a group. Unlike being a member of the Corredor program, being a Juntos beneficiary is viewed as a tax, a payment for the transfer that comes with discrimination and occasional insults from other members of the community.16 These prejudices reduce social interaction among beneficiaries, and may be partially responsible for the observed lack of internalization of the objectives of the program. Like Corredor, the ISIAT project – focusing on interventions against TB – contains a strong socialization component. Indeed, the project regularly organizes workshops for TB affected patients. Some of these workshops aim towards community mobilization; others have specific objectives, mainly income-generating activities. The design of the project’s evaluation does 15  Original quote: Nos ha gustado en lo que nos han sensibilizado, el curso o taller, el intercambio de experiencias con compañeros de otros lugares cuando íbamos de pasantía, eso era lo más interesante. Por ejemplo, fuimos a Puno y allá nos contaban que ahorraban en dólares, te daba una sensación de seguir adelante. (…) Habían bastantes talleres de educación financiera, donde nos enseñaban que el Estado protegía nuestro dinero, qué nos daba el banco, de todo nos sensibilizaban y enseñaban, entonces, nosotras aprendíamos. 16  Policy Lesson 6 discusses the importance of the involvement of the community in improving the effects of social programs. not allow to disentangle the specific contributions of the workshops to the overall positive effects of the ISIAT interventions (ranging from successful TB treatment to preventive therapy initiation). However, the interviews with the beneficiaries suggest that the mutual support and inspiration that they drew from the workshops contributed to their adherence to treatment and ultimately, to improvements in their health. The power of leaders The Corredor project paid particular attention to community leaders as a valuable resource. For example, at the initial stages of the project, the staff carried out a series of workshops targeting leaders – community members who later became champions of savings. The goal of the workshops was to transfer some of the coordination responsibilities from the staff to the beneficiaries. However, given existing evidence on the positive effects of engaging leaders in the community (Macours and Vakis, 2009), cultivating leaders may have had far more important impacts, beyond lowering the administrative burden of the program’s staff. First, delegating some of the administrative responsibilities to the members of the community may have increased the sense of ownership of the project of other members of the community. Several beneficiaries emphasized during the interviews the importance of seeing women similar to them in a different setting – as ahorristas, comfortable with the complexities of 38 the banking system. Such examples of indigenous, sometimes illiterate women becoming capable of opening bank accounts and dealing on equal terms with the financial system had a tremendous power to “wake them up� to new possibilities. Thus, the engagement of leaders provided examples of achievements from the community, which served as powerful source of inspiration for many beneficiaries. Secondly, beneficiaries report that leaders were highly effective in resolving the lingering doubts of fellow-residents who were not convinced by the presentations of the program staff; and sometimes even in mitigating intra-household conflicts. This effectiveness probably stems from the leaders’ deep knowledge of the community and subsequent ability to contextualize the messages of the program, far superior to that of the outsider staff. The value of feeling happier The fact that beneficiaries simply felt happier may have been an important factor in the success of the program. There is a vast psychological and medical literature that suggests that positive emotions make people more flexible and more creative, they help build the resources that people need to thrive – strength, wisdom, resilience – and are instrumental in coping with stress.17 Indisputably, such an effect on beneficiaries is likely to contribute positively to fulfill the objectives of any social program. The interviews with beneficiaries suggest that the programs that encouraged communication between beneficiaries in fact contributed to the overall happiness of the community. 17  See, for example, Fredrickson (2000). “Sadness was what we had before the program because the women stayed alone at home. Now we get together, we talk and we learn things all together. In the past everyone was walking by themselves, but now at least we support each other.18� Group interviewed in the Community of Huáscar The powerful effect of happiness and its converse, depression, are strikingly exemplified by the findings of the ISIAT project. Depression among TB patients constituted a prominent barrier to project activities and medical care. Even though the project’s objectives were seemingly unrelated to happiness, interventions of psychological counsel to address agency – manifested as depression and despair – became a key component of the project, as described in more detail in Box 9. Box 9. Addressing the hazards of depression - the ISIAT project In enquiries to members of TB affected households, the project’s staff was struck by how TB patients often cried, sometimes even describing thinking about suicide. This prompted several activities, including adding a standardized depression questionnaire to the project 39 (Karlin et al. 2011). The results were as surprising as they were concerning: almost a third of TB patients were moderately or severely depressed and one in thirty admitted to thinking about killing themselves. The depression of TB patients was independently associated with female gender, poverty, incomplete schooling, illicit drug use, and perceived discrimination – that is, depression particularly affected people who were already at high risk. TB alone was found to be associated with a 53 percent increase in the risk of moderate or severe depression. In addition to the human suffering this caused, depression was thus a prominent barrier to the project activities and TB care. Indeed, TB patients who were depressed at the time of diagnosis had a 43 percent greater likelihood of abandoning treatment before they were cured. Attempts to treat depression with drugs seemed to be ineffective, unsustainable, and difficult to integrate with TB treatment. Instead, the ISIAT project incorporated clinical psychologists into the team. The psychologists’ contribution became central to the household, small group and community workshop activities and a frequently cited driving force for beneficiary participation in the project. In this way, even when the ISIAT project aimed to address specific objectives seemingly unrelated to happiness, intervening to address lack of agency manifested as depression and despair, became a key component of the socioeconomic intervention. 18  Original quote: Tristeza era antes del programa, porque las mujeres antes estaban solas en sus casas. Ahora nos juntamos, y conversamos, y aprendemos cosas juntas. Antes todas caminaban de manera individual y ahora al menos nos apoyamos. Opportunities to enrich the program through communal activities Finally, encouraging socializing among beneficiaries may create a jumpboard, which allows enriching the program with additional activities. For example, social ties fostered by the Corredor program served as a solid basis for the organization of additional activities: pandero, canasta navideña. These activities were not incorporated in the original design of the program, and were mostly carried out by beneficiaries, with little or no intermediation by the program’s staff. Nevertheless, they strengthened the financial knowledge of the beneficiaries and contributed to the objective of facilitating savings and productive investments. “You can use a savings account; but you can also put money into the account that you cannot touch for eight months… in a fixed term account. They have taught us about many possibilities. To organize a “pandero�, to organize a Christmas basket raffle, not only about the account. (…) For instance in a Pandero, a group of friends get together and each one puts 100 soles, if the amount is 1,000 soles, I can win this money in a raffle one month, next month another friend can get it, but I continue contributing each month until the end of the year. For the Christmas Basket, you deposit money every month until Christmas, and the basket can be a large bag of rice, or two large bags of sugar, many things for which you have been saving during 40 the whole year.19� Fausta, the Corredor program Stigma, social capital and equity in the ISIAT project The importance of socialization is powerfully illustrated by the ISIAT “fighting poverty to control TB� project. TB is an infection that is relatively easy to diagnose and treat, albeit with a course of antibiotics that must be taken for a long time to be curative. However, it is also a social issue because TB is a feared, stigmatized disease. In this project, more than 80 percent of patients reported experiencing stigma because of their disease, some even keeping their TB as a secret from their families. Several women who did not keep their illness a secret, were thrown out of their homes because of their TB. The feeling of being stigmatized is greater in poorer people and in women, concentrating in already vulnerable groups. People who felt stigmatized were more likely to stop their treatment. 19  Original quote: “No solo se puede ahorrar en la cuenta, también se puede poner una plata que en ocho meses no se puede sacar… de plazo fijo. Hay muchas posibilidades que nos han enseñado. Hacer el pandero, jugar una canasta navideña, pero no solo en la cuenta. (…) Pandero es, por ejemplo, entre compañeras se juntan y ponen 100 soles cada una, si son mil [Soles] me puede tocar a mi sacarlos por sorteo, en otro mes a otra compañera le va a tocar, pero sigo poniendo yo aunque ya me lo haya sacado hasta terminar el año. La canasta navideña es que depositan plata y juntan hasta navidad para tener su canasta, o sea, un saco de arroz, dos sacos de azúcar, muchas cosas para las que ahorraste durante el año.� Stigma also makes people feel more marginalized. That is, in social capital terms, they felt less safe and supported by their communities. Box 10 describes how, just like with stigma, people who felt more marginalized – or had lower social capital – were much more likely to abandon treatment, and even die during treatment. These findings show the powerful influence that social experiences can have in peoples’ access and following of medical care, even predicting survival rates. They imply that interventions to increase agency by providing social support to people living with TB might improve health, as well as wellbeing. This was clearly demonstrated by the ISIAT project. All newly diagnosed TB patients and their households were invited to attend community workshops. Most individuals arrived furtively, sat at the back and maintained a low profile when they first attended these meetings. As they realized that everyone present was also living with TB, the change was striking. Box 10. Stigma, marginalization and treatment abandonment in the ISIAT project In addition to the human suffering that stigma causes, the patients who felt stigmatized were five-times more likely to inadequately adhere to their treatment, making their cure 41 more improbable. Furthermore, stigma makes people living with TB feel marginalized. This marginalization can be measured in terms of social capital, defined as how safe and supported people feel in their communities. In association with the stigma that TB patients experienced, they also felt much more marginalized than other individuals in their communities, and even than the people they lived with. This marginalization was more likely in more vulnerable people, women who lived in poorer households and who had completed less schooling. Just like stigma, marginalization was found to be strongly associated with abandoning treatment and even predicted patients who were more likely to die during their TB treatment, as shown below (Figure 2). Figure 2: health outcomes depend on social capital 20 15 % of patients 10 5 0 Treatment abandonment Death during therapy Source: Maritz et al., 2011. Lowest Middle Highest The community workshops gradually transformed people living with TB from a marginalized group keeping their disease a secret, into a civil society that actively supported one another and advocated locally and nationally for improvements in their care. This simple ISIAT intervention to increase agency in a high-risk, marginalized group led to decreases of stigma and increases in access to medical care. Furthermore, access to care became more equitable (Iberico et al. 2011). In regions where the ISIAT project intervened, the completion of therapy to prevent TB more than doubled and the more vulnerable households were more likely to complete the therapy to prevent TB (vis-à-vis the case of nearby communities where ISIAT did not intervene, even though treatment was also offered and recommended free of charge). These findings show how a community intervention centered on getting marginalized people together into better-informed, better-interacting, mutually supportive groups improved equitable access to potentially life-saving medical care. Further details on the comparison between regions where the project intervened and where it did not, can be found in the Appendix. 