61314 v2 POVERTY AND SOCIAL EXCLUSION IN INDIA Overview THE WORLD BANK Washington, D.C. This booklet contains the Overview as well as a list of contents from the book, Poverty and Social Exclusion in India. To order copies of the full-length book, published by the World Bank, please use the form at the back of this booklet. ©2011 The International Bank for Reconstruction and Development I The World Bank 1818 H Street NW Washington, DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org All rights reserved. This volume is a product of the staff of the International Bank for Reconstruction and Development I The World Bank. The findings, interpretations, and conclusions expressed in this volume do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement 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 publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development I The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. Contents Preface and Acknowledgments IX Abbreviations XI 1 Overview 1 Organization of the Report 6 Main Findings 8 Common Themes 31 Notes 32 References 33 2 Adivasis 37 Poverty and the Gap between Adivasis and Non-Adivasis 40 The Survival Disadvantage: Mortality among Adivasi Children 47 Other Processes Related to Higher Mortality among Adivasi Children 56 Land and Natural Resources: A Central Role in Explaining Tribal Deprivation 64 Adivasi Deprivation Is Tied Up with the Limited Voice of Adivasis 72 Notes 74 References 77 3 Dalits 83 Educational Expansion for All: Yet More for Some than for Others 86 How Does the Labor Market Behave toward Dalits? 89 Social Networks: Does It Boil Down to Whom You Know? 109 Voice and Agency Have Accompanied Economic Change 112 Conclusion 117 Annex 119 111 iv CONTENTS Notes 121 References 121 4 Women 127 The Survival Disadvantage 128 Markets and Assets: Some Progress, but Continuing Disadvantage 149 Voice and Visibility in Public Spaces 160 Conclusion 167 Notes 168 References 168 Boxes 1.1 Roma in Europe: A Large Ethnic Minority Excluded at High Cost 4 2.1 Melghat: Years of Reporting Malnutrition 48 2.2 Mistrust Is a Barrier to Adivasi Access to Health Services 62 2.3 Distress Migration among Adivasis 68 3.1 Intergenerational Mobility Is Visible, but Restricted among Dalits 104 3.2 State Support and Entrepreneurial Spirit Can Change Outcomes 108 4.1 Conditional Cash Transfers to Improve Survival and Development Outcomes among Women and Girls 141 4.2 Kudumbashree: Innovation for Results 160 4.3 Legislation Can Have Unintended Consequences, Both Good and Not So Good 163 4.4 The Challenge of Giving Excluded Groups Space in Political Decision Making in Latin America 164 4.5 What Do You Want to Do? Who Do You Want to Be? What Is Stopping You? The Aspirations of Young Girls in a South Delhi Slum 166 Figures 1.1 Rural Adivasi Children: Lower Risk of Dying at Birth, but Greater Risk by Age 5 12 1.2 More ST Children Are Severely Stunted and Wasted within the First 10 Months of Birth, 2005-06 14 1.3 Change in Postprimary Education, by Caste and Gender, 1983-2005 17 1.4 A Small Labor Market Transition among Dalit Men: Out of Casual Labor into Self-Employment 19 CONTENTS v 1.5 Wage Differentials between Dalits and Others Are Higher in Salaried Work than in Casual Work 21 1.6 Only in India and Nepal Is Infant Mortality Higher among Girls than among Boys 25 1.7 The Experience of Violence Is Associated with Worse Outcomes among Women and Their Children 27 2.1 Consumption Growth among Urban STs Was Highly Skewed with Bigger Gains Near the Top of the Distribution 46 2.2 The District-Level Relationship between Infant Mortality and the Concentration of STs 51 2.3 Adivasi Children: Lower Risk of Dying at Birth, but Greater Risk by Age 5 51 2.4 More ST Children Are Severely Stunted and Wasted within the First 10 Months of Birth, 2005-06 57 2.5 Tribal Children: Less Likely to Fall Sick with Cough or Fever, but Much Less Likely to Be Treated 58 2.6 Mothers of Tribal Children Are Less Likely to Receive Health Care 60 3.1 Change in Postprimary Education, by Caste and Gender, 1983-2005 87 3.2 Dalit Men: More Likelihood of Casual Labor in Rural Areas; Less Likelihood of Self-Employment in Urban Areas 93 3.3 Where Does Education Take Rural Men in the Labor Market? 98 3.4 Where Does Education Take Urban Men in the Labor Market? 99 3.5 Salaried Work: Wage Differentials between Dalits and Others Are Higher 100 3.6 Adivasis and Dalits Have Few Social Networks in the Formal Sector 105 4.1 Only in India and Nepal Is Infant Mortality Higher among Girls than among Boys 130 4.2 Fertility Is Declining: Rates in Many Indian States Resemble Rates in Developed Countries 131 4.3 Nearly One-Fourth of Indian Women 20-24 Had Their First Child by the Age of 18 135 4.4 A Majority of Indian and Bangladeshi Parents Do Not Consider Daughters as Sources of Old Age Support 138 4.5 Indian Women Face a 1 in 70 Chance of Dying in Childbirth 139 VI CONTENTS 4.6 Occupational Structure Is Changing More among Men than among Women, and the Differences across Social Groups Are Large 152 4.7 Women Are Paid Less than Men across the Wage Distribution 155 4.8 The Aspirations of Women Doing Only Domestic Work 156 4.9 Almost Half of MGNREGS Participants Are Women, but Some States Do Better than Others, 2010 159 Tables 1.1 Poverty Rates: STs Are 20 Years behind the Average Population 11 1.2 Adivasi Child Mortality Exceeds the Relative Population Share in Rural Areas 12 1.3 The Representation of Dalits in Government Jobs, 2006 20 1.4 Three in Five Women Cite Lack of Credit as a Reason for Not Doing the Work They Want 30 2.1 In Terms of Poverty, STs Are 20 Years behind the Average Population 41 2.2 The Depth and Severity of Poverty Have Declined More Slowly among STs than among Other Social Groups 43 2.3 The Increasing Concentration of STs in the Poorest Wealth Deciles, 1993-2005 44 2.4 Trends in Poverty Incidence in States with a High Proportion of Adivasis 45 2.5 Child Mortality among Adivasis Far Exceeds Their Relative Share in the Rural Population 49 2.6 Child Mortality Rates among STs, by State, 2005 50 2.7 STs and Non-STs: The Gap in the Relative Hazard Rate 52 2.8 Early Childhood Mortality Rates by Residence and Ethnicity among Different Cohorts, 2005 54 2.9 ST Children: The Breadth of Immunization Coverage Improved, but the Intensity Expanded More Slowly 60 2.10 The Average Landholding in Rural Areas 65 3.1 Odds Ratios Predicting Labor Force Participation 91 3.2 Small Labor Market Transitions among Dalit Men, While Dalit Women Are Exiting Casual Work and the Labor Market 92 CONTENTS VII 3.3 The Representation of Dalits in Government Jobs, 2006 102 3.4 What Do Self-Employed Men Do? 106 3A.1 Type of Employment, Rural Men 119 3A.2 Type of Employment, Urban Men 120 4.1 The Decline in Reported Son Preference in India, 1992-93 to 2005-06 133 4.2 The Slow Decline in Maternal Mortality Rates, India, 2001-03 to 2004-06 140 4.3 Antenatal Care Improved in 1998-2005, but Most Women Still Do Not Find Giving Birth in Medical Facilities Necessary 143 4.4 Indian Women Who Experience Domestic Violence Also Report -Poorer Reproductive Health Outcomes 145 4.5 Women in Rapidly Growing Regions (West and South) Show Higher Odds of Labor Force Participation 151 4.6 Three in Five Women Cite a Lack of Credit as a Reason for Not Doing the Work They Want 157 4.7 The Poor Awareness of Women about Available Credit Facilities 158 Preface and Acknowledgments This report has been prepared by a team led by Maitreyi Bordia Das (SDV, World Bank), under the guidance of Mansoora Rashid (Sector Manager, SASSP, World Bank) and N. Roberto Zagha (Country Director, India, World Bank). Core team members are Soumya Kapoor (Independent Consul- tant), Rinku Murgai (SASEP, World Bank), and Denis Nikitin (SASSP, World Bank). John Prakash Badda (SASSD, World Bank), Gertrude Cooper (SASHD, World Bank), Sujata Pradhan (SASSD, World Bank), and Tanusree Talukdar (SASHD, World Bank) have provided administrative and document layout support. This report draws on background papers written by the core team members with Gillette Hall (Georgetown Public Policy Institute) and with Puja Vasudeva Dutta (SASSP, World Bank). Individual papers were prepared by Neera Burra (Independent Consultant); Sonalde Desai, James Noon, and Reeve Vanneman (University of Maryland); Amaresh Dubey, Vegard Iversen, Adriaan Kalwij, Bereket Kebede, and Arjan Verschoor (respectively, Jawaharlal Nehru University, University of Manchester, Indian Statistical Institute, University of Utrecht, and University of East Anglia); Iffath Sharif (SASSP, World Bank) and Jeffrey Witsoe (University of Pennsylvania). Early work also drew on data analysis by Kiersten Johnson (ICF Macro) and Pinaki Joddar (Independent Consultant). Fieldwork commissioned for this report resulted in two papers on Dalits by Surinder Jodhka (Jawaharlal Nehru University), and the Dalit chapter heavily draws on these. In addition, the report takes qualitative evidence from the World Bank study, Moving out of Poverty. This report has been prepared alongside a parallel World Bank India Poverty Assessment led by Peter Lanjouw (DECRG, World Bank) and Rinku Murgai. The two reports have a few common team members and often draw on common material. Background work for this report also features in the World Bank's Indigenous Peoples, Poverty, and Development. The generous financial contributions of the U.K. Department for International Development-India and the Swedish International Development Cooperation Agency are gratefully acknowledged. ix X PREFACE AND ACKNOWLEDGMENTS Valuable comments during the preparation of the report were provided by Junaid Ahmad (AFTUW, World Bank), Tara Datt (lAS, Joint Secretary, Cabinet Secretariat), A. K. Shiva Kumar (United Nations Children's Fund), Martin Macwan (Navsarjan), Aasha Kapur Mehta (Indian Institute of Public Administration), Dev Nathan (Indian Institute for Human Development), Harry Patrinos (HONED, World Bank), Kumar Amarendra Narayan Singh (SASSO, World Bank), Shekhar Singh (Centre for Equity Studies), Varun Singh (SASSO, World Bank), and Sukhadeo Thorat (University Grants Commission). Initial discussions on framing the issues have benefited from the views of an informal advisory group meeting held at the Planning Commission in July 2008 and comprising Drs. B. L. Mungekar (ex- Member, Planning Commission), Surinder Jodhka, Mary John (Centre for Women's Development Studies), and Virginius Xaxa (University of Delhi). The report also benefited from comments received from Mr. Amitabha Bhattacharya (Planning Commission) and Dr. T. K. Oommen (Jawaharlal Nehru University) during the presentation of draft findings to the Planning Commission in April 2010 and from representatives of various ministries and agencies of the government of India at a seminar organized by the Department of Economic Affairs, Ministry of Finance. The team thanks peer reviewers Alaka Mal wade Basu (Cornell University andJawaharlal Nehru University), Alexandre Marc (SDV, World