2012 world development report O v e rv i e w Gender Equality and Development 2012 world development report Gender Equality and Development 2012 world development report Gender Equality and Development Overview © 2011 The International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org All rights reserved 1 2 3 4 14 13 12 11 This document summarizes the World Development Report 2012. It is a product of the staff of the International Bank for Reconstruction and Development / The World Bank. The find- ings, 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 / 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, Dan- vers, 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. Cover photo: Arne Hoel, World Bank Photo credits: World Bank Cover design: Critical Stages Figures design and infographics: Design Symphony, Cymetrics, Harkness Design, and Naylor Design Contents Foreword   vii Acknowledgments   ix Overview Why does gender equality matter for development?   2 What does this Report do?   6 Where has there been the most progress in gender equality?   8 Where have gender inequalities persisted and why?   13 What is to be done?   22 The political economy of reforms for gender equality   35 A global agenda for greater gender equality   36 Notes   38 References   40 v Foreword The lives of girls and women have changed dramatically over the past quarter century. Today, more girls and women are literate than ever before, and in a third of developing countries, there are more girls in school than boys. Women now make up over 40 percent of the global labor force. Moreover, women live longer than men in all regions of the world. The pace of change has been astonishing—indeed, in many developing countries, they have been faster than the equivalent changes in developed countries: What took the United States 40 years to achieve in increasing girls’ school enrollment has taken Morocco just a decade. In some areas, however, progress toward gender equality has been limited—even in devel- oped countries. Girls and women who are poor, live in remote areas, are disabled, or belong to minority groups continue to lag behind. Too many girls and women are still dying in child- hood and in the reproductive ages. Women still fall behind in earnings and productivity, and in the strength of their voices in society. In some areas, such as education, there is now a gen- der gap to the disadvantage of men and boys. The main message of this year’s World Development Report: Gender Equality and Develop- ment is that these patterns of progress and persistence in gender equality matter, both for development outcomes and policy making. They matter because gender equality is a core development objective in its own right. But greater gender equality is also smart economics, enhancing productivity and improving other development outcomes, including prospects for the next generation and for the quality of societal policies and institutions. Economic develop- ment is not enough to shrink all gender disparities—corrective policies that focus on persist- ing gender gaps are essential. This Report points to four priority areas for policy going forward. First, reducing gen- der gaps in human capital—specifically those that address female mortality and education. Second, closing gender gaps in access to economic opportunities, earnings, and productivity. Third, shrinking gender differences in voice and agency within society. Fourth, limiting the reproduction of gender inequality across generations. These are all areas where higher incomes by themselves do little to reduce gender gaps, but focused policies can have a real impact. Public actions need to address the underlying determinants of gender gaps in each prior- ity area—in some cases, improving service delivery (especially for clean water, sanitation, and maternal care), for others, tackling constraints that originate in the workings of markets and institutions to limit progress (for example, in reducing gender gaps in earnings and productivity). Development partners can complement public action. In each of the four priority areas, efforts need more funding (particularly to support the poorest countries as they address female mortality and gender gaps in education); better gender-disaggregated data; more experimentation and systematic evaluation; and broader partnerships that include the private sector, development agencies, and civil society organizations. vii viii F O R E WO R D Gender equality is at the heart of development. It’s the right development objective, and it’s smart economic policy. The World Development Report 2012 can help both countries and international partners think through and integrate a focus on gender equality into develop- ment policy making and programming. Robert B. Zoellick President The World Bank Group Acknowledgments This Report has been prepared by a core team led by Ana Revenga and Sudhir Shetty, and comprising Luis Benveniste, Aline Coudouel, Jishnu Das, Markus Goldstein, Ana María Muñoz Boudet, and Carolina Sánchez-Páramo. Research assistance was provided by Rabia Ali, María Inés Berniell, Rita Costa, Nina Rosas, and Lucía Solbes Castro. The multi- country qualitative assessment was coordinated by Patti L. Petesch and