42 Policy lesson 3 in a nutshell It pays to socialize! Why does this make sense? Social interactions and social capital contribute to a variety of important development outcomes, including: poverty reduction, health, happiness, life satisfaction, lower transaction costs of economic activities and improved performance of public institutions. How does it work? • Social interactions improve enrollment in the programs by triggering “chain reactions�: 43 once a few early adopters join, their word of mouth becomes a highly effective enrollment tool • Socializing with other beneficiaries improves program sustainability through mutual support, increased motivation and better internalization of the program’s objectives • The presence of a program leader from the community develops a sense of ownership of the program in all community members • Community leaders of a program are likely to combine understanding of the program objectives with knowledge of local context, and may become the most effective program champions • Social interactions contribute to overall happiness and life satisfaction. Greater life satisfaction is associated with lowering stress and depression, which may be serious impediments to achieving program objectives Try this… 1. Targeting early adopters and helping them spread the word may be a winning strategy, particularly at the enrollment stages 2. Encourage socialization among beneficiaries, for example through the formation of self-help groups, organization of workshops or setting up fairs/competitions among participating communities 3. Collaborate with community leaders, for example, by delegating some of the managerial tasks to them – they may be the most effective champions of the program! POLICY LESSON 4 MANAGING YOUR OWN RESOURCES CREATES AGENCY Two of the programs reviewed in this study encourage their respective beneficiaries to manage resources. Juntos does so by increasing the resources available to participants: the transfer of 100 soles constitutes approximately 30 percent of a beneficiary’s family average consumption. The Corredor program, on the other hand, provides its participants with tools and techniques to better administer and plan their savings. The findings from quantitative evaluations and field interviews suggest that, in addition to accomplishing their objectives, both programs have also boosted agency. Juntos, for instance, has contributed to decrease physical and domestic violence against women and girls20, a distressing manifestation of lack of agency (see Box 11 for details). On the other hand, evidence from the qualitative studies suggests that participating in the Corredor program significantly raised the ahorristas’ level of control over resources. Interviews with the beneficiaries after the program suggest that maintaining the savings accounts played an important role in this process. Box 12 illustrates how women who kept their accounts open after the program ended, report greater autonomy in the decision of how to spend, not only their money, but that of their partner’s – including in purchasing decisions that affect the whole household. What triggered these changes? The qualitative interviews suggest a number of channels, as do existing theories on the bargaining that takes place inside a household. These theories assert that the balance of power within a marriage depends on the partner’s outside options. 45 If a woman does not have credible opportunities to leave a marriage, her partner is less likely to have any incentive to sacrifice his conveniences, desires or privileges to accommodate her preferences. Leaving an abusive relationship may be incredibly difficult if there are no economic opportunities to do so – for example, when a woman’s earning potential is far below the minimum to support the family, or when there are no jobs. But, once such opportunities arise – for example, once a woman has more economic power –her partner may become more agreeable to bargain as his prospect of losing her becomes more real. The receipt of cash transfers or the opportunity to save securely – without facing the risk of having their savings used by their partners - may have triggered a shift in the bargaining power of women within families. An increase in bargaining power may have contributed to the expansion of agency – manifested in a decrease in physical and emotional abuse, and in greater control over resources. 20  It is important to note that some evidence suggests that CCT programs may trigger increase in some forms of domestic violence. For example, Bobonis et al. (2011) find that decline in physical violence is accompanied in increase in emotional in case of Oportunidades in Mexico. These effects disappear in the long run (Bobonis and Castro, 2010). Hidrobo and Fernald (2013) find increases in emotional violence among women with less than primary education whose partners have less education than they in case of Bono de Desarrollo Humano in Ecuador. “This happens because women do not know how to earn money, if she knew how to make money, she would make decisions, because if you depend on your husband and on what he is earning, this becomes an hurdle in your way, I can see that.21� Fausta, the Corredor program “Now my husband has also changed, because before he used to tell me that everything in the house was provided by him, but now that I also earn money, I tell him that I have money.22� Genara, 38 years old, 1st grade of primary education, 3 children, the Corredor program Box 11. The impact of Juntos on decreasing domestic violence Econometric analysis suggests that enrollment in the Juntos program is associated with a 9 percent decrease in physical violence and an 11 percent decrease in emotional violence, (Perova, 2009). Domestic violence is indisputably one of the most striking manifestations of lack of agency. It is highly implausible that violence may be a desired outcome, brought to life by a consciously and freely made choice. More subtle indicators of agency –such as self- 46 esteem and a feeling of control in one’s future – are not available for a quantitative analysis of the effects of the program. Yet, the sociological and psychological literature suggests that a change in the perception of one’s self, rights and worthiness are generally required to break the cycle of violence. Thus, a decrease in self-reported abuse likely goes hand-in-hand with improvements in more subtle manifestations of agency. Figure 3: The impact of Juntos on domestic violence 18 16 14 12 10 8 6 4 2 0 Physical violence in the last 12 months Emotional violence in the last 12 months Before Juntos With Juntos Source: Perova (2010). 21  Original quote: eso pasa porque la mujer no sabe ganar su plata, si supiera ganar decidiría ella, porque al final de cuentas si estas dependiendo de tu esposo y de lo que está ganado te pone obstáculos en tu camino, eso yo veo. 22  Original quote: ahora también mi esposo ha cambiado porque primero me decía que todo lo de la casa sale de él y ahora que yo gano también le digo que yo tengo plata. Nevertheless, the Juntos transfer, as mentioned, equals only 30 percent of the average household consumption among beneficiaries – unlikely enough to enable female beneficiaries to separate from their husbands and start a new household… It may be possible however, that the Juntos transfer makes a separation from the partner possible, if combined with other sources of income, becoming in this way a step on the way to economic autonomy – crowning the ladder of other steps, such as earnings, savings, etc. Alternatively, the opportunity to manage one’s own resources may be expanding agency through other channels, in addition to bargaining mechanisms. The interviews with the beneficiaries of the Corredor project suggest a few. Box 12. Beneficiaries from Corredor report greater autonomy in spending decisions Of the group of women who kept their accounts after the program ended, a higher fraction reports making the ultimate decisions about how to spend their money, compared to the group of women who chose to terminate their accounts: 40 vs. 22 percent. Similarly, the fraction of women who have made decisions that affect the entire household – such as spending of the partner’s earnings or making minor and major purchases for the household – is also higher among women who maintained their savings accounts after the termination 47 of the program. Table 1. Decision-making over household spending Do you still have the account that you opened as part of the Corredor program? Yes No N Percentage of women who decide how to 11.6 6.3 303 spend their partner’s earnings Percentage of women who decide on minor 48.1 30.8 389 household purchases Percentage of women who decide on major 13.8 12.8 131 household purchases Source: Survey of beneficiaries, the Corredor program, 2010; authors’ calculations. First, the expanded access to owning resources changed women’s perceptions about the future, as related in the interviews. The beneficiaries summarize it in three simple words: “Ahora puedo soñar� (now I can dream). Indeed, more resources (as in the Juntos program) or greater security of and better tools to manage existing ones (as in the Corredor program) make achieving future goals more attainable. Not only wishing for a better education for their children, but being able to foresee the necessary steps to make this wish a reality may be very empowering and inspiring. This vision may deeply affect women’s perceptions of themselves, potentially triggering a change in their relationship with their partners. Second, a better management of resources may challenge the views of the women’s partners on gender equality. With a savings account and better knowledge of the financial system, women may suddenly obtain an advantage over men in an area that is not traditionally considered to be a “woman’s domain�, in the way cooking and childcare are. The beneficiaries confirm that their partners appreciate and respect these new skills and financial acumen. Potentially, this change in how a woman is perceived also contributes to the shift in family dynamics. “Before the savings program my husband did not listen to me, I almost wanted to leave him because I was not feeling well due to all the humiliation. But now he has changed because I have changed too, I am more alive, I talk in a more decided 48 way and now he helps me in the business and in the work at home, he reviews the school home work of our children and takes more care of them. If he is going to say humiliating things to me now, that would not he right because now I work and I sell things in order not the remain quiet.23� Rosalía, 42 years old, 6th grade of primary education, 5 children, the Corredor program The qualitative and quantitative research used as a background for this study, suggests that and increased access to owning income has a profound empowering effect, boosting the ability to make independent decisions and to act on them. While multiple channels may be at play – ranging from shifts in the family bargaining power to changes in gender stereotypes – improving the beneficiaries’ skills to manage their resources appears to be an effective tool to enhance agency. 