Bank), and Lant Pritchett (Harvard University) for their com- ments on a draft of this report and for participating in the Review Meeting. The team also thanks all other participants of the Review Meeting who provided comments. Finally, the team gratefully acknowledges the support and guidance received from Dr. Narendra Jadhav (Member, Planning Commission and National Advisory Council) during the preparation of the report. The Report has been discussed with the Government of India, but does not necessarily bear their approval for all its contents, especially where the Bank has stated its judgment, opinion, or policy recommendations. Abbreviations DHS Demographic and Health Survey MGNREGS Mahatma Gandhi National Rural Employment Guarantee Scheme NFHS National Family Health Survey NSS National Sample Survey NTFP nontimber forest product OBC Other Backward Class PESA Panchayats Extension to the Scheduled Areas Act RCH II Reproductive Child Health Survey II sc Scheduled Caste SHG self-help group ST Scheduled Tribe XI 1 Overview In India ... we must aim at equality. That does not mean and cannot mean that everybody is physically or intellectually or spiritually equal or can be made so. But it does mean equal opportunities for all, and no political, economic or social bar- rier.... It means a faith in humanity and a belief that there is no race or group that cannot advance and make good in its own way, given the chance to do so. It means a realization of the fact that the backwardness or degradation of any group is not due to inherent failings in it, but principally to lack of opportunities and long suppression by other groups. Jawaharlal Nehru, The Discovery of India (1946) Nehru would have been proud of India's stellar performance in the creation of a new economic system and its self-confidence relative to other nations, though not necessarily of the acquisitive society that goes against the collectivist spirit he espoused (Nehru 1946). There are other developments to be proud of as well. Poverty has declined among all groups; enrollments in schools have increased; health out- comes have improved; and fertility rates in many Indian states now resemble those of developed countries. Moreover, the 73rd and 74th amendments to the Indian Constitution that mandate the devolution of powers and responsibilities to rural panchayats (local govern- ments) and urban local municipalities have set the agenda for local ownership of and participation in governance. In parallel to these positive developments, rising inequality in India has been a subject of concern among policy makers, academics, and activists alike. The recent India poverty assessment of the World Bank (2011) shows that, when inequality is based on income rather than consumption measures, India is not so different relative to the 1 2 POVERTY AND SOCIAL EXCLUSION IN INDIA inequality levels recorded in countries such as Brazil and South Africa, countries commonly singled out as global outliers. In other words, if inequality in India is measured on the basis of per capita income, it stands among those countries with the highest recorded inequality rates. The report also decomposes inequality according to differences within and across states, according to urban-rural status, and according to levels of educational attainment. In our report, we focus on social exclusion, which has its roots in historical divisions along lines of caste, tribe, and the excluded sex, that is, women. These inequalities are more structural in nature and have kept entire groups trapped, unable to take advantage of opportunities that economic growth offers. Culturally rooted systems perpetuate inequality, and, rather than a culture of poverty that afflicts disad- vantaged groups, it is, in fact, these inequality traps that prevent these groups from breaking out. 1 Therefore, cultural factors can play a role in sustaining inter-group differ- ences in wealth, status and power. Where the mechanisms involved are self-enforcing this can be considered to be an "inequality trap." Where such an inequality trap exists, it implies that subordinate groups are maintained at least in rela- tive poverty, and that these are associated, in part, with cultur- ally shaped behaviors, including endogenous preferences that can limit the prospects of poorer, or subordinate, groups. (Walton 2007, 2) So, the question we ask in this report is: who has gained and who has not from India's growth surge? Well, then, why "social exclusion" and why not "inequality"? This is a question often asked by those who have read early chapters of this report. India is not alone in having social groups that have been traditionally excluded; yet, the structure of the caste system and its ramifications on employment, education, and the rules of social and economic exchange are distinctive and shared only with Nepal. Both caste and gender inequalities are rooted in a philosophical tradition that justifies these through religious texts that provide systematic rules for exclusion. This is what makes these inequalities particularly durable (Tilly 1999). And, for this reason, our report focuses not on inequality, but on exclusion. The relevance of social exclusion has been incorporated into the development lexicon in a number of ways. The term "social exclu- sion" was first used in France in the 1970s to distinguish the excluded, who then comprised a wide variety of people: the disabled, suicidal and elderly persons, and abused children, among others (Silver 1994). OVERVIEW 3 Since then, it has been used in the social science literature to distin- guish from and add to the concept of poverty and to denote rules of exchange and practices that keep groups out; the term can also some- times mean different things to different people. Globally, excluded groups tend to be ethnic or religious minorities that, by virtue of their distinct cultural practices, are considered the "other." Social exclu- sion as a concept perceives the individual as an entity embedded in society and groups. The focus is thus not on outcomes such as increased consumption or income or education alone, but on rela- tions that constrain individuals from achieving these outcomes (de Haan 1997). Nobel Laureate Amartya Sen (1998) calls these the "relational roots of deprivation," whereby membership in a particu- lar group (women, lower castes, indigenous people, or persons with disabilities) limits the "functionings" of individuals to acquire or use their capabilities. Social exclusion is therefore not about outcomes alone, but about the processes that lead to these outcomes. DeHaan (1997), for instance, calls income poverty only one element of social exclusion and refers to its eradication as a part of the broader agenda of social integration. This is because poverty focuses only on an absolute measure of indi- vidual well-being: income, consumption, or human development. In contrast, social exclusion focuses on processes or noneconomic means that exclude certain groups from equal access to basic goods and ser- vices that, in turn, determine their well-being. As a result, these groups have unequal access to labor markets and social protection mecha- nisms through formal and informal institutions. Even for people with equal levels of human capital and skills, there appears to be an impor- tant element of discrimination that is part of what one would define as social exclusion beyond purely economic considerations. Finally, exclu- sion involves unequal access to the full exercise and protection of rights and liberties, including, sometimes, the denial of basic human rights. India is not alone in grappling with issues of social exclusion. While many other countries include similarly excluded minorities, the most relevant empirical evidence and conceptual clarity come from Latin America and the Caribbean and from Eastern Europe. Recently, the flagship report ofthe Inter-American Development Bank, as part of its annual series, Economic and Social Progress, focused on social exclu- sion. The report, Outsiders? The Changing Patterns of Exclusion in Latin America and the Caribbean, is a regional analysis of ways in which different groups have been excluded during a period of macro- economic transformation (see Marquez eta!. 2007). It documents the economic and welfare costs of exclusion, suggesting that inclusion policies be viewed as an investment rather than as a handout to the least well off in society. Inclusion policies are more than new 4 POVERTY AND SOCIAL EXCLUSION IN INDIA programs or new institutions to redress past injustices; they imply fundamental changes in the way decisions are made, resources are allocated, and policies are implemented in democratic societies. In Eastern Europe and Central Asia, the focus on the Roma, for instance, distinguishes between poverty and exclusion (see box 1.1). Box 1.1 Romain Europe: A Large Ethnic Minority Excluded at High Cost The Roma (Gypsies) number about 10 million-12 million people in Europe and represent the largest transnational minority in the region. The most significant representation of Roma is in Central and Eastern European countries, where, overall, the Roma account for the main poverty risk group, suffering from low educational attainment, high unemployment, and poor human development outcomes. One of the most frequent sites of Roma discrimination is the labor market, where even qualified Roma are discriminated against in favor of non-Ro- mani applicants. In general, however, Roma lack sufficient education to participate in the labor market. A recent study across four Central and Eastern European countries-Bulgaria, the Czech Republic, Romania, and Serbia-suggests that as little as 12.5 percent of work- ing-age Roma in Bulgaria, Romania, and Serbia are educated at or above the secondary level (World Bank 2010}. The low educational levels are reflected in employment rates of 60 percent, on average, relative to the majority population and, not surprisingly, much lower earnings. The study estimates an annual economic loss of €5. 