23  Original quote: Antes del ahorro mi esposo no me escuchaba, casi me separo porque no me sentía bien por las humillaciones. Ahora él ha cambiado porque yo he cambiado, soy más viva, hablo más, más empeñosa y ahora el me ayuda en los negocios y en las tareas de la casa, revisa las tareas de los hijos, los cuida más. Si ahora él me dice algo, me humilla eso no estaría bien porque ahora trabajo y vendo cosas para no quedarme tranquila. Policy lesson 4 in a nutshell Managing your own resources creates agency Why does this make sense? Expanding beneficiaries’ ability to manage their resources – through transfers or training – has profound empowering effects, boosting the ability to make independent decisions and to act on them. Through these empowering effects, managing one’s own resources triggers positive impacts on other outcomes, including domestic violence. How does it work? 49 • Managing one’s own resources shifts bargaining power within the family • An expanded access to owning resources changes the perception of the future thus increasing aspirations and motivation • A better management of resources lies outside the traditional domain of women’s skills, and may challenge gender stereotypes Try this… Changing “outside options� may be an effective tool for reducing domestic violence. POLICY LESSON 5 CLEAR, ACCESSIBLE, AND CONTINUOUS INFORMATION CAMPAIGNS FOR BENEFICIARIES CAN INCREASE SENSE OF OWNERSHIP, AGENCY AND FINAL OUTCOMES It would be surprising to hear an argument against the importance of information. However, how to ensure that the intended audience absorbs and internalizes information? The answer to this question varies depending on context, and frequently lies outside the domain of common sense and general knowledge. The interviews with the beneficiaries of the three programs highlighted three requirements that information campaigns must have to be successful – that is, to improve beneficiaries’ agency and the programs’ final outcomes. Information campaigns need to: (i) target beneficiaries; (ii) be clear and accessible; and (iii) be continuous. Targeting beneficiaries Many social programs provide extensive information about their objectives, missions, and goals to external audiences: civil society, governmental institutions, and potential funding sources. However, given that their beneficiaries are unlikely to enjoy access to much of the communication media used in these endeavors (e.g. websites), they rarely receive this information. Additionally, information delivered to beneficiaries is often highly practical in nature: for instance, it centers on providing instructions to comply with the requirements, but neglects to elaborate on the broad objectives of the program. The lack of access to information can clearly reduce the potential effectiveness of the programs. 51 For example, interviews with Juntos beneficiaries showed that a gross misunderstandings of certain objectives of the program resulted in opposite behaviors to what the program is trying to encourage. Specifically, in some locations, the beneficiaries of the program do not use their Juntos accounts to save. Those who do not need to withdraw the entire bi-monthly transfer to cover current expenses, still take out all of their money in order “not to lose the benefits of the program�. There is a deeply ingrained conviction that, as the program targets the poor, leaving some money in the account would jeopardize their status as a poor household, and deprive them of future benefits. Interviewed women reported withdrawing money and converting it to cuys to save for a rainy day – precisely what the program is trying to discourage. The reason behind the misunderstanding is a lack of information. The primary sources of information on the program for beneficiaries refer to promotoras, program employees who are the primary liaison between beneficiaries and the program. Given the workload of the promotoras – attending hundreds of families, frequently in very disperse geographical locations - it is difficult to expect flawless, effective and personalized communication. Promotoras interviewed for this study admitted that there are families on their rosters that they have not seen even once. It is difficult to offer an easy solution to this problem. The excessive workload of the promotoras may partially stem from budget restrictions. However, it is important to emphasize that communication campaigns targeted to beneficiaries rarely seem to be placed high on the priority lists of the programs’ management. Recognizing the important role of clear communication flows between the beneficiaries and the program would be the first step to take advantage of the power of information. Once the role of information is recognized, different options can be pursued to achieve fluid and clear communication with beneficiaries under tight budget constraints. Options include, for instance, the use of information technology. Although the target population of social programs typically has limited access to technology, recent research shows an expansion of cell phone ownership even in remote areas of Peru, while places that do not have electricity may have radio coverage. Second, taking advantage of the local leadership, and delegating some of the liaising responsibilities to local leaders may reduce the workload of promotoras whilst improving information flows. Communication through mass media can also be further explored, including tele- and radio- novelas, or cine móvil. Clarity and accessibility Poor communication flows can not only result in gross misunderstanding of the program’s objectives, as exemplified above, they can also decrease the agency of the beneficiaries. The Latin origin of the word – agere means to drive, lead or act – suggests an active role. The opposite of an agent, or somebody in possession of agency, is a patient – the passive 52 recipient of a treatment. Lacking information can constraint the beneficiaries’ ability to take on an active role, becoming instead compliers to – sometimes misinterpreted – requirements. Indeed, while one of the objectives of the Juntos program is to build up human capital in the next generation, this is rarely communicated to the beneficiaries. Juntos recipients interviewed for this study stated that they were receiving the transfer of the program because they were poor, and in order to improve cleanliness in the households. Probed further, they did not mention any improvement of their parenting practices or attempts to provide better nutrition or mental stimulation for their children. An opportunity to jointly work on a common goal seems to have been lost. This example stands in stark contrast to the experience of the Corredor program. As described in earlier sections, participation in the Corredor program gave impetus to new initiatives carried out by beneficiaries independently, but generally aligned with Corredor’s objectives. The beneficiaries continued using the financial system, including credits, and some requested assistance from other associations for different productive trainings. Is it possible that the difference in communication with the beneficiaries was partially responsible for such sustainability of incomes and joint work on common goals, without the need of a “carrot� of material remuneration? Differences in the information campaigns between programs are indeed striking. The Corredor program placed special importance on the clarity and accessibility of its campaign. This was firstly achieved through the contextualization of the materials: all the training modules were translated into Quechua. Interviews with Credinka management showed that the brochures and other hand-out materials on the use of savings accounts were translated into Quechua and contextualized. Second, beneficiaries had ample opportunities to try out what they were asked to do: go to the bank, withdraw money, deposit money, etc. The staff at Credinka Banks was adequately prepared to deal with the new clients, which lowered potential frictions. The opportunity to practice the actions required by the program, in a safe setting, unquestionably helped participants to master and internalize these skills. Continuity Another distinguishing feature of the Corredor information campaign refers to its continuity. The workshops took place regularly, which according to the reports of the former beneficiaries, was instrumental in helping them overcome lingering doubts or family pressures. Box 13. The role of repeated peer accompanied communication in the ISIAT project The renowned Peruvian national TB program advises new patients on the importance of 53 actions such as health insurance registration, and TB screening for people living in the same household and for people who are undergoing HIV testing – all of which are available free of direct charges. Nevertheless, prior to the ISIAT intervention, the existing approach of informing patients what they and their families should be doing during a clinic consultation was only partially effective. The addition of regularly repeated similar health promotion messages provided by the ISIAT project, in the beneficiaries’ own homes and communities, markedly increased the completion of the recommended behaviors. The success was also supported by linking the health promotion messages, essentially ‘what we want the TB-affected households to do’, with income-generating opportunities that addressed the beneficiaries’ own priorities. This shift of emphasis also involved peer support – the recommendations came from peers who had themselves survived TB due to following the recommendations that they were now making. Together, these aspects of continued, community-based, peer accompanied communication achieved much greater impact than the health center-based advice alone. Insurance registration, and screening of TB for household members and patients undergoing HIV testing markedly increased after the socioeconomic interventions; for details see Appendix. Two factors contributed to the success of continuous campaigns: (i) continuity helped develop relationships with the program staff, thus forging a sense of trust. Former beneficiaries of the Corredor program refer to the workshop leaders by their names, still speak of them fondly, and highlight the importance of the communication with them; (ii) indisputably, continuity reinforces the messages. The beneficiaries agree that such reinforcement was crucial in helping them absorb and internalize the concepts of financial literacy, and the economic autonomy granted by it. It is worth noting, that even though regular workshops imposed additional demands on the beneficiaries’ time, they welcomed the initiative. Indeed, turnout at the workshops in Corredor as reported by the beneficiaries of the program was remarkably high: more than 90 percent of the beneficiaries did not miss a single of these non-obligatory workshops!24 Continuity, peer-accompaniment and accessibility in communication also played a key role in the success of the ISIAT project. Box 13 describes how the project’s communication strategy contributed to achieve positive behavior changes in terms of increasing insurance registration and screening for TB. 54 24  Source: Beneficiary survey, Corredor Project, 2010, authors’ calculation. Policy lesson 5 in a nutshell Clear, accessible and continuous information campaigns FOR beneficiaries can increase sense of ownership, agency and final outcomes Why does this make sense? Programs designed to trigger behavioral changes are more likely to succeed when the beneficiaries want the change. This is easier to achieve when participants fully understand the program’s goals and objectives. How does it work? 55 • Internalizing the objectives of the program is more likely to trigger sustained behavioral change, than “stick and carrot� incentives that ensure compliance only as long as incentives last • Misunderstanding of the programs’ objectives may trigger behaviors that are counterproductive to the objectives themselves Try this… 1. Take advantage of the local leadership, and delegate some of the liaising responsibilities to local leaders to improve information flows without having to increase the program budget 2. Make sure that information campaigns are translated into the language of the program’s potential participants and contextualized; for instance, use examples that potential participants can identify with 3. Spread the campaign over prolonged periods of time – the continuity of the campaign is key to internalizing its messages! POLICY LESSON 6 INVOLVING THE COMMUNITY AT LARGE MAY HAVE AN ADDED POSITIVE IMPACT ON AGENCY AND ON OVERALL PROGRAM OUTCOMES Why is it important to take into consideration the community, beyond beneficiaries, in social programs targeted to a specific group – for instance, poor, indigenous women? Would it not be a waste of resources? To date, qualitative work suggests the opposite. In many cases, involving the community at large can improve the benefits of the program. Not doing so – focusing exclusively on the beneficiaries and neglecting the rest – can be fraught with tensions that are likely to reduce the effectiveness of the program. Targeting criteria, in particular, are not always clearly communicated to the communities. The selection of beneficiaries25 may be at odds with the perception of poverty within the community. Additionally, there is the risk that the process of community validation may be captured by the community elite, further aggravating the sense of injustice in the distribution of benefits. The sad outcome of such a sense of injustice may be discrimination against beneficiaries. The interviewed Juntos beneficiaries reported being constantly ridiculed by the rest of their communities. For example, they are called “amantes del gobierno� (government lovers). Participation in the program becomes a badge of shame. Not surprisingly, this does not add to a greater sense of ownership of the program, nor a desire to cooperate on achieving its objectives beyond the bare minimum towards receiving the transfer. This is particularly true in communities where the differences in poverty levels are minor – sometimes it is not even clear to the beneficiaries why they are receiving the Juntos benefits, while their neighbors are not. 57 Jealousy and the sense of injustice of those who marginally missed qualifying in the program stems from a lack of transparency in targeting. On the other hand, the lack of understanding of the objectives of the program by the community at large breeds prejudices among its less poor members, for example, service providers. Beneficiaries reported discrimination at the health posts, where they are forced to wait longer than other visitors, with the haughty explanation: “You are getting money from the government? You can wait to justify it, and don’t you dare complaining�. Indeed, there is a popular belief that giving money to the poor contributes to perpetuating poverty as it eliminates incentives to work. Important objectives of the program such as investing in the human capital of the future generation are hardly ever mentioned. Not surprisingly, prejudice and discrimination at the health posts, schools or banks lowers the effectiveness of the program. The program aims at long-run behavioral change, which would ideally remain in place after the short-term incentive of the transfer finishes. Yet, it can be difficult to imagine that former Juntos beneficiaries will continue to regularly attend health posts once the transfers are withdrawn if these have become a place of humiliation for them. 25  The selection of beneficiaries of the Juntos program is carried out in 3 stages: first, qualifying districts are selected based on several measures of monetary and not monetary poverty. Second, within each district beneficiaries are selected based on the proxies of poverty calculated by statisticians. Finally, a community validation takes place, when the members of the community confirm or reject selection of beneficiaries carried out in the second stage based on the proxies. What are the practical solutions to overcome these impediments to the effective functioning of the programs? The qualitative evidence combined with international experience suggests at least two channels: (i) clarity and transparency of the targeting process and criteria; and (ii) better information and communication campaigns, targeted beyond beneficiaries, to the community at large. Clarity and transparency of the targeting process Most prejudices against beneficiaries stem from a feeling of injustice in the awarding of the program. Several mechanisms can improve the transparency of the targeting process; for instance clearly communicating the existence of a targeting criteria and emphasizing the program’s adherence to the rule. The assignation of the program by lottery, leaving the matter to luck, can have more acceptance in the community. Box 14 describes these mechanisms in more detail. Box 14. Mechanisms to improve the transparency of the targeting process 58 In the situations where a formula to approximate poverty status (based on observed characteristics) is used, it is difficult to balance transparency with the need to avoid a manipulation of the score. However, a clear explanation of the existence of the formula, and emphasis on adherence to the established – although not perfect - rule, may be in important step in reconciling those left out with their status of non-beneficiaries. Interestingly, experiences in Nicaragua suggest that when luck is responsible for awarding some and leaving others out, tensions in the community do not arise. In Nicaragua, cash grants for productive activities were assigned through a lottery. The lottery was carried out on the village square, where everybody participated and eye-witnessed that no manipulation was involved. Although not every village resident benefited from the grants, the community fabric and the social capital of the lucky recipients remained intact. Alternatively, in the case of communities with a low variation in poverty levels, geographic targeting may be the best option. The added benefit of reducing tensions may more than compensate the granting of transfers to a few residents who may be not as poor as the rest of the community. This seems particularly reasonable in locations where non-monetary poverty is widespread: such as lack of access to clean water or electricity. Such non-monetary dimensions of poverty equally affect the entire community, and equally reduce the chances of upward mobility. Notably, interviews with current beneficiaries suggest that they are in favor of geographical targeting. Many of them noted that it would have been better if “either everyone in the community received the program, or none did�. Better information and communication campaigns, targeted beyond beneficiaries The qualitative evidence suggests that many prejudices against the program are due to a lack of information about its objectives and functioning. For example, one of the major objectives of the Juntos program – investment in the human capital of the future generation – comes as a surprise to the program’s adamant critics in the communities. Repositioning the program as an investment necessary for Peru to advance into the 21st century; and promoting the evidence that reducing extreme poverty benefits communities at large, increases demand and is a prerequisite for growth would increase its acceptance in the society at large. Such acceptance would definitely be helpful for the beneficiaries. 59 Policy lesson 6 in a nutshell Involving the community at large may have an added positive impact on agency and on overall program outcomes Why does this make sense? Misunderstanding of the program’s objectives or targeting processes may trigger discrimination against beneficiaries, which is likely to lower their agency and willingness to internalize the objectives of the program, consequently undermining the impacts of the program. How does it work? 60 • Lack of clarity in the targeting processes fosters a feeling of injustice in program assignment among the non-beneficiaries. • Lack of understanding of the objectives of the program breeds prejudice among the less poor members of the community whom beneficiaries have to interact with, such as service providers, teachers, local authorities. Try this… 1. Make sure that targeting processes are transparent and understood by all members in the community. 2. In some cases, especially when overall poverty levels are high, geographical targeting may be preferable. Conclusion T his study, focused on the role of agency in development, started with an example from the 19th century literature. The story of Jean Valjean’s encounter with the bishop of Digne, which triggered a transformation of an ex-convict into a model of righteousness, successful entrepreneur, benefactor of the poor and policy maker in his role of a mayor of Montreuil-sur-Mer, is the story of agency. The bishop has lifted two constraints on Valjean’s agency, or ability to make choices to achieve desired outcomes. The first, internal constraint yielded due to bishop’s treatment of Valjean as a deserving man, worthy of unconditional trust. By forgiving Valjean’s robbery, sparing him harsh punishment, and treating him as an honest person in front of police guards, the bishop has forever changed Valjean’s perception of self. He lifted the second, external constraint on agency by giving Valjean a means to live up to this newly found perception of self, newly born aspirations: a generous gift of silverware and candlesticks afforded needed investment to start a new life 61 and a new enterprise. An example of an equally breathtaking journey – from a poor family in Junin to London Olympics – resonates. In her interviews, Gladys Tejeda acknowledges the importance and relevance of external constraints – for example, the absence of proper shoes have been making her less competitive, bringing her to the finishing line second, not the first… “I remember that once I could not be the winner because I did not have sport shoes�26. However, she recognizes that the unswerving belief of Gladys’s mother in her daughter’s talent and possibilities granted her motivation that was an incredible driving force behind her success and achievements. “I was always surprised to see that very few Peruvians were participating in the Olympic Games. And my mother too, she says, what about year 2012 are you in the next Olympic Games?�27. Economic development policies generally center on external constraints, addressing the lack of capital, education, labor market opportunities… A burgeoning strain of research has been suggesting equal importance of tackling the internal constraint, too. Even in Valjean’s story it is hard to believe that a handout of silverware and candlesticks would have had the same transformative effect on the ex-convict, had not it been complemented with the gift of new aspirations and vision. Nevertheless, discussions of how to lift internal constraint on agency – 26  Original quote: Recuerdo una vez cuando no pudo ganar porque, o sea, no tenia yo zapatillas. 27  Original quote: Siempre me sorprendió porque había tan pocos atletas peruanos en los juegos Olímpicos. Mi mama también, igual dice ella, Que tal el 2012 estas en las Olimpiadas siguientes?. changing aspirations, self-esteem and self-efficacy – rarely enter policy debates, despite their potential to improve overall program outcomes. This study has traced stories of many Peruvian women, which bear some resemblance to Valjean’s story. A combination of resources and some sort of psychological awakening have been instrumental in their journeys from saving in cui to navigating financial system, from isolation of TB patients to active participation in the community or to greater gender equality in their families and communities. Synthesizing interviews and analytical work, the study attempts to elucidate mechanisms through which such awakening took place, and present them in the form of simple and concrete policy lessons, easily transferable to other programs. The potential of simple interventions, aimed at inspiring, motivating, fostering social interactions, remains largely untapped. Few programs incorporate such psychological modalities in their design, although their costs are low and potential to improve program outcomes is high. The objective of the study was to transform an amalgam of beneficiary experiences and research into an idea-kit for policy makers, a set of simple how-tos of lifting internal constraints on agency, and thus improving overall program outcomes. Hopefully, it succeeds in making a contribution to unleashing tremendous power of motivational, inspirational and other psychological interventions in development work. 62 References Abadie, Alberto, D. Drukker, J. Herr and Guido Imbens. 2004. “Implementing Matching Estimators for Average Treatment Effects in Stata.� The Stata Journal, 4, 290-311. Abadie, Albeto and Guido Imbens. 2006. “Large Sample Properties of Matching Estimators for Average Treatment Effects.� Econometrica 74, 235-267. Alkire, Sabina, 2008. “Concepts and Measurement of Agency�. OPHI working paper. Alsop, R., Bertelsen, M. and Holland, J. (2006) Empowerment in Practice From Analysis to Implementation (Washington, D.C., World Bank). Altonji, Joseph G., Todd E. Elder and Christopher R. Taber. 2005. “Selection on Observed and Unobserved Variables: Assessing the Effectiveness of Catholic Schools.