7 billion and a fiscal loss of €2 billion to the four countries as a result of reduced productivity and additional costs incurred to finance the social security of unemployed Roma. In fact, it warns that these losses will only increase over time as younger Roma join the working-age population. The annual fiscal gains from investing in the education of Roma are significantly higher than the costs incurred even if all Roma people were to be educated {World Bank 2010). The Roma Education Fund is one such initiative, funding specific efforts that can make formal schools more responsive to the needs of Roma children (see Roma Education Fund, at http://www.romaeducationfund.hu/). The empha· sis is on giving Romani children a good educational start by focusing on access to preschool and successful transitions into and through primary education. While the impact of the fund is yet to be assessed, funded projects reached about 30,000 Roma students in 2008, among whom 800 graduated. Source: Roma Education Fund website, http:/!www.romaeducationfund.hu/. OVERVIEW 5 A 2005 report on multiple deprivations in Serbia and in Montene- gro found a strong ethnic dimension to exclusion (World Bank 2005). It found that the Roma population, particularly those Roma residing in settlements, was among the most deprived. A detailed analysis of Nepal has, to a large extent, influenced the manner in which the government collects data and designs policies for people who have been historically and ritually excluded (World Bank 2006). Other important analyses of social exclusion have also influenced multilateral agencies and national governments and have, in turn, been influenced by national debates on the issue (see Baker 2002 on Uruguay, for example). In Brazil, Gacitua-Mari6 and Woolcock (2008) find that the poor, especially Brazilians of African descent, have been systematically excluded from the economic growth pro- cess, despite laudable improvements in education levels and broad reductions in poverty. This exclusion has persisted (even intensified) over decades. Addressing social exclusion for its intrinsic value is sufficient rea- son. Many studies have shown the instrumental value of addressing social exclusion (see Akerlof 1976; Scoville 1991 ). This report argues, however, that addressing social exclusion because exclusion is morally and socially unjust is reason enough. Moreover, mere ethnic or linguistic heterogeneity is not the same as exclusion. How- ever, if such heterogeneity succeeds in keeping groups out, its exclu- sionary role plays out in the overall development outcomes of a country or a region. In India, as this report shows, substantially higher mortality among Adivasi children in the 1-4 age-group is affecting India's overall advances in lowering child mortality. This report does not explore or adhere to any of the myriad formulations of social exclusion, which, by themselves, are concep- tually challenging. In some sense, the focus is quite simple. It is based on the established existence of three axes of exclusion in India: caste, tribe, and gender. These axes, particularly caste and gender, have a normative foundation in historically grounded pro- cesses and are responsible for a number of unequal outcomes and processes. For instance, norms of son preference, derived from the value of sons in supporting their parents and in undertaking the rituals around parental death, are the foundation for a consis- tent neglect of female children and women in their reproductive years. Similarly, the occupational logic of caste has meant that the traditionally lower castes remain typed into traditional trades and jobs, making mobility difficult. The geographical and cultural iso- lation of tribal groups has meant that they are still difficult to reach and that their traditional patterns of landownership are at variance with new developments. 6 POVERTY AND SOCIAL EXCLUSION IN INDIA The overarching themes in this report revolve around three inter- related aspects: services, markets, and political spaces. In most coun- tries, excluded groups face discrimination, lower access to health and education, and lower returns to education and assets. They are also more likely to be poor, and this likelihood is passed down through the generations. This is true in India as well, but there are some unique aspects of the Indian trinity of caste, tribe, and gender that lead to dif- ferent processes and, hence, different outcomes. For instance, unlike countries in Latin America, where excluded groups face higher unem- ployment rates, open unemployment in India has been historically low, and the real labor market consequence of exclusion is allocation to low-level occupations and lower mobility. Excluded groups are less likely to participate in political processes in other countries as well, but, in India, this is a more nuanced process. Social movements represent- ing excluded groups are much more potent, and affirmative action, which is fairly new in many Latin American countries, has been consti- tutionally mandated in India since the country's independence. Finally, the representation of excluded groups in political spaces has also been secured, although the chapters in this report show that the impacts of this representation on group welfare have been mixed. Organization of the Report The focus. Exclusion operates along multiple and interrelated dimen- sions. Because addressing them all would be an insurmountable task, this report does not attempt to do so. It does not, for instance, focus on exclusion because of religion or disability, though there is extensive research to show that these are important axes along which people face deprivation. Even in the case of caste, tribe, and gender, the litera- ture is vast, multidisciplinary, and so prolific that it appears to grow every day. The evidence contained in this report builds on this vast aca- demic and activist literature, but the report is by no means an attempt to provide a comprehensive survey of the evidence or even a review of all the issues involved. The fact that each of these groups is highly heterogeneous and that outcomes and processes differ by state and district and by type of caste or tribe makes the task even more chal- lenging and generalizations that much more difficult. The report focuses on three select groups that face exclusion in India. In addition to Scheduled Tribes (STs) and women, the report dis- cusses outcomes among Scheduled Castes (SCs), even though recent data suggest that Other Backward Classes (OBCs) also fare poorly on some indicators. 2 There are three reasons for this. First, as we show in the report, SCs face structural inequalities that have a ritual OVERVIEW 7 backing in historical processes such as untouchability. In contrast, OBCs do not face such historical deprivation. Instead, they comprise- and are defined so by the Constitution of India-as castes that are not SCs, but that are backward in most respects. Second, the issues related to OBCs, who were traditionally peasant cultivators, are dif- ferent from those affecting outcomes among SCs and merit a sepa- rate analysis. Finally, National Sample Survey (NSS) data on OBC outcomes is available only after 1999-2000, meaning that it is dif- ficult to monitor the performance of OBCs along a range of develop- ment indicators over time (for example, since 1983, as the report does for other groups). For all these reasons and because of the fact that this report deals with more deeply rooted structural inequali- ties, the focus is primarily on the outcomes among SCs. However, this should not suggest that caste can be conflated with SCs or that other caste-based issues are not important. Outcomes among OBCs are discussed wherever data are available. Similarly, even though exclusion by religion is significant, the issues affecting Muslims merit an entire analysis, and religion is discussed here only partially (in the context of the labor market). For each group, too, the focus is on a subset of issues that offer some new insights. For SCs, the report focuses on poor labor market out- comes despite the expansion in education. For STs, it focuses on a dis- tinct disadvantage in survival and on the correlates, which go beyond the health sector and include geographical isolation and the removal of tribals from their traditional lands and assets. For women, it focuses on poor health and survival outcomes, but also the disadvantage of women in the labor market and the vulnerability of women to violence within the family and insecurity in public spaces. Thus, even though there are missing variables in the analysis, it is our hope that, by deeply probing one or two of the kinds of exclusion faced by each category, we provide tools that can be applied to study other forms of exclusion or issues not covered in this report. The format of our analysis is such that we compare outcomes among one subgroup with outcomes among other subgroups. However, since the report also tracks changes in outcomes over time, the finer aspects of development reality are sometimes lost. Objectives and data. The report is organized around three chap- ters, in addition to this overview, each one dealing with an excluded group: STs, SCs, and women. The objective is to provide a diagnos- tic of how the three excluded groups under analysis have fared along various development indicators during a period of rapid economic growth in the national economy. In seeking this objective, the report also addresses correlates and the processes that explain how and why these groups have fared the way they have over a period of time. 8 POVERTY AND SOCIAL EXCLUSION IN INDIA The report is not intended as a policy document, although some of the empirical findings lend themselves to policy conclusions. Data for the report come primarily from multiple rounds of national-level household survey data (the NSS, the National Family Health Survey [NFHS], the Reproductive Child Health Survey, and the India Human Development Survey), as well as international sur- veys (Demographic and Health Surveys and the database of the Population Reference Bureau). Large national data sets, while allow- ing for national-level and, to some extent, state-level generalizations, do not allow analysis of outcomes below the state level or within group differences and processes. It is well known, for instance, that there is substantial diversity among SCs and STs even within states. This heterogeneity is widely known because of microstudies and ethnographic, anthropological, and journalistic accounts. While there is a rich body of such qualitative work in India, the results provide contextual specificity, but cannot be added up to paint an aggregate picture. Moreover, such studies or accounts generate results that are limited to one tribe, village, or state and cannot be generalized. Nonetheless, we do draw upon such studies, in particu- lar, qualitative work commissioned for this report, to capture under- lying processes and triangulate with the quantitative analysis. A word on usage: the terms "Scheduled Caste (SC)" and "Sched- uled Tribe (ST)" come from the Constitution Order passed in 1950, which contains the names of castes and tribes that are earmarked for special treatment, such as reservations in legislation, public sector employment, and government-run educational institutions. How- ever, SCs and STs have taken names that signify both political asser- tion and a rejection of the previously accepted idioms that reduced them to "reserved categories." For the purposes of the tables and graphs presented in this report, we use the terms "SC" and "ST" because these are standard administrative and survey categories. In the text, we use the terms Dalits and Adivasis (or tribals) interchange- ably with SCs and STs, respectively. Main Findings During the period of rapid economic growth, what happened to entrenched group inequalities? Were there ways in which traditionally excluded