� Journal of Political Economy 113(1): 151-184. Appadurai, Arjun, 2004. “The Capacity to Aspire�. In Rao, V., and M. Walton (eds.) 2004. 63 Culture and Public Action. Palo Alto, CA: Stanford University Press. Acemoglu, D., S. Johnson, and J. A. Robinson. 2002. “Reversal of Fortune: Geography and Institutions in the Making of the Modern World Income Distribution,� Quarterly Journal of Economics, 117: 1231-1294. Banerjee, A. and L. Iyer (2005). “History, Institutions, and Economic Performance: The Legacy of Colonial Land Tenure Systems in India,� American Economic Review, 95: 1190-1213. Bellows, John and Edward Miguel. 2009. “War and Local Collective Action in Sierra Leone.� Journal of Public Economics 93(11-12): 1144-1157. Bendor, Jonathan, Dilip Mookherjee, and Debrj Ray. 2001. “Aspiration-Based Reinforcement Learning in Repeated Games: An Overview.� International Game Theory Review 3: 159-174. Bobonis, Gustavo, Melissa González-Brenes, and Roberto Castro (2012). Public Transfers and Domestic Violence: The Roles of Private Information and Spousal Control. American Economic Journal: Economic Policy (forthcoming). Bobonis, Gustavo, Roberto Castro, and Juan S. Morales (2010). Conditional Cash Transfers and Spousal Abuse: Short-Term and Long-Term Effects. Under revision. Borgers, Tilman, and Rajiv Sarin. 2000. “Naïve Reinforcement Learning with Endogenous Aspirations.� International Economic Review 41, 921-950. Dell, Melissa. 2010. “The Persistent Effects of Peruvian Mining Mita�. Econometrica. Deneulin, Séverine and Shahani, Lila. 2009. An Introduction to the Human Development and Capability Approach: Freedom and Agency. Sterling, VA: Earthscan. Access online at: http:// www.idrc.ca/openebooks/470-3/. Diecidue, Enrico, and Jeroen van de Ven, 2008. “Aspiration Level, Probability of Success and Failure, and Expected Utility�, International Economic Review, 49(2): 683-700. Fredrickson, B.L. “Why Positive Emotions Matter in Organizations: Lessons from the Broaden- and-Build Model�. Psychologist-Manager Journal 4, no. 2 (2000): 131-42. Glaeser, E. and A. Shleifer. 2002. “Legal Origins,� Quarterly Journal of Economics, 117: 1191- 1230. Helliwell, J. F. 2006. “Well-Being, Social Capital and Public Policy: What’s New?� The Economic Journal116(510): C34-C45. 64 Helliwell, J. F. and R. D. Putnam. 1995. “Economic Growth and Social Capital in Italy.� Eastern Economic Journal21(3): 295-307. Hidrobo, Melissa and Lia Fernald. 2013. Cash Transfers and Domestic Violence. Journal of Health Economics, 32(1): 304-319. Iberico MM, Zevallos K, Boccia D, Tovar MA, Gavino B Valiente A, Montoya R, Rivero M, Curatola A, C Rocha, Evans CA. Uptake and equity of TB preventive therapy was increased by a socio-economic intervention. Oral presentation published in the conference proceedings, the Global Health Conference, Montreal, November 13-15th, 2011. Open access http://www. xcdtech.com/ghc2011/18.003.html. Imbens, Guido. 2004. “Nonparametric Estimation of Average Treatment Effects under Exogeneiry: A Survey.� Review of Economics and Statistics, 86, 4-30. Karlin DM, Fernandez F, K Zevallos K, C Loiselle C, R Montoya R, Allen NR, Boccia D, Evans CA. Depression is common in TB patients and predicts treatment abandonment. International Journal of TB & Lung Disease 2011:15(11);S70. Open access http://www.theunion. org/images/stories/journal/AbstractBook2011_Web.pdf page 70. Oral presentation OP-432- 28 in session 02. The spectrum of TB management issues at the International Union Against TB & Lung Disease annual conference in Lille Oct. 2011. A recording of this presentation with the projected slides has open access at http://uwclh.conference2web.com/content/1286. Knack, S. and P. Keefer. 1997. “Does Social Capital Have an Economic Payoff? A Cross-Country Investigation.� The Quarterly Journal of Economics112(4): pp. 1251-1288. Macours, Karen and Renos Vakis, 2009, “Changing Households’ Investments and Aspirations through Social Interactions: Evidence from a Randomized Transfer Program�, (2009). World Bank Policy Research Working Paper #5137. Maritz M, Bayer AM, Zevallos K, Acosta CD, Montoya R, Rivero M, Ford CM, Evans CA. 2011. “Tuberculosis treatment adherence and mortality are predicted by low social capital�, International Journal of Tuberculosis and Lung Disease, 15(11). Narayan, D. (2005) Measuring Empowerment: Cross-Disciplinary Perspectives (Washington, The World Bank). Nunn, N. (2008). “The Long-Term Effects of Africa’s Slave Trades�. Quarterly Journal of Economics, 123: 139-176. Onifade D, Bayer A, Montoya R, Haro M, Alva J, Franco J, Sosa R, Valiente B, Ford C, Acosta C, Evans CA. A study of obstacles and opportunities for gender equity in TB control. BMC Public Health 2010 Jun 29;10:381. [Journal Open Access] [Pubmed Open Access] Open access http://www.ncbi.nlm.nih.gov/pubmed/20587044. 65 Perova, Elizaveta, 2010. “Buying out of Abuse – how Changes in Women’s Income Affect Domestic Violence�, working paper. Perova, Elizaveta and RenosVakis, 2009. “Welfare Impact of the Juntos Program in Peru: Evidence from a Non-Experimental Evaluation�, Mimeo, World Bank. Phipps, S. 2003. Social Cohesion and the Well-Being of Canadian Children. The Economic Implications of Social Cohesion. L. Osberg, University of Toronto Press. Ray, Debraj, 2006. “Aspirations, Poverty and Economic Change�, in Banerjee, Abhijit V., Roland Benabou, and Dilip Mookherjee (eds.), Understanding Poverty, Oxford University Press, Oxford. Robeyns, Ingrid. 2005. “Sen’s Capability Approach and Gender Inequality: Selecting Relevant Capabilities�. In “Amartya Sen’s Work and Ideas. A Gender Perspective�, edited by Bina Agarwal, Jane Humphries and Ingrid Robeyns, Routledge, New York. Rocha, C., R. Montoya, K. Zevallos, A. Curatola, W. Ynga, J. Franco, F. Fernandez, N. Becerra, M. Sabaduche, M. A. Tovar, E. Ramos, Tapley, N. R. Allen, D. A. Onifade, C. D. Acosta, M. Maritz, D. F. Concha, S. G. Schumacher, C. A. Evans, 2011. “The Innovative Socio-economic Interventions Against Tuberculosis (ISIAT) project: an operational assessment�. International Journal of Tuberculosis and Lung Disease, 15(5):S50–S57. [Journal Open Access] & [Pubmed]. Open access http://www.ncbi.nlm.nih.gov/pubmed/21740659. Ryan, Richard M. and Deci, Edward L. “Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being�. American Psychologist, Vol 55(1), Jan 2000, 68-78. Samman E. and Santos, M. E. 2009: Agency and Empowerment: A review of concepts, indicators and empirical evidence. Background paper for the LAC-Regional HD Report 2010. Access online: http://www.ophi.org.uk/wp-content/uploads/OPHI-RP-10a.pdf?cda6c1. Sen, Amartya. Development as Freedom. Anchor, 2000. Sen, Amartya. 1985: Well-being, Agency and Freedom. The Journal of Philosophy LXXXII, pp. 169-221. Sen, Amartya. 1992: Inequality Reexamined, Oxford. Stanley, D. and S. Smeltzer. 2003. Many Happy Returns: How Social Cohesion Attracts 66 Investment. The Economic Implications of Social Cohesion. L. Osberg, University of Toronto Press. Trivelli, Carolina y Johanna Yancari. “¿Inclusión a través de una cuenta de ahorros? El caso de las mujeres indígenas y rurales y el sistema financiero�. En Revista Argumentos, año 2, n° 3, setiembre 2008. World Development Report, 2012. Zhuang, Juzhong, Herath Gunatilake, Yoko Niimi, Muhammad Ehsan Khan, Yi Jiang, Rana Hasan, Niny Khor, Anneli S. Lagman-Martin, Pamela Bracey, and Biao Huang, 2009. “Financial Sector Development, Economic Growth, and Poverty Reduction: A Literature Review�. ADB Economics Working Paper Series. Appendix 1: Agency: what do we understand by it, and why does it matter? The concept of agency is closely linked to Amartya Sen’s capabilities approach. Following Sen, the objective of development is to enhance people’s capabilities, in the present and future, in all areas of life – economic, social, political and cultural28. The approach combines a focus on outcomes with a focus on processes. As freedom lies at the core of development, agency constitutes a process freedom (Sen, 2000), while capabilities refer to opportunity freedoms. ‘The expansion of both types of freedoms – processes and opportunities – is the objective of development and therefore, of intrinsic value’ (Samman and Santos, 2009). Agency can be defined as the ability to pursue goals that one has reason to value (Sen, 1985). The notion is thus related to the idea of driving or leading – an agent is ‘someone who acts and brings about change’ (Sen, 2000). It stresses the idea that people are not passive recipients of social programs but that they are active subjects of their own destiny. As one of the primary dimensions of gender equality, the World Development Report 2012 recognizes agency as “…the process through which women and men use their endowments and take advantage of economic opportunities to achieve desired outcomes. Thus, agency is key to understanding how gender outcomes emerge and why they are equal or unequal�. 67 Notably, Sen stresses the importance of the reason to value. Thus identifying agency entails an assessment of whether the agent’s goals are in some way reasonable – a person who harms or humiliates others would not, in this view, be exerting agency (Deneulin and Shahani, 2009). In the words of Robeyns (2005), it is important “to scrutinize our motivations for valuing specific lifestyles, and not simply value a certain life without reflecting upon it�. Among the characteristics of agency is that it is multidimensional: it can be exercised in different spheres, domains and levels; and relational – it does not occur in a vacuum (Samman et al. 2009). Agency implies an emphasis on choice; it is exercised with respect to the goals that are valued. Identifying agency thus entails an assessment of whether the goals pursued are reasonable – a person who harms or humiliates others would not, in this view, be exerting agency (Deneulin and Shahani, 2009). Agency is not collective, “…individuals, and only individuals, are the ultimate units of moral concern� (Robeyns, 2005). The notion of agency frequently overlaps with the concept of empowerment. Empowerment has been defined as “the expansion of assets and capabilities of poor people to participate in, negotiate with, influence, control, and hold accountable institutions that affect their lives� 28  Capability refers to the freedom to enjoy various functionings - the set of activities and achievements that a person may value doing or being. That is, functionings are the valuable activities and states that make up people’s well-being – such as being healthy and well-nourished, being safe, being educated, having a good job, being able to visit loved ones. (Narayan, 2005; see also Alsop, Bertelsen and Holland, 2006)29. The WDR 2012 based on Sen, understands empowerment as the expansion of agency, which, alongside the expansion of opportunities, constitutes development. The conception of agency in this case thus refers to both direct control and effective power. Aspirations are a central aspect of agency. Appadurai (2004) and Ray (2006), for instance, suggest that the lack of capacity to aspire – one of the manifestations of lack of agency - may hinder upward mobility among the poor by preventing them from investing in human capital and production technology. The role of aspirations for decision-making has been incorporated into theoretical microeconomic literature (Borgers and Sarin, 2000; Diecidue and Van de Ven, 2008), and aspirations-based learning has been introduced into game theory (Bendor, Mookherjee and Ray, 2001). This capacity to aspire is linked to social norms. Appadurai (2004) describes how ‘development ‘needs’ are always grounded in culture. Social norms thus define and constrain the space for women (and for men) to exercise their agency, imposing penalties on those who deviate from the norms. Why is agency significant in the context of gender equality? Agency matters in itself: as a freedom to live a life of one’s choosing. Additionally, agency is vitally important for policy 68 design as a catalyst of other development outcomes: “it has been hypothesized and many times confirmed, that [agency] can serve as a means to other development outcomes. The agency of women for instance, has been shown to affect positively the wellbeing of all those around them� (Sen, 1999). The WDR 2012 indicates that women’s agency matters at three levels. Agency has intrinsic relevance for women’s individual wellbeing and quality of life. It has instrumental relevance for actions that improve the wellbeing of their families. And it is required if women are to play an active role in shaping institutions, social norms, and the well-being of their communities. Sen has claimed that ‘‘the issue of gender inequality is ultimately one of disparate freedoms’’ (Sen, 1992). This perspective is key when assessing gender equality, as it challenges the assumption that the assessment of women’s wellbeing can be subsumed under wider entities such as the household or the community. 