groups such as SCs, STs, and women broke out ofthe traps or did the traps trump the opportunities? The report emphasizes that the changes wrought in the 20 years beginning in 1983 were complex, and social groups transformed themselves in different ways in response to new opportunities. For instance, far from being the immutable frame OVERVIEW 9 that Weber (1958) seemed to suggest, "caste," in fact, has evolved and mutated. Nonetheless, by no means are we seeing the annihilation of caste, as Ambedkar (1936) exhorted Indians to achieve. I still think that after a long and convoluted path, after many a stumble and wrong turn, a different kind of moment seems to be upon us ... there is a sense of hope across the country, which I believe is universal. There is a momentum for change. Nandan Nilekani (2008, 484) Adivasis The self-preferred term "Adivasi" is commonly translated as "original inhabitants," and literally means Adi, beginning or earliest time, and vasi, resident of. There is active discussion around the question of "original inhabitant," but this is not directly relevant to our report. According to the 2001 Census, the population of India includes 84.3 million STs, comprising 8.1 percent of the total population of the coun- try. However, the proportion of STs has been increasing over census periods partly because more tribes are being included in the ST cate- gory and partly because of higher fertility rates in this category (Maha- ratna 2005). While the Constitutional Order declared that 212 tribes located in 14 states are STs, the Government of India today identifies 533 tribes as STs, of which 62 are located in the state of Orissa. 3 Chapter 2 in this report focuses on the Adivasis or STs. In most analyses, this topic is addressed after the Dalits, but we have placed it first for analytical and organizational purposes. There are two reasons for this: tribal groups are not strictly within the caste system, and the bonds of rituals do not affect their relations with the world in general. Also the report shows that outcomes among Adivasis are among the worst, despite considerable variation across places of residence and tribal groupings. Tribal groups in India are often conflated with castes, and "Sched- uled Caste" and "Scheduled Tribe" are sometimes said in the same breath, although they are actually distinct social categories. As Andre Beteille writes (1998, 187): "Even the best ethnographers [in India] habitually confused tribe with caste, which, on any reasonable assumption, is a different kind of social category." The major differ- ence between SCs and STs is that, while the former were subject to historical, ritualistic discrimination, STs were excluded from the national context because of their physical isolation. Inhabiting topo- graphically inaccessible areas, STs still face difficulties in accessing services such as health care and education. Even though they own 10 POVERTY AND SOCIAL EXCLUSION IN INDIA more land than SCs, a complex set of contravening forest laws means they are confronted by barriers in selling their produce from the land, perhaps explaining the high poverty rates among STs in rural areas. Most tribals therefore end up migrating, leaving their forest lands to work as casual laborers in urban centers. The report focuses on two major markers of tribal deprivation: poverty levels, which are higher than the national average and in comparison with other groups (including SCs), and child mortality. We argue that the roots of Adivasi deprivation lie in two main fac- tors. One, their physical segregation renders problematic the deliv- ery to them of services such as health care and education. Two, their traditional dependence on land and forests and their widespread displacement from these areas have changed the nature of the rela- tionship tribals share with their land. This, in part, may help explain rising poverty among STs in some states. ADIVASIS: MAIN fiNDINGS Despite a decline in poverty rates, Adivasis in 2004-05 were 20 years behind the average. Poverty among STs (Adivasis) has fallen, but the gap between them and the average Indian is large and growing. The poverty headcount index for STs fell by 31 percent between 1983 and 2004-05, compared with a more rapid decline of 35 per- cent among SCs and an average overall decline in India of 40 percent (table 1.1 ). In 2004-05, a little less than half the ST population remained in poverty (44 percent), while, nationwide, the poverty rate had been reduced to almost one-quarter of the population (27.5 percent). STs in urban areas fared better than those in rural areas, with a lower poverty rate and steeper reductions since 1983. How- ever, given the low share of tribal population in urban centers, ST poverty rates as a whole remained closer to the rural average. The relatively slower declines in poverty among STs also meant that they were increasingly concentrated in the poorest deciles of the popula- tion; their poverty levels in 2004-05 resembled poverty levels expe- rienced by the average population 20 years earlier. More worryingly, in states with high tribal populations (more than 10 percent of total population), ST households exhibited poverty rates that were higher than the rates across the nation as a whole in 2004-05. In Orissa, almost 75 percent of ST households fell below the poverty line. One should read these numbers while keeping heterogeneity in mind. STs are a highly diverse category, and the heterogeneity within the category is so striking that often the same surname may belong to different tribes in different states. Outcomes among tribals have also been shaped by the region in which they live. Development outcomes among tribals in the northeastern states, for instance, OVERVIEW 11 Table 1.1 Poverty Rates: STs Are 20 Years behind the Average Population population below the poverty line, % Location, social %change group A. 1983 1993-94 B. 2004-05 (A, B) Rural ST 63.9 50.2 44.7 -30 sc 59.0 48.2 37.1 -37 Others 40.8 31.2 22.7 -44 All 46.5 36.8 28.1 -40 Urban ST 55.3 43.0 34.3 -38 sc 55.8 50.9 40.9 -27 Others 39.9 29.4 22.7 -43 All 42.3 32.8 25.8 -39 Total ST 63.3 49.6 43.8 -31 sc 58.4 48.7 37.9 -35 Others 40.5 30.7 22.7 -44 All 45.6 35.8 27.5 -40 Source: Staff estimates based on Schedule 1.0 of the respective NSS rounds and official poverty lines. appear quite different relative to the outcomes among tribals in central India. We therefore keep the former out of the purview of this diagnostic and draw upon microstudies, where relevant, to explain processes that lead to deprivation among STs. High child mortality is the starkest marker of tribal deprivation. Every monsoon season, the Indian media are rife with stories of child deaths in tribal areas, which are frequently reported as malnutrition deaths. Kalahandi District in Orissa, for instance, became a metaphor for starvation because of press reports dating back to the 1980s. Under-5 mortality remains a stark marker oftribal deprivation in India: nearly 96 tribal children die for every 1,000 live births, compared with an under-5 mortality of74 per 1,000 for all India. ST children make up 12 percent of all children under 5 in rural areas, but account for almost 23 percent of deaths in the 1-4 age-group (table 1.2). For all other social groups, the representation of deaths is proportionate to the representation of the group in the population. Previous analyses of the correlates of under-5 mortality did not find any difference with respect to what we already know: the effect of ST status is insignificant once we control for poverty; that is, poverty is the primary cause of excess child mortality among tribals. However, most analyses lump all under-5 mortality together. Our 12 POVERTY AND SOCIAL EXCLUSION IN INDIA Table 1.2 Adivasi Child Mortality Exceeds the Relative Population Share in Rural Areas percent Share of Share of all children Share of child deaths under-5 Social group under 5 age 1-4 deaths sc 21.6 28.1 24.6 ST 11.7 23.0 13.9 OBC 41.8 35.5 39.6 Other 24.9 13.4 21.9 Total 100 100 100 Source: Das, Kapoor, and Nikitin (2010), based on 2005 NFHS data. Note: The table refers to children born during the five years prior to the survey in rural areas only. Figure 1.1 Rural Adivasi Children: Lower Risk of Dying at Birth, buf Greater Risk by Age 5 120 100 -5..., 80 :.0 v E 60 0 0 o. .., 40 v ~ 20 0 ~ ~ \~<:- '" ~-- o<:- ~-- 0<:- -$" <:-'- § 0.5 0: ·- "' 'iii --5 !:l 0.4 ·- "U ::: v § c 0.3 ·- t: "' E g_·c 0.2 0 & 15.. &0.1 0+--------,---------.~------.--------,--------~ 1983 1987-88 1993-94 1999-2000 2004-05 year - - non-ST/SC men - - SC men ----- non-ST/SC women ········ ST men -·- ·- SC women ----- ST women Source: Calculations based on five-year NSS rounds on the population aged 15-59. it was the preserve of the upper castes and reflected an elaborate ideology that excluded Dalits from its pale. This has begun to change, especially among SC men, whose educational attainment above postprimary has grown at a pace similar to that of non-SC/ST women (figure 1.3 ). Nonetheless, even today, upper caste parents may not allow Dalit students to sit with their own children or to eat together, and other such issues are common in some parts of the country (Nambissan 2007). Historical stereotyping also affects the perfor- mance of Dalit students. Some years ago, Hoff and Pandey (2004 ), using controlled experiments with Dalit junior high school students in a village in Uttar Pradesh, found that beliefs shaped by a history of prejudicial treatment can have a significant impact on children's responses when opportunities are presented to them. Dalits are slightly more likely to participate in the labor force compared with non-SCs/STs, but these effects are stronger among rural women than among other groups. The analysis in this report looks at labor force participation rates separately for rural and urban areas and men and women. Among men, there is little variation; most men, regardless of caste status, report themselves employed. However, Dalit and Adivasi women in rural areas show much higher labor force participation rates than other rural women. The ethno- graphic evidence also points to the lower mobility restrictions on Dalit and Adivasi women, which, if combined with higher poverty 18 POVERTY AND SOCIAL EXCLUSION IN INDIA rates, makes it more likely that these women work outside the home. For the most part, Dalits do not own land and have historically been workers in the fields of landed castes. In 2004-05, according to NSS data on usual principal status, over 41 percent of Dalit men and 20 percent of Dalit women were engaged in casual labor com- pared with, respectively, 19 percent of non-SC/ST men and 8 per- cent of non-SC/ST women. Their landless status also excludes the Dalits from the large employment category of farm-based self- employment, and, within casual labor, the Dalits mostly remain farmworkers. At the multivariate level, too, the main Dalit effect is in the assignment to casual wage work. Predicted probabilities