29  Women’s empowerment, however, encompasses additional elements. Women are not just one group amongst several disempowered ones (the poor, ethnic minorities, etc.); they are a crosscutting category of individuals. The household and interfamily relations are a central element of women’s disempowerment in a way that is not true for other disadvantaged groups. While empowerment in general requires institutional transformation, women’s empowerment fundamentally requires transformation in those institutions that support patriarchal structures. Appendix 2: Overview of interventions analyzed in this study The study focuses on three initiatives which were launched in Peru over the last decade: the national Conditional Cash Transfer (CCT) program Juntos; the financial literacy program Corredor Puno-Cusco and the ISIAT project – a conjunction of socioeconomic interventions in TB affected communities. Juntos program The CCT program Juntos is a flagship social program of the government of Peru. It was started in 2005 in 110 districts, covering about 37,000 households. Since then, the program has grown to operate in 638 districts, covering approximately 600,000 households. As most CCT programs, Juntos pursues two broad objectives: (i) in the short-run, to reduce poverty by providing households with cash transfers; and (ii) in the long-run, to break the intergenerational transmission of poverty through increasing investments in human capital via improved access to education and health services. The program seeks to achieve these objectives by providing eligible households with a monthly cash transfer of S./ 100 (soles). 69 On average, this amount represents 13 percent of monthly household consumption (Perova and Vakis, 2009). Unlike other CCT programs, this is a lump-sum payment and does not differ across households depending on the number of children. In order to receive the transfer, households need to comply with a number of requirements. These “co-responsibilidades� vary depending on the age and gender of the beneficiaries, and are listed in Box 1. Box A.1: Program co-responsibilidades For children under 5 years: Attend regular health controls For children 6-14 years with primary School attendance of at least 85% of the school incomplete: school year For pregnant and breast-feeding mothers: Attend prenatal and post-natal check-ups Financial literacy program: Corredor Puno Cusco The Corredor Puno-Cusco program provides technical assistance in the use of financial services. Focused on the promotion of savings through incentives among rural women with low incomes, it pursued a dual objective. First, it aimed to start the process of financial inclusion through incentivizing demand for passive financial services, such as savings and insurance. Its other objective was to check whether such financial instruments were in demand by traditionally excluded groups, specifically poor rural indigenous women. To achieve these objectives, the program relied on four instruments: 1. Personalized and regular financial education. This component included two types of workshops. The first type aimed to provide information on the formal financial system, as well as to raise self-esteem. Specifically, it covered eight basic themes: (i) development of stronger self-esteem and organizational skills; (ii) civic consciousness and consumer culture; (iii) advantages of the savings accounts and savings groups; (iv) risk management, liquidity management and management of personal financial resources; (v) introduction to the financial system; (vi) basics of personal investment; (vii) culture of living up to financial and moral commitments; (viii) dealing with administrative documents and processes. The second type of workshops aimed to develop leadership skills among beneficiaries, with the objective of transferring some of the managerial responsibilities of running the project to members of the savings groups. In addition to the workshops, the Project organized visits to other communities or fairs, where the beneficiaries could exchange experiences and reinforce mutual efforts and commitment to move forward. 2. A series of monetary incentives to open savings accounts during a three-year period. There 70 were three types of incentives: (i) an initial incentive to open an account, equal to 100 percent of the initial deposit and capped at 100 soles per account; (ii) an incentive to increase the balance during the period from 36 to 48 months, equal to 25 percent of the monthly increase in the balance and capped at 25 soles per account per month30; (iii) incentives to withdraw savings to finance the education of children, health or productive investments, equal to 20 percent of the withdrawal and capped at 700 soles. The incentives could not be withdrawn prior to the termination of the program (usually 36 to 48 months after opening the account). The maximum amount of the incentives constituted 1,520 and 1,340 soles for the beneficiaries who remained in the program during 48 and 36 months, respectively. 3. Formation of savings groups that would help the program beneficiaries to manage their individual accounts, and would create the space for exchange, skills building, and support among beneficiaries. 4. Continuous work with financial intermediaries, in order to ensure quality of service and support learning processes among the beneficiaries. Between 2002 and 2007, the program reached over 7400 women, who opened approximately 7000 savings accounts in the partnering financial intermediaries.31 30  If a beneficiary decided to join a saving group, the incentive increased to 33% of the increase in the balance, with a cap of 15 soles per month. Beneficiaries would receive the payment only if all members lived up to their savings commitments. 31  Caja Credinka en Cusco y Caja Los Andes en Puno. Support to TB affected families: Innovative Socioeconomic Interventions against Tuberculosis (ISIAT) The ISIAT program offers TB-affected families integrated community and household socio- economic interventions aiming to: 1. Enhance uptake of TB care by health promotion, community mobilization and psychosocial support; and 2. Reduce poverty through food and cash transfers, microcredit, microenterprise and vocational training. The program is based in the sixteen shantytown communities that make up Ventanilla, a region of Northern Lima/Callao, Peru. Beneficiaries Interventions focused on TB patients and their household contacts because they constitute a high-risk group for poverty and future TB recurrence and transmission. Household contacts were defined as people spending more than two hours thrice weekly in the home with the TB patient during their illness, so extended beyond residents of the household but work-place contacts were not considered because stable employment is uncommon in this setting. 71 Evaluation Many systems provide TB patients and their households with social and economic support, including the renowned Peruvian national TB control program. A distinguishing feature of the ISIAT project was a strong prospective emphasis on process and impact evaluation. All TB- affected households in the 16 participating communities were monitored prospectively prior to, throughout and after the socioeconomic intervention was rolled out with community- randomized phased-implementation in eight of the sixteen communities. Specifically, the catchment area of the intervention expanded, community-by-community to facilitate impact evaluation. Intervention in whole communities at one time allowed the creation of an active civil society by introducing interventions simultaneously to geographically clustered TB- affected households. In each community, after interventions commence all subsequently diagnosed patients and their household contacts were invited to participate. Intervention The project team consisted of nurses, health promoters, vocational trainers, microcredit / microenterprise officers, a sociologist, a psychologist, a nutritionist, diagnosticians, equity advocates, physicians and logisticians. The main mechanisms by which this team intervened were household visits, community workshops and partnering with existent organizations such as microcredit and training organizations to connect them with TB-affected households. Health promotion activities aimed to help patients overcome psychosocial barriers to increase their uptake of TB-related prevention, diagnosis and treatment through regular peer-to-peer information provision linked with incentives and enablers to care. Socio-economic activities were: 1) visits to the homes of all TB-affected households; 2) regular community-mobilization workshops and other workshops with specific objectives, principally income-generation; 3) individual and small-group psychological counseling, principally for depression and substance abuse; 4) poverty reduction activities involving poverty mitigation with food and cash transfers; and income generation by microenterprise, microcredits and vocational training. The evolution of the emphasis of the income-generating activities from large-scale microcredit and vocational training to principally small-scale household production is described above. Food and cash transfers constituted an important part of most project activities, and averaged 42% of per capita, 10% of TB-affected household median income). Food (23% of transfer value) was 72 provided at all project events, and food packages were provided in patients’ homes. Cash transfers averaged 13% of transfer value for TB diagnosis and treatment costs supplementary to the main aspects of TB care, which are provided free of direct charges; 25% of transfer value for TB- related transport expenses; and 39% of transfer value for poverty reduction (excluding microcredit loans). Outcomes Outputs of these activities were: equitable healthcare uptake for TB-related prevention, diagnosis and treatment; poverty reduction; and empowerment uniting TB-affected people into an active civil society for mutual support advocating for improvements in their care. Biosafety for the project participants and staff was increased by only invited to group activities after they have become noninfectious. From 2007 until 2012 the ISIAT project included 3,068 beneficiaries of the intervention and a further 10,487 people were recruited to facilitate impact evaluation by providing the same information but without receiving socio-economic intervention. Appendix 3: Overview of Studies used as Background Research This Appendix provides a brief description of the evaluations used in drawing the conclusions of this study. It provides more details on the first paper (Perova, 2010) to give a sense of the type of research used as a background for analysis presented in the study. Readers interested in technical details of other papers are encouraged to refer to originals. Buying out of Abuse – How Changes in Women’s Income Affect Domestic Violence (Perova, 2010) This paper develops a model, which takes into account two distinct aspects of the relationship between domestic violence and the income of the victim. Violence may be used as a rent- extraction mechanism; in this case increases in the victim’s income create greater incentives for an abuser to extract rents. Alternatively, higher income of the victim improves her outside option and therefore bargaining threat point, and may reduce violence. The model yields a set of conditions under which increases in a victim’s income decrease violence. The model is tested using the data on domestic violence and participation in a Conditional Cash Transfer 73 program Juntos in Peru. Specifically, the paper estimates the effect of increase in income due to the Juntos transfers on domestic violence. The paper uses self-reported data on domestic violence from the Encuesta Nacional de Salud, which spanned from 2004 through 2009. It estimates the average effect of exposure to the Juntos program at the district level on the incidence of physical, emotional and sexual violence. The paper uses difference-in-difference framework as a primary method. Specifically, in order to estimate the impact of Junots on the incidence of different types of domestic violence, the following regression is used: Vijt = αj + βt + γJit + Xijt θ + εijt (1) Where Vijt is a dummy variable equal to 1 if a woman i in district j in year t experienced physical, sexual or emotional violence. αj and βt are district and year fixed effects. Jit is a treatment variable which captures exposure to Juntos: it is equal to 1 if a district j received the program in year t, and 0 otherwise. Xijt θ is a vector of individual and household characteristics. It contains data on age, number of children, indicators for completion of primary and secondary education, literacy indicator and indicator for cohabitation (as opposed to formal marriage), a dummy equal to 1 if a woman is insured in order to proxy access to medical services and a dummy equal to 1 if a woman has a cash-paying job. Given theoretical evidence that there is intergenerational transmission of violence (Pollack, 2004), Xijt θ also includes a dummy equal to 1 if a woman has ever seen her father beat her mother. Household characteristics include Juntos eligibility score and its square as well as the variables used in the construction of the score: access to running water, hygienic restroom and electricity, indicator equal to 1 if a household uses industrial fuel, indicators of the types of materials used in the construction of the dwelling, and a variable used in the calculation of eligibility score which captures the number of durable goods (such as TV, vehicle, bicycle) that a household does not have. In addition, an indicator for land ownership and an indicator equal to 1 if the household is in an urban area are included. The identification strategy relies on the assumption that in the absence of the program the changes in incidence of domestic violence would not have been different in treated and control districts. To increase the likelihood that this assumption holds the sample is limited to 880 districts designated as the districts in the highest need of governmental assistance under the CRECER initiative. Of these districts, 637 were integrated in Juntos between 2005 and 2008; however, the government of Peru considers the remaining 243 districts similarly in need of governmental assistance and plans to incorporate them in the program in the future. Furthermore, it is possible to assess the validity of the identification assumption, taking advantage of the fact that ENDES survey was administered several years prior to the rollout of the program. Specifically, using ENDES 2000 and 2004, one can test whether there are systematic changes 74 in violence outcomes that are different between treatment and control groups prior to the rollout of the program. Unfortunately, the question about violent experiences in the last 12 months does not appear in ENDES prior to 2004; moreover, there were no questions about emotional and sexual violence in ENDES 2000. Consequently, it will only be possible to test for systematic differences in the dynamics of domestic abuse using the question about having ever experienced physical violence. To do so, one can estimate equation (1), where the indicator t takes values “2000� and “2004�. Jit takes the value of 1 if an observation is from a district that received the program after roll-out was initiated, and comes from 2004 survey. Otherwise, the notation remains the same. Coefficient γ captures differences in changes in outcome variable from 2000 to 2004 between Juntos districts and districts that never received the program. Given that none of the Juntos districts was enrolled in the program prior to 2005, as long as there are no systematic changes in violence outcomes that are different across treatment and control groups, the coefficient γ should not be significantly different from 0. Table A1 presents the results from estimation of regression (1) with 2000 and 2004 data. The coefficient on the interaction between the 2004 dummy and the indicator of participation in Juntos program indicates that violence have been increasing in the Juntos districts prior to the implementation of the program. Thus one can rule out the possibility that the changes in the incidence of domestic violence in Juntos districts after the introduction of the program are due to the trend. Table A1: Identification validity check - difference-in-difference estimates with 2000 and 2004 data Regressors any physical violence 0.12** dummy, =1 for Juntos districts in 2004 (0.05) -0.02 dummy, = 1 for 2004 observations (0.06) -0.27 proxy means score (0.29) 0.34 proxy means score squared (0.23) 0.01*** Age (0.00) -0.00 primary education completed (0.02) -0.07** secondary education completed (0.03) -0.09** 75 dummy, =1 if illiterate (0.04) -0.03* dummy, =1 if living together (not married) (0.02) 0.04*** number of children (0.01) 0.13*** dummy, =1 if her father used to beat her mother (0.01) 0.05 household has electricity (0.03) 0.02 household has access to hygienic restroom (0.04) -0.02 household has access to piped water (0.02) number of assets from Juntos score list that the household -0.01 does not have (0.02) note: *** p<0.01, ** p<0.05, * p<0.1. all regressions include district and year fixed effects; robust standard errors clustered at district level reported in parenthesis. Table A1: Identification validity check - difference-in-difference estimates with 2000 and 2004 data (cont.) Regressors any physical violence 0.00 type of materials 1 (0.07) -0.01 type of materials 1 (0.06) 0.11 type of materials 1 (0.07) 0.05 type of materials 1 (0.08) 0.00 household uses industrial fuel (0.05) -0.03 dummy, =1 if urban (0.06) 0.10 Constant (0.10) 76 Number of observations 4,889 note: *** p<0.01, ** p<0.05, * p<0.1. all regressions include district and year fixed effects; robust standard errors clustered at district level reported in parenthesis. Nevertheless, the test described above does not address the possibility that the error term in equation (1) may conceal time-variant omitted variables, which simultaneously affect the outcome variable and the roll-out of the program. For example, consider a hypothetical situation where Juntos districts were also more likely to enjoy an expansion of medical services. A visit of a battered woman to the doctor makes public the abuser’s behavior. Such exposure may serve as a deterrent to exert violence, and the proposed estimation would run the risk of erroneously attributing its effect to the increases in the discretionary income of women due to the CCT program. In order to mitigate this risk, a rich set of controls is included. Additionally, it is possible to assess the importance of omitted variables compared to the variables included into regression as controls, using a methodology proposed by Altonji (2005) et al. and modified by Bellows and Miguel (2009). As in Bellows and Miguel (2009), to derive the estimate of the importance of omitted variables compared to observed controls, this paper examines the changes in the magnitude of γ as the controls are added. Table A2 presents the results of estimating regression (1) with and without controls. Table A2: Juntos impact on domestic violence - difference-in-difference regressions No controls regressions experienced in the last 12 months any physical any sexual any emotional violence violence violence dummy, =1 if a district enrolled in Juntos -0.10** -0.04 -0.11** in a given year (0.04) (0.03) (0.05) 0.24*** 0.08*** 0.24*** Constant (0.03) (0.03) (0.04) All controls regressions experienced in the last 12 months dummy, =1 if a district enrolled in Juntos -0.09** -0.03 -0.11** in a given year (0.04) (0.03) (0.05) 0.26 -0.13 0.29 proxy means score (0.22) (0.15) (0.24) -0.08 0.08 -0.17 proxy means score squared 77 (0.17) (0.12) (0.17) 0.00 0.00** 0.00* Age (0.00) (0.00) (0.00) 0.01 0.00 0.01 primary education completed (0.02) (0.01) (0.02) 0.00 -0.01 -0.01 secondary education completed (0.02) (0.01) (0.03) -0.01 -0.01 -0.02 dummy, = 1 if has health insurance (0.01) (0.01) (0.01) -0.06* 0.01 -0.02 dummy, =1 if illiterate (0.04) (0.02) (0.04) 0.04*** 0.01 0.04*** dummy, =1 if living together (not married) (0.01) (0.01) (0.01) 0.00 0.01* -0.00 number of children (0.00) (0.00) (0.00) dummy, =1 if her father used to beat her 0.07*** 0.02*** 0.06*** mother (0.01) (0.01) (0.01) note: *** p<0.01, ** p<0.05, * p<0.1. all regressions include district and year fixed effects; robust standard errors clustered at district level reported in parenthesis. Table A2: Juntos impact on domestic violence - difference-in-difference regressions (cont.) All controls regressions experienced in the last 12 months any physical any sexual any emotional violence violence violence 0.02 0.02 0.03* has a job and is paid in cash (0.02) (0.01) (0.02) 0.05** 0.01 0.03 household has electricity (0.02) (0.01) (0.02) household has access to hygienic -0.01 -0.02 0.00 restroom (0.03) (0.02) (0.03) -0.01 -0.01 -0.02 household has access to piped water (0.02) (0.01) (0.02) number of assets from Juntos score list -0.01 0.00 -0.01 that the household does not have (0.01) (0.01) (0.02) 78 household owns land usable for -0.02 -0.00 -0.02 agriculture (0.02) (0.01) (0.02) -0.05* -0.01 -0.03 type of materials 1 (0.03) (0.01) (0.03) -0.01 -0.01 -0.05 type of materials 1 (0.03) (0.01) (0.03) 0.01 0.00 0.02 type of materials 1 (0.04) (0.02) (0.04) -0.00 -0.03 0.05 type of materials 1 (0.05) (0.03) (0.06) 0.02 -0.02 -0.02 household uses industrial fuel (0.04) (0.02) (0.04) -0.07** 0.01 -0.08** dummy, =1 if urban (0.03) (0.03) (0.03) 0.14 0.05 0.13 constant (0.10) (0.06) (0.10) Number of observations 3,904 3,904 3,903 note: *** p<0.01, ** p<0.05, * p<0.1. all regressions include district and year fixed effects; robust standard errors clustered at district level reported in parenthesis. The results suggest that Juntos had significant negative impact on the prevalence of physical violence and emotional violence. District enrollment in Juntos decreases the fraction of women exposed to physical violence by 9 percentage points on average, and the fraction of women exposed to emotional violence by 11 percentage points. These are over 50 percent decreases relative to pre-treatment levels of these types of violence: in 2005 14 and 17 percent of ENDES respondents in CRECER districts reported having experienced physical or sexual violence during 12 months prior to the survey. The magnitude of these coefficients hardly changes when the extensive set of controls is added - it drops from 0.1 to 0.09 in case of physical violence and from 0.1066 to 0.1062 in case of emotional violence. Based on Bellows and Miguel (2009) methodology, one can conclude that in order to explain away the impact of Juntos, the influence of omitted variables should exceed the influence of all observed variables by more than 7 times in case of physical violence, and by more than 200 times in case of emotional violence. The point estimates of the coefficients on district enrollment in Juntos in the regressions with sexual violence are negative, but are not precisely estimated. Overall, the data present strong evidence that district participation in the CCT program decreased domestic violence. As a robustness check, the impact of Juntos on domestic violence is evaluated using the matching estimator developed by Abadie and Imbens (2006) to the 2008 round of ENDES survey. In 2008 a question whether a respondent was receiving Juntos transfers was added to the survey. This allows for identifying individual beneficiaries and estimating the impact of 79 the program by comparing them to individuals with similar observable characteristics who were not receiving the program. For the matching estimator to provide unbiased estimate of the impact of Juntos on domestic violence, an “ignorable treatment� assumption should hold (Imbens (2004)). This assumption requires that potential outcomes, conditional on the observed characteristics, are independent of treatment assignment. In other words, matching is an appropriate estimation technique for the settings where, conditional on observed variables, one can assume the treatment to be exogenous. In the context of the Juntos program participating districts and beneficiary households were selected according to a clearly specified rule. Moreover, while a subset of CRECER districts - 637 - were incorporated in the program between 2005 and 2008, the government plans to expand the program to the remaining CRECER districts. All 880 CRECER districts were selected on basis of the same criteria and are regarded equally in need of governmental assistance. If the matches are drawn according to the same selection criterion that was used by the Juntos program to select the current beneficiaries and the pool of control units is limited to the districts included in the program expansion plans, the ignorable treatment assumption is likely to hold. Intuitively, beneficiaries will be matched to the individuals that would have become beneficiaries, had the program already arrived to their districts. More formally, the impact of Juntos on domestic violence is estimated as: If respondent i is a beneficiary, is an observed outcome, and is the weighted average of outcomes of similar individuals who are not receiving the transfer. If respondent i is a beneficiary, then is an observed outcome, and is the weighted average of outcomes of similar individuals who are receiving the transfer. The matching estimator of the average treatment effect is derived as the average of the differences between and over all the respondents in the sample: Juntos beneficiaries and non-beneficiaries from CRECER districts32. Juntos beneficiaries are matched to individuals in CRECER districts that are still not incorporated in the program on basis of “indicador sintetico� – district level poverty score, the score used to determine eligibility and its square as well as all the covariates that were used in the difference-in-difference regression estimation. Table A3 presents the results. The matching coefficients are consistent with the results from difference-in-difference estimation: the impacts on physical and emotional violence are negative and similar in magnitude: 8 and 10 percentage points, respectively; the impacts on sexual violence is not statistically distinguishable from 0. 