cal- culated from multinomial regression models that estimate the assignment of Dalit men to various employment groups show that Dalit men do not look so different from other men except that they are much more likely to be casual laborers. While many have received the benefits of public sector employ- ment, Dalits still lag behind non-Dalit/Adivasi/OBCs in regular salaried work and in nonfarm self-employment. The difference between Dalits and non-Dalits in regular salaried work is not huge, because, overall, salaried work forms a small proportion of all employment. At the univariate level, against about 17 percent non- SC/STs/OBCs, a little over 13 percent of Dalits are regular salaried workers. However, as the report shows, the real differences are not in the proportions of Dalits and non-Dalits in regular salaried work, but in the kinds of jobs the Dalits land even within salaried work. The assignment to low-end jobs in the salaried market leads to wage differentials as a result mainly of occupational segregation. Despite the changes, aggregate data indicate that Dalit men con- tinued to be in casual labor over the 20 years beginning in 1983. Shifts in labor force activity among Dalits show, overall, a slight decline in casual labor and a slight increase in both nonfarm self-employment and opting out of the labor force. Over the 20 years or so beginning in 19 8 3, while the proportion ofDalit men in casualla bor declined slightly (from 44.6 to 41.7 percent) and in nonfarm self-employment increased slightly (from 11.0 to 15.6 percent), these changes were small (figure 1.4 ). Dalit men are restricted to menial, low-paying, and often socially stigmatized occupations, while upper caste groups are concentrated in preferred occupations. Among women, the major change is the fact that they have moved out of casual labor, but are also withdrawing from the labor market. In general, while SCs were represented proportionately to their population overall at each employment level within central govern- ment services, they were vastly overrepresented in the least skilled OVERVIEW 19 Figure 1.4 A Small Labor Market Transition among Dalit Men: Out of Casual Labor into Self-Employment 100 90 80 70 ;:: 60 OJ u 1-; 50 OJ 0. 40 30 24 16 18 20 11 16 11 t: ~ 0 5 8 o/:: 10 0 :::::: 12 .·.·.· ,i, ~~. ~ ·... 15 :::::: u ·.·.· % 13 :::::: 18 :::::: :l ~ ?lt).. ~':l ~ ?lt).. ~':l ~ ?lt).. ~':l ":( ":( ¥ ":( ":( ¥ ":( ":( ¥ ~% ~0., "vr;::,c;::, ~% ...,_o.,OJ "vr;::,c;::, ...,_o.,'tl ~0., "vr;::,c;::, ST sc non-ST/SC population groups B not in labor force, unemployed II casual wage B agricultural self-employed D nonfarm self-employed 1:::1 regular Source: Staff calculations based on 2004-05 employment-unemployment schedule NSS data. Note: Statistics pertain to men between 20 and 65 years of age. occupational categories at the lowest employment level. In 2006, almost 60 percent of the sweepers in central government ministries were SCs (table 1.3 ), indicating that SCs are more likely to under- take ritually unclean, manual work, although qualitative and small- area studies indicate that this is changing in many places. Recent work suggests that there is also subtle stereotyping along caste lines in private sector hiring. New research establishes empirically that SC applicants face particular difficulty in passing through the screening questions (such as those on family background) set up by prospec- tive employers (Thorat and Attewell 2007; Jodhka and Newman 2007). Dalits also seem to have lower returns to education. Our report suggests that, while all men benefit from education, Dalit men in rural areas who have completed primary education or postprimary education are less likely to be employed. This may perhaps reflect, 20 POVERTY AND SOCIAL EXCLUSION IN INDIA Table 1.3 The Representation of Dalits in Government Jobs, 2006 percent Group SCs A 13.0 B 14.5 c 16.4 D (excluding sweepers) 18.3 Sweepers 59.4 Average across groups A-D, excluding sweepers 24.3 Average across groups A-D, including sweepers 15.6 Source: Calculated from data of Ministry of Personnel, Public Grievances, and Pensions (2009). Note: Data pertain to central government services as of January 1, 2006 and exclude two ministries. Group A indicates the highest level, and Group D, the lowest. first, that all men with education in rural areas are penalized, second, that Dalit men feel these effects especially if they have postprimary education, and, third, that the growth of jobs in rural areas does not keep pace with the increase in the supply of educated Dalit men. In urban areas, while all men show a much higher likelihood of par- ticipating in the labor force if they have education, the effects among Dalit men are not statistically significant. The wage differentials between Dalits and others are a testimony to the continued disadvantage of Dalits in the labor market. The influence of caste affiliation on wages is a much noted feature of the Indian labor market (see, for example, Banerjee and Knight 1985; Das 2006; Unni 2001). Kernel density plots of wages by social group in 2004-05 largely bear out the expectation that SC workers in regular jobs are less likely to hold the more well paying jobs. The distribution of regular wages for general caste workers lies to the right of the distribution for SC and OBC workers (figure 1.5). Oaxaca-Blinder decompositions of wage differentials indicate that nearly 60 percent of the wage gap between SC and general caste workers is caused by unobserved factors or cannot be explained by human capital endowments. Glass walls? Occupational segregation can explain much of the Dalit disadvantage in the labor market. The concentration of Dalits in casual work or in low-pay occupations compared with other groups is linked to relative differences in educational level, but this is only part of the story. Das and Dutta (2008) propose the notion of glass walls, whereby occupationally slotted castes cannot leave their traditional trades or jobs. Castes are clustered around occupa- tions, whether by choice or compulsion. Microlevel studies, for OVERVIEW 21 Figure 1.5 Wage Differentials between Dalits and Others Are Higher in Salaried Work than in Casual Work a. Regular workers b. Casual workers 0.8 0.8 "' bn "' bn "' ::: 0.6 "' ::: 0.6 >- >- ::;2 ::;2 "' 0.4 "' ::: "' "' ::: 0.4 ...... ...... 0 0 ,q ,q :g 0.2 :g 0.2 0) 0) " " 0 0 2 4 6 8 10 12 2 4 6 8 log weekly wage log weekly wage 1-- SC - - OBC ----·genera! I Source: Das and Dutta (2008), based on NSS data, 61st round, 2004-05. instance, point to the possibility that small-scale Dalit entrepreneurs, especially in rural areas, are being prevented from moving out of caste-based occupations into self-employed ventures through social pressures and ostracism (Thorat 2007). There is also a possibility that people may not want to leave their caste networks because they may be advantageous in responding to new opportunities (Munshi and Rosenzweig 2005; Kishwar 2002). The issue of Dalit entrepreneurship has been in the national dis- course for some time. Despite impressive microlevel studies of Dalit entrepreneurship, the national data do not seem to capture a shift into self-employment. It is also likely that the national data of the NSS date back to 2004, while changes that we are unable to capture may have taken place in the last six or seven years. Using the 61st round of the NSS, we find that Dalit men seem, overall, less likely to have their own enterprises, but self-employment is rare among most men in India, and the hurdles for small (often household) enterprises are well documented, most recently by Kishwar (2002) and Nilekani (2008). It seems that, in rural areas, Dalit men have a huge disad- vantage in self-employment. This could be caused as much by the fact that self-employment is rare in rural areas and that, where it exists, it may be ancillary to agriculture, and Dalits typically do not own land. Equally, however, rural Dalit men have fewer social net- works and less access to credit, markets, and raw materials than 22 POVERTY AND SOCIAL EXCLUSION IN INDIA their urban counterparts. In a study on Dalit entrepreneurship com- missioned for this report, Jodhka and Gautam (2008) find that, while those who venture into nonfarm activities are able to circumvent caste norms, significant barriers exist in accessing networks, particu- larly social networks that would enable access to credit. Poor access to networks reflects in national data as well. Only 12 percent of SC households had access to three contacts in the formal sector in 2005 compared with 26 percent among forward caste households (see chapter 3, figure 3.6). There are cracks in the glass walls, but relative disadvantages remain. In sum, we find evidence of subtle changes in education out- comes and the labor market among the Dalits, particularly Dalit men. However, these changes appear to be mostly cracks in the glass wall. Mobility among the Dalits is constrained largely because of the poor initial conditions Dalits face (for instance, lack of assets and poor access to markets). IN SUMMARY • There has been some convergence in education outcomes, par- ticularly in postprimary education, between Dalit and non-SC/ ST men. However, while the advance has been impressive, Dal- its still lag behind because of their low starting points. Dalit women, in particular, do worse, faring only marginally better than Adivasi women. • Dalits are slightly more likely to participate in the labor mar- ket. However, they remain largely in casual jobs. Over time, there has been only a slight shift away from casual labor into nonfarm employment among men. More Dalit women, how- ever, seem to be withdrawing from the labor force, perhaps on account of more education or social mobility. • The picture that emerges in the employment options for edu- cated Dalit men is more complex. In rural areas, education mostly hampers the employment opportunities for Dalits, mainly because there are few regular salaried jobs available. In urban areas, too, the combined effects of caste and education indicate that SC men have fewer chances (relative to other groups) of exiting casual labor and moving into regular salaried jobs if they have postprimary education. This may be a corol- lary of an increasing supply of educated SC men over time, thereby creating a system of job rationing among SCs, who cannot compete in the nonreserved salaried job market. • The wage differentials between Dalits and others are a testi- mony to the continued disadvantage of Dalits in the labor mar- ket. Differences in access to occupations-or glass walls-are OVERVIEW 23 an important determinant of the wage gap. Recent research suggests that there is subtle caste-based stereotyping in private hiring. • In addition to new economic opportunities, Dalit solidarity movements and affirmative action policies have helped the Dal- its claim political space. In this, they have been more successful than other excluded groups. Women While gender per se is a relational category, and men are arguably disadvantaged along some indicators in some areas (for example, health and mortality), our report focuses on women-who, in India, are an excluded category-in major development outcomes and pro- cesses. 