80 Table A3: Juntos impacts on domestic violence - matching estimator any physical any sexual any emotional violence violence violence -0.08* 0.01 -0.10** Average treatment effect on the treted (0.04) (0.02) (0.04) Number of observations 929 929 928 note: *** p<0.01, ** p<0.05, * p<0.1. Welfare impacts of the “Juntos� program in Peru: Evidence from a non-experimental evaluation (Perova and Vakis, 2009) This paper carries out a non-experimental impact evaluation of the Juntos program, focusing on the outcomes in poverty, consumption, utilization of the health services and some final health outcomes for children under five (captured in anthropometric data). The impact evaluation uses propensity score matching, and estimates impacts in the regression framework on the matched sample. To improve the quality of the matching, the sample is limited to CRECER districts, i.e. districts considered by the Government of Peru to be in equal need of 32  See Abadie et al. (2004) for more formal presentation of the matching estimator. social assistance as the original 638 Juntos districts, and included in the program expansion plans. The impact evaluation concluded that Junots has improved a number of key welfare dimensions among beneficiaries, including moderate poverty reduction, utilization of health services, and increase in school attendance at transition points. Survey evidence on the financial literacy program The Corredor Puno-Cusco financial literacy program has carried out a number of studies to improve the design of the program. Although none of these studies has been designed as a rigorous impact evaluation, they provide a highly valuable source of information for understanding the mechanisms behind the success of the program. Prior to the launch of the program, women in the target areas were interviewed about their perceptions of the formal banking system and their savings practices. The women were interviewed shortly after the launch of the program, as well as after the program was completed, on their use of the financial services. The interviews combined survey questionnaires with unstructured interviews, thus providing a glimpse at the motivations, fears, and other emotional component of the beneficiaries’ experiences. A study of obstacles and opportunities for gender equity in TB control (Onifade D et al 2010) 81 This qualitative project formed part of the ISIAT baseline studies and was prompted by the observation that fewer women than men were diagnosed with TB by health clinics whereas active door-to-door screening diagnosed more TB in women than men. Interviews with TB patients and healthcare workers revealed a common perception that women’s TB care was of secondary importance to that of men. This reflected societal gender values and occurred despite apparent gender equality in care provision. The paper concludes that the greatest opportunities for improving women’s access to TB care appear to be in improving social, political and economic structures, more than TB program modification. Depression is common in TB patients and predicts treatment abandonment (Karlin DM et al 2011) This report briefly presents data collected in the ISIAT baseline studies and from the members of TB-affected households in regions that were not offered socio-economic interventions. This provided important context to the ISIAT study and demonstrates the importance of mutual and psychological support in the ISIAT intervention. Specifically, moderate to severe depression was common among newly diagnosed TB patients and predicted patients who later abandoned TB treatment. These findings imply that the burden of depression on TB patients and its impact in TB control represents an unmet human need that warrants targeted intervention. The Innovative socioeconomic interventions against TB (ISIAT) project – an operational assessment (Rocha C et al) This paper describes the ISIAT project and an interim impact evaluation focused on the eight communities that received the socio-economic intervention. The analysis compared pre- intervention versus during-intervention data because it was done prior to completion of the intervention and follow-up. This described the process evaluation of how the project evolved, what lessons were learnt during implementation and how the socio-economic interventions successfully engaged most TB-affected households in an active civil society. These activities were associated with marked improvements in several objective quantitate measures of uptake of TB prevention, diagnosis and treatment and poverty-reduction activities. Uptake and equity of TB preventive therapy was increased by a socio-economic intervention (Iberico et al 2011) This report briefly summarizes the first findings from the controlled impact evaluation of ISIAT comparing concurrent data from communities that were and were not receiving the socio-economic interventions. This was possible prior to completion of the project follow- up thanks to data collected and shared by the Peruvian national TB control program during 82 the treatment of each patient in the project region. This demonstrated that TB preventive therapy was least likely to be accessed by households at highest risk of TB, despite this being recommended to all TB-affected households and offered free of direct charges. The ISIAT project almost doubled rates of completion. Additionally, with ISIAT the highest risk households tended to be more likely to complete TB preventive therapy than families that were better off. Thus, ISIAT increased uptake and equity of TB preventive therapy in impoverished households. Tuberculosis treatment adherence and mortality are predicted by low social capital (Maritz et al., 2011) The paper hypothesizes that TB may be associated with diminished social capital, potentially impairing access to care. It assesses social capital and demographic factors in 1226 TB patients, a 329 healthy household contacts of TB patients, and 478 randomly selected community controls (n = 478) in impoverished peri-urban shantytowns in Peru. A questionnaire was developed consisting of 83 questions encompassing the micro, meso, and macro-levels of social capital including support (government, community, and individual), perceived safety and trust. Exploratory principle component analysis was used to identify correlated variables. Indicators of social capital were lowest in TB patients, higher in their contacts and highest in controls. A social capital score was created based on principle component analysis. Patients with low social capital at the time of diagnosis were more likely to abandon treatment or die during treatment. These associations were independent of markers of disease severity. Appendix 4 More equitable access to care - the ISIAT project The ISIAT intervention to increase agency in a high-risk, marginalized group led to more equitable access to care. This is illustrated in the graph below: Figure A1: Socio-economic intervention increases equity of TB care 50 45 Children completing chemoprophylaxis 40 35 30 (% + 95% CI) 25 20 15 10 83 5 0 Control TB-affected families TB-affected families receiving (without intervention) socio-economic intervention Poorer Less poor The bars on the left shows that in nearby communities where the ISIAT project was not intervening there was poor uptake of therapy to prevent TB, even though this was recommended and offered free of direct charges. Furthermore, the poorer and most overcrowded households that were at highest risk of further TB and were most in need of this preventive therapy were least likely to complete it. In contrast, in the regions receiving the ISIAT project intervention the completion of therapy to prevent TB more than doubled and the more vulnerable households were more rather than less likely to complete the therapy to prevent TB. Appendix 5 The role of repeated peer accompaniment Prior to the ISIAT intervention, the existing approach of telling patients what they and their families should do during a clinic consultation was only partially effective, as shown by the shaded bars in the graph below. The addition of regularly repeated health promotion messages provided by the ISIAT project in the beneficiaries’ own homes and communities markedly increased completion of the recommended behaviors, as shown by the white bars in the graph below. Figure A2. Socio-economic interventions improve patient health insurance registration, screening of contacts and testing of HIV patients 100% 90% 80% % (+95% con�dence interval) 70% 84 60% 50% 40% 30% 20% 10% 0% Patient health Contact > 18 years old Patient insurance registration screening HIV testing Pre-interventions (baseline) After socio-economic interventions T his study focuses on the instrumental role of agency – or ability to make choices to achieve desired outcomes – in economic development. The study argues that in addition to having inherent value for development, agency is a powerful catalyst for improving welfare, as measured by concrete and widely used metric of development success, such as poverty reduction, asset accumulation and investments in human capital. It also shows that improving development outcomes through agency may be highly cost- effective, and proposes concrete suggestions on how to improve effectiveness of policies and interventions through incorporating agency-enhancing elements into their design. These suggestions are based on qualitative and quantitative analysis of three Peruvian programs: a financial literacy program Corredor Puno Cusco, a national Conditional Cash Transfer (CCT) program Juntos, and the Innovative Socio-economic Interventions against Tuberculosis (ISIAT) project. Through interviews with beneficiaries, staff and community members and econometric analysis, the study attempts to disentangle mechanisms behind changes in beneficiaries’ agency, and distil them into concrete policy lessons, applicable to other circumstances and programs. Renos Vakis is a Lead Economist in the Poverty Reduction and Economic Management Unit of the Latin America and the Caribbean Region at the World Bank. His research focuses on poverty reduction, equity and gender issues in Latin America and the Caribbean. He also supports the design, targeting and impact evaluation of anti-poverty interventions. He has written extensively on issues related to poverty dynamics and mobility, risk management, social protection, gender, market failures and rural development, especially in Latin America and South Asia. He has also taught economics at Johns Hopkins University (SAIS). He holds a Ph.D. from the University of California, Berkeley. Elizaveta Perova is an Economist in the Poverty Reduction and Economic Management Unit of the Latin America and the Caribbean Region at the World Bank, and a Regional Gender Coordinator for Latin America and the Caribbean. Her research focuses on issues of poverty, labor, economic mobility and, of course, gender. She received her Ph.D. from the University of California, Berkeley. 1818 H Street, NW, Washington, DC 20433, USA.