5 Female disadvantage is well documented and finds its ratio- nale, as does the caste system, in Hindu law books (Deshpande 2002). It plays out in several spheres of economic and social life: women's lower labor force participation and wages relative to men, poorer health and education outcomes, less voice in the political or general public arena, and less access to markets. In an infamously Indian pattern, we find that, in terms of sex ratios, India lags behind many countries at the same income level. This represents stark testimony of female disadvantage and the disincentive for parents to have daugh- ters, although there may be evidence of an incipient turnaround in some parts of India (Das Gupta, Chung, and Shuzhuo 2009). Yet, when girls and women do survive, they do better today than did the girls and women of the generation of their mothers, which is another way of saying that key indicators of gender equality are improving if one discounts the "missing women" (Sen 1992). The absolute levels of the indicators, however, continue to be poor, especially for Dalit and Adivasi women, who suffer from multiple disadvantages. Chapter 4 focuses on select outcomes among women. In particu- lar, it addresses women's survival disadvantage both in childhood and in their reproductive years, as well as some of the processes, such as marriage, that mediate women's access to markets, services, and spaces. In addressing women's voice and agency, the chapter also looks at an oft-omitted variable: women's vulnerability to violence or threats of physical harm in the home and outside the home. It cor- relates the likelihood of experiencing violence with a number of human development outcomes among women and their children. WOMEN: MAIN FINDINGS India has made substantial investments in human development, and the changes are evident. Today, most Indian families would not be able 24 POVERTY AND SOCIAL EXCLUSION IN INDIA to identify with the anachronistic Laws of Manu, which laid down the lower status of women. Indian women are now much more visible in public spaces and in positions of authority; fertility rates in several states are now below replacement levels and are similar to levels in developed countries; contraceptive prevalence is much higher than even a decade ago; and maternal mortality, while at stubbornly high levels across South Asia except in Sri Lanka, is showing a decline that is greater in India than in other countries. Also, there seems to be evi- dence of an incipient turnaround in reported son preference at least in some parts of India. Yet, the roots of gender inequality are still strong and affect a range of outcomes among women. Despite the strides in education and health, many outcomes are worse in India than in neighboring Bangladesh and relative to some other countries at a similar income level. The Human Development Index brings this to public attention annually. Female disadvantage is most starkly apparent in the lower survival chances of infant girls compared with infant boys. India and, to a lesser extent, Nepal are the only two countries in which the sur- vival rate among infant girls is lower than the rate among boys (figure 1.6). Declines in mortality rates overall have occurred more slowly in India than in Bangladesh and Nepal. Finally, the rate of progression of girls through to secondary school is much lower in India than in Bangladesh, although the latter spends a smaller pro- portion of gross domestic product on education (Das 2008). But, as in other areas, in gender inequality, too, India is highly heteroge- neous and Adivasi, Dalit, and Muslim women tend to show much poorer outcomes than other women. There are also large regional variations in most of these indicators. Adverse child sex ratios in many Indian states have received con- siderable attention, but is there evidence of an incipient turnaround? The strong preference for sons among Indian families plays out in the neglect of daughters and, over the last few decades, in the selec- tive abortion of female fetuses. This has led to massive outrage; Amartya Sen (1992) famously drew attention to the missing women in China and India. Das Gupta, Chung, and Shuzhuo (2009) draw comparative evidence from the Republic of Korea to show that the fertility preference for sons is declining and that this may well be followed by better childhood sex ratios in some parts of India, which could be on the cusp of a turnaround. Similar evidence is provided by John et al. (2008), who have carried out research in some of the districts with the poorest childhood sex ratios in India. At the core of this preference for sons seems to be a number of cultural practices, including the taboo against receiving financial help from daughters or considering daughters as old age insurance. Figure 1.6 Only in India and Nepal Is Infant Mortality Higher among Girls than among Boys = 29 ..s "' .s -B Q.J ·== -;;:;..o E g: -2:..:::: 1l 6 19 <"0 a;,_;' -;;:; '-< E 15. t:: a; a; "' a; ::: :. ~ '-< 9 a; >. ..0 ~ J1_, L. ullll ~,~ ~ ·,~ .,.~ ·,~ ~" ~- ·,~ ~" ~- 0 ~ . ~ ~ o.~ . ~ .:.~ ~' " 0 . "' " ~ 0 . ~ 0 ~ ~ ~ . ~ . ~ ~ ~ ~~ ..::;.8 ,..,~~ ~ <$'~~ o.v'- ~~ ~~~,<10~~ '<,~o;·,o,'-~~-"';Q,~ ~~ ~~00 ~~ l~' 0"<:$>~~' o,'-~ <.00 o'-G ~'-"o'1i o~" lt ~~4, oo" o<:i.' ~o.,~~ "'.~~~"1;-'- . .,.o~ ~<$':~0o ..~-,~~"''If~ ~<$'~~2Yo+o +~~'<.y~o; 0~'c.p ~ ~'<;~.;:;' v -9 ,. "~:- ,}!<' . "'-o"' v "'"' . "'-'~>' c? ..,.. "'~~ c? O' ~'-' ~ "';:'' "';:'' v UO~~ country Source: Selected Demographic and Health Surveys, 2003-06. N v, 26 POVERTY AND SOCIAL EXCLUSION IN INDIA Together with high dowry rates, this implies that daughters are per- ceived as a net loss in economic terms. In such a scenario, the easy availability of technology that detects the sex of the child allows families to abort female fetuses. However, not all states fare so badly. Andhra Pradesh, Assam, Kerala, and West Bengal show better indi- cators. In Haryana and Punjab, the sex ratios are so adverse that there is now a shortage of marriageable women, and families with sons have to import daughters-in-law from across the country and across caste lines. Whether these trends will have longer-term conse- quences is yet to be determined (Kaur 2004). Despite many positive demographic outcomes among women, childbearing remains a high-risk event. Indian women face a 1 in 70 risk of dying in childbirth. This falls at the high end of the global spec- trum, and other comparators have much better outcomes. For instance, Chinese women face a 1 in 1,400 risk of maternal death, while the risk among Vietnamese women is 1 in 280. Within India, there is consider- able variation across states, and some have done worse than others over the first few years of the new millennium. Assam, Bihar, Chhat- tisgarh, Jharkhand, Madhya Pradesh, Orissa, Rajasthan, Uttaranchal, and Uttar Pradesh together accounted for about 65 percent of all maternal deaths in 1997-2003 (Registrar General 2006). Childbear- ing is entangled with many cultural factors such as the early age of marriage, which, combined with poor access to health care, may be responsible for large numbers of maternal deaths. The government is cognizant of the slow pace of decline in maternal mortality overall, and the National Rural Health Mission has put in place several inter- ventions to address the issue (Registrar General 2006). Less than half of Indian women receive complete antenatal care, and 60 percent of all childbirths take place at home. A majority of the women who give birth at home feel it is not necessary to deliver in a formal medical facility. The public health system has agonized over the low demand for maternal health services; cultural and behavioral fac- tors have been blamed (see Basu 1990 ). Medical practitioners often cite ignorance as the reason for the poor outcomes among women (Khan- dare 2004 ). It is true that, among the women who gave birth at home, a majority felt that to give birth in a medical facility was not necessary (see chapter 4, table 4.3). However, the low demand for health care may also be triggered by gaps in supply, the inability to reach a health center in the moment of need, and the lack of information on whether the health centers would be open. Violence could be one of the omitted variables that explains why many women do not access reproductive health services. According to the 2005 NFHS, over one-third of Indian women reported that they had experienced spousal violence at some point, and about one-fourth OVERVIEW 27 had experienced violence in the previous year. Violence against women is a marker of extreme inequality in gender relations, and addressing the problem has an intrinsic value because it has important implica- tionsfor human rights, but addressing it also has an instrumental value. Nearly 81 percent of the women who reported that they had never experienced violence also said that they had received antenatal care; in contrast, only 67 percent of those who had experienced violence had received antenatal care. At the multivariate level, after controlling for wealth quintile, educational levels, and other background character- istics, one finds that the experience of spousal violence increases the odds of non-live births, lack of antenatal care, lack of iron supple- ments and tetanus shots before delivery, stunting among children under the age of 5, and diarrhea for the last child to be born in families (figure 1.7). Our report also highlights the fact that access to land, assets, and education among both men and women and campaigns to change attitudes that condone violence could go a long way to addressing the prevalence of spousal violence. The labor market is one of the most important sites of gender inequality. Women's visibility in the high-end urban labor market in India is much greater today than it was a decade ago, but this is far Figure 1. 7 The Experience of Violence Is Associated with Worse Outcomes among Women and Their Children child had diarrhea child stunted (last birth) given iron supplements in pregnancy woman had tetanus injection before birth one or more antenatal visits (last pregnancy) 0 0.2 0.4 0.6 0.8 1.2 1.4 1.6 odds ratios from logistic regression models controlling for individual and socioeconomic characteristics Source: Staff calculations based on 2005 NFHS data. 28 POVERTY AND SOCIAL EXCLUSION IN INDIA from representing a trend that may be considered general. Female labor force participation rates have remained low; only 40 percent of women are employed in full-time work (NSS 2004-05). The stag- nation in female labor force participation is driven mainly by rural areas; urban areas have seen a 7 percent increase, from 20.3 percent in 1993 to 21.8 percent in 2005. There is also considerable diversity by state and by caste and tribal status. Women in the more rapidly growing regions where the norms of female mobility are more lib- eral, such as the south and west, are much more likely to be employed than are women in the central states. Similarly, SC and ST women participate more in the labor force out of necessity; the latter mostly take up self-employment in agriculture, while SC women undertake casual labor. Entry into the labor market, rather than wages, is the critical marker of the employment trajectories of Indian women. This is because most Indians do not work for wages anyway, and self- employment is the default option where it is economically feasible. Casual labor is the option of last resort. In a pattern peculiar to India and much of South Asia, education lowers the likelihood that women will participate in the labor force. Two arguments have been articu- lated to explain the low labor force participation of Indian women. The first i~ a supply-side argument according to which the supply of secure well-paid jobs for educated women is low. Hence, educated women, who also belong to the higher socioeconomic strata, prefer to opt out of the labor force rather than accept low-status (manual) jobs. The second is a demand-side argument that rests on the cul- tural mores and values of status and seclusion in the region; this may prevent higher-status households from allowing women to work or demand jobs. Family honor in most parts of India, for instance, requires that women be restricted to the home, thus affecting the ability of women to work outside the home (Chen 1995). Muddy waters: an income effect or a discouraged worker effect? The literature on women's labor force participation, particularly in developed countries, indicates that women's employment decisions are often contingent upon the employment status and earnings of hus- bands (see Cohen and Bianchi 1999). Das (2006) finds that the educa- tion and wages of husbands do lower the probability of employment among women. Moreover, after controlling for the incomes of hus- bands, one finds that women's postprimary education shows a positive correlation with women's labor force participation. In fact, it takes the form of a "U," with high labor force participation among uneducated women, the lowest labor force participation among women who have completed primary education, and rising labor force participation among women with postprimary education. OVERVIEW 29 Thus, women with higher education may perhaps stay out of the labor force because of an income effect, but this conclusion is muddied by the lack of jobs that these women would want to take. In the absence of regular salaried jobs, the only options available to educated women, especially in rural areas, are low-status, low-paying manual work, such as work on family farms or work as petty vendors, domestic servants, or day laborers. In the face of such unsuitable employment opportunities, households decide to withdraw female labor if there are other earning members. (If women are household heads, they are more likely to be employed.) Because educated women are usually married to edu- cated men and are likely to have some financial resources, they stay out of the labor force instead of accepting poorly paid jobs as casual wage workers. Inequalities in wages and opportunities are an additional disin- centive for women to work. While wages have risen in the aggregate for all over the last 10 years or so, lower wages among women com- pared with men are an added disincentive for women to work outside the home. Women's nominal weekly wages are, on average, 71 percent of men's wages in regular salaried work and 56 percent of men's wages in casual work (NSS 2004-05). Oaxaca-Blinder decompositions con- ducted on the wages of male and female casual workers in Bangladesh and India indicate that unobserved factors account for over 70 percent ofthedifferenceinwages(Das2006;WorldBank2008).0ftheseunob- served factors, a large proportion are likely to be accounted for by dis- crimination. Low and unequal wages and the concentration of these in agricultural labor and in" female" occupations are added disincentives for women to work outside the home. Counting aspirations: Indian women clearly want to work out- side their homes. One of the explanations for women's low partici- pation in the labor market is that they would prefer not to work, given the pervasive culture of seclusion, especially in states where women's employment rates are low (see Das 2006). We find quite the reverse in women's responses to the question in the NSS about why they did not work outside the home. Over 89 percent of the women are doing only domestic work say they are obliged to do so. One-third say they would accept paid work, in addition to their household duties. Clearly, their household responsibilities are para- mount, but the majority say they would like regular part-time jobs in, for example, dairying or tailoring. Of these, almost 60 percent cite the lack of finance and credit as the most important constraint on finding the work they desire (table 1.4 ). Further analysis using 2005 NFHS data suggests that less than 40 percent know about the existence of credit facilities; of these, only 10 percent had actually applied for the credit. To a large extent, this may explain why women 30 POVERTY AND SOCIAL EXCLUSION IN INDIA Table 1.4 Three in Five Women Cite Lack of Credit as a Reason for Not Doing the Work They Want percent What do you need to facilitate the work you want? Share Nothing 4.24 Finance and credit 58.50 Raw materials 3.68 Assured markets 7.58 Training 15.63 Accommodation 0.87 Other 9.50 Source: Staff calculations based on the 6lst NSS round, 2004-05. Note: The sample consists of women who are currently doing household chores, but who say they would also like to do market work. are so poorly represented in nonfarm employment and why they are transitioning so slowly out of agriculture. Voice and visibility can change the outcomes among women. Data of the 2005 NFHS indicate that, at home, women do not have much voice in major decisions. For instance, they are least likely to participate in decisions about major household purchases and more likely to participate in decisions regarding their own health care or visits to their own families. Mandatory legal provisions reserving seats in legislatures for women have enabled women's participation in public spaces. However, the evidence on whether this has helped improve outcomes among a majority of women is mixed. One of the factors that hinders women's visibility and voice is the threat of physical harm and lack of security outside the home. Threats to women's security also influence the ability of women to access markets and services and claim spaces for themselves. This is an area in which policy can have a huge effect. Making public spaces safe for women is a major step forward in enhancing women's access to these spaces. IN SUMMARY • Women today are doing better than their mothers' generation along a range of outcomes that include health, education, voice and visibility, and, to an extent, participation in the urban labor market. • There are large inequalities among states in most areas of wom- en's status and gender equity. OVERVIEW 31 • Poor access to reproductive health means that too many women die unnecessarily in childbirth. • Violence against women is a strong correlate of a number of poor outcomes among women and their children. • Despite a period of dramatic economic growth, labor force par- ticipation rates among women virtually stagnated from 1983 to 2004-05. This is driven especially by rural areas. • Heterogeneity across social groups is pronounced, and espe- cially Dalit and Adivasi women are left out of nonfarm self- employment. • Women appear to be stuck in farm-based employment; a major r~ason why they are not transitioning into nonfarm self- employment is their poor access to credit facilities. • Over time, women's voice and agency both in the home and in public have increased, but the extent to which this has impacted aggregate outcom~s is unclear even at the local level. Common Themes This report has several common themes and messages, as follows: • First, it shows that, while growth has touched everyone, it has not done so equitably; traditional hierarchies have remained stubborn against growth. In the aggregate, STs appear to have done more poorly than other groups; they show the slowest pace of improvements in a range of areas. • Second, although caste seems to be reinventing itself in response to economic opportunities and is far from a static stereotype, we find that SCs are still held back by, among other factors, their initial disadvantage and lack of social networks. • Third, female disadvantage in India persists despite high rates of economic growth. Women are dying unnecessarily both in infancy and in motherhood; the outcomes are poorer among Dalits and Adivasis. At its root, exclusion can be explained by inequality in opportuni- ties, inequality in access to markets (for example, labor and credit), and inequality in voice and agency. Voice and agency have played out in different ways among all three groups (Adivasis, Dalits, and women), but are particularly salient drivers of exclusion. For instance, violence against women is not merely a variable that explains some of the poor outcomes; it is also a mechanism of control. Among the Adivasis and, in the past, the Dalits, militancy has represented a form of voice and assertion. 32 POVERTY AND SOCIAL EXCLUSION IN INDIA India is not alone in grappling with serious challenges in reaching its most excluded populations. In other countries as well, this chal- lenge is a formidable one. An Inter-American Development Bank report on Latin America states that "inclusion is not just about changing outcomes, but crucially about changing the processes that produce and reproduce exclusionary outcomes" and that "in order to make normative changes effective, institutions must change the ways in which they operate, hire employees, and enforce laws and regulations. This in turn materializes as changes in the implementa- tion of programs and policies" (Marquez et al. 2007, 14 ). The Indian Constitution has set the stage for almost unparalleled affirmative action and other forms of positive actions. These have been translated into laws, programs, and procedures. Yet, the com- bination of identity politics, the inflexibility of the systems that seek to promote inclusion, and the attendant poor implementation have resulted in patchy impact, affecting some groups more than others. To describe the real challenge is to state a truism: the implementation of policies and the reform of institutions are the key to ensuring that economic growth becomes more equitable. Notes 1. Coined by anthropologist Oscar Lewis, the term 'culture of poverty' refers to a unique value system of the poor. It was Lewis's belief that the poor are socialized into believing they deserve to be poor, leading to low aspirations, low effort and inability to escape poverty. 2. For instance, improvements in antenatal care have been particularly slow among OBC women. 3. See http://www.tribal.nic.inlindex.asp. 4. The Hindu hierarchy is said to have evolved from different parts of the body of Brahma, the creator of the universe. Thus, the Brahmans, who originated from the mouth of Brahma, undertake the most prestigious priestly and teaching occupations. The Kshatriyas (from the arms) are the rulers and warriors; the Vaishyas (from the thighs) are traders and mer- chants; and the Shudras (from the feet) are manual workers and servants of other castes. Below the Shudras and outside the caste system, the lowest in the order, the untouchables, engage in the most demeaning and stigma- tized occupations (scavenging, for instance, and dealing in bodily waste). 5. It is acknowledged, however, that male-female differences along a range of indicators need to be considered, and more so if one is thinking of designing interventions to lessen the gender gap. OVERVIEW 33 References Akerlof, G. 1976. "The Economics of Caste and of the Rat Race and Other Woeful Tales." Quarterly Journal of Economics 90 (4 ): 599-617. Ambedkar, B. R. 1936. Annihilation of Caste, 2nd ed. Mumbai: Govern- ment of Maharashtra. Baker,]. L. 2002. "Social Exclusion in Urban Uruguay." En Breve 2 (April). World Bank, Washington, DC. Banerjee, B., and J. B. Knight. 1985. "Caste Discrimination in the Indian Urban Labour Market." Journal of Development Economics 17 (3): 277-307. Basu, A.M. 1990. "Cultural Influences on Health Care Use: Two Regional Groups in India." Studies in Family Planning 21 (5): 275-86. Beteille, A. 1998. "The Idea of Indigenous People." Current Anthropology 39 (2): 187-191. Chen, M. 1995. "A Matter of Survival: Women's Right to Employment in India and Bangladesh." In Women, Culture and Development: A Study of Human Capabilities, ed. M. C. Nussbaum and J. Glover, 37-58. Oxford: Clarendon Press. Cohen, P. N., and S.M. Bianchi. 1999. "Marriage, Children and Women's Employment: What Do We Know?" Monthly Labor Review 122 (12): 22-31. Das, M. B. 2006. "Do Traditional Axes of Exclusion Affect Labor Market Outcomes in India?" South Asia Social Development Discussion Paper 3, World Bank, Washington, DC. - - - . 2008. "What Money Can't Buy: Getting Implementation Right for MDG3 in South Asia." In Equality for Women: Where Do We Stand on Millennium Development Goal3?, ed. M. Buvini, A. R. Morrison, A. W. Ofosu-Amaah, and M. Sjoblom, 261-92. Washington, DC: World Bank. Das, M. B., and P. Dutta. 2008. "Does Caste Matter for Wages in the Indian Labor Market? Caste Pay Gaps in India." Paper presented at the Third Institute for the Study of Labor-World Bank Conference on Employment and Development, Rabat, Morocco, May 5-6. Das, M. B., S. Kapoor, and D. Nikitin. 2010. "A Closer Look at Child Mortality among Adivasis in India." Policy Research Working Paper 5321, World Bank, Washington, DC. Das Gupta, M., W. Chung, and L. Shuzhuo. 2009. "Is There an Incipient Turnaround in Asia's 'Missing Girls' Phenomenon?" Policy Research Working Paper 4846, World Bank, Washington, DC. de Haan, A. 1997. "Poverty and Social Exclusion: A Comparison of Debates." PRUS Working Paper 2, Poverty Research Unit, University of Sussex, Brighton, United Kingdom. 34 POVERTY AND SOCIAL EXCLUSION IN INDIA Deshpande, A. 2002. "Assets Versus Autonomy? The Changing Face of the Gender-Caste Overlap in India." Feminist Economics 8 (2): 19-35. Gacitua-Mari6, E., and M. Woolcock, eds. 2008. Social Exclusion and Mobility in Brazil. Washington, DC: World Bank. Government oflndia. 2008. "Development Challenges in Extremist Affected Areas: Report of an Expert Group to Planning Commission." Report, April, Planning Commission, New Delhi. http://planningcommission .gov. in/reports/publications/rep_dee. pdf. Hoff, K., and P. Pandey. 2004. "Belief Systems and Durable Inequalities: An Experimental Investigation of Indian Caste." Policy Research Working Paper 3351, World Bank, Washington, DC. Jodhka, S. 2008. "A Forgotten 'Revolution': Revisiting Agrarian Change in Haryana." Paper prepared for the study, Poverty and Social Exclusion in India, Indian Institute of Dalit Studies, New Delhi. Jodhka, S., and S. Gautam. 2008. "In Search of a Dalit Entrepreneur: Bar- riers and Supports in the Life of Self-Employed Scheduled Castes." Paper prepared for the study, Poverty and Social Exclusion in India, Indian Institute of Dalit Studies, New Delhi. Jodhka, S., and K. S. Newman. 2007. "In the Name of Globalisation: Mer- itocracy, Productivity and the Hidden Language of Caste." Economic and Political Weekly 42 (41): 4125-32. John, M. E., R. Kaur, R. Palriwala, S. Raju, and A. Sagar. 2008. Planning Families, Planning Gender: The Adverse Child Sex Ratio in Selected Dis- tricts of Madhya Pradesh, Rajasthan, Himachal Pradesh, Haryana, and Punjab. New Delhi: ActionAid and International Development Research Centre. Kaur, R. 2004. "Across-Region Marriages: Poverty, Female Migration and the Sex Ratio." Economic and Political Weekly 39 (25): 2595-2603. Khandare, L. 2004. "Korku Adivasis in Melghat Region of Maharashtra: A Socio-economic Study; a Course Seminar." Department of Humanities and Social Sciences, Indian Institute of Technology, Mumbai. http:// lalitreports.blogspot.com/2004/12/korku-adivasis-in-melghat-region-of .html. Kishwar, M. 2002. "Working under Constant Threat: Some Setbacks and Some Steps Forward in Sewa Nagar." Manushi 130 (May-June). Maharatna, A. 2005. Demographic Perspectives on India's Tribes. New Delhi: Oxford University Press. Marquez, G., A. Chong, S. Duryea, J. Mazza, and H. Nopo, eds. 2007. Outsiders: The Changing Patterns of Exclusion in Latin America and the Caribbean. Economic and Social Progress in Latin America 2008. Wash- ington, DC: Inter-American Development Bank. Ministry of Personnel, Public Grievances, and Pensions. 2009. Annual Report 2008-2009. New Delhi: Ministry of Personnel, Public Griev- ances, and Pensions. OVERVIEW 35 Munshi, K., and M. Rosenzweig. 2005. "Why Is Mobility in India So Low? Social Insurance, Inequality, and Growth." Draft working paper, Center for International Development, Harvard University, Cambridge, MA. Nambissan, G. B. 2007. "Exclusion, Inclusion and Education: Perspectives and Experiences of Dalit Children." Paper prepared for the Indian Insti- tute of Dalit Studies, New Delhi. Nehru,]. 1946. The Discovery of India. Calcutta: The Signet Press. Nilekani, N. 2008. Imagining India: Ideas for the New Century. New Delhi: Penguin Books India. North, D. 1990. Institutions, Institutional Change and Economic Perfor- mance. Cambridge: Cambridge University Press. Prasad, C. B. 2009. "Sir, the World Turned Upside Down: Radical Changes in Dalits' Occupations, Food Habits and Lifestyle Observed in Two Blocks of UP." Paper presented at the Planning Commission of India, New Delhi, November 5. Registrar General. 2006. "Sample Registration System; Maternal Mortality in India, 1997-2003: Trends, Causes and Risk Factors." Working Paper 753, Registrar General, New Delhi. Scoville,]. G., ed. 1991. Status Influences in Third World Labor Markets: Caste, Gender, Custom. New York: Walter de Gruyter. Sen, A. 1992. "Missing Women: Social Inequality Outweighs Women's Sur- vival Advantage in Asia and North Africa." British Medical journal 304 (6827): 586-87. - - - . 1998. "Social Exclusion: A Critical Assessment of the Concept and Its Relevance." Paper prepared for the Asian Development Bank, Manila. Silver, H. 1994. "Social Exclusion and Social Solidarity: Three Paradigms." IlLS Discussion Papers 69, International Labour Office, Geneva. Thorat, S. 2007. "Economic Exclusion and Poverty: Indian Experience of Remedies against Exclusion." Paper presented at the International Food Policy Research Institute and Asian Development Bank policy forum, "Agricultural and Rural Development for Reducing Poverty and Hunger in Asia: In Pursuit of Inclusive and Sustainable Growth," Manila, August 9-10. Thorat, S., and P. Attewell. 2007. "The Legacy of Social Exclusion: A Cor- respondence Study of Job Discrimination in India." Economic and Polit- ical Weekly 42 (41): 4141-45. Tilly, C. 1999. Durable Inequality.Berkeley, CA: UniversityofCaliforniaPress. Unni,]. 2001. "Earnings and Education among Ethnic Groups in Rural India." NCAER Working Paper 79, National Council of Applied Eco- nomic Research, New Delhi. Walton, M. 2007. "Culture Matters for Poverty, but Not Because of a Culture of Poverty: Notes on Analytics and Policy." Unpublished paper, 36 POVERTY AND SOCIAL EXCLUSION IN INDIA Centre for Policy Research, New Delhi; John. F. Kennedy School of Government, Harvard University, Cambridge, MA. http://www.afd.fr/ j ahia/we bda v/users/ admin istra teur/public/ eudn2007/walton. pdf. Weber, M. 1958. The Religion of India: The Sociology of Hinduism and Buddhism. New York: The Free Press. World Bank. 2005. Poverty, Social Exclusion, and Ethnicity in Serbia and Montenegro: The Case of the Roma. Washington, DC: World Bank. - - - . 2006. "Unequal Citizens: Gender, Caste and Ethnic Exclusion in Nepal." National Planning Commission (Nepal), Department for Inter- national Development (United Kingdom), World Bank, Kathmandu. - - - . 2008. Whispers to Voices: Gender and Social Transformation in Bangladesh. Washington, DC: World Bank. - - - . 2010. "Economic Costs of Roma Exclusion." Note, April, Europe and Central Asia, Human Development Department, World Bank, Washington, DC. - - - . 2011. Perspectives on Poverty in India: Stylized Facts from Survey Data. India Poverty Assessment. New Delhi: Poverty Reduction and Economic Management Network, World Bank. Four easy ways to order Phone: Mail: Customer outside the US: Online: Fax: Pubhcalions.worldbank.org +1-703-661-1501 +1-703-661-1580or P.O.Box960 publications.worldbank.orglbookse!lers 1-800-645-7247 Herndon, VA20172~. USA Subtotal * Geographic discounts apply- dependmg on ship-to country. See http:!lpu blications. worldbank. orgldiscounts Geographic discount* **Within the US, charges on prepaid orders are $8 per order, plus $1 per Item. Institutional customers using a purchase order will be charged actual shipping costs. Outside of the US, customers have the option to choose Shipping and Handling** between nontrackab/e airmail delivery (US$7 per order plus US$6 per item) and trackab/e couriered airmar1 delivery (US$16.50 per order plus US$8 per item). Nontrackable delivery may take 4-6 weeks; trackab/e delivery takes Total US$ about 2 weeks MAILING ADDRESS METHOD OF PAYMENT Name 0 Charge my Organization Ovisa 0 Mastercard 0 American Express Address Credit card number City Expiration date State Zip Name Country Signature Phone 0 Enclosed is my check in US$ drawn on a U.S. Fax bank and made payable to the World Bank Email For information on geographical discounts please visit publications.worldbank.org/discounts THANK YOU FOR YOUR ORDER! 1I"""• ~::,:) ~, wP ObR 1n RANK Ll 1ICiHIOI\'> I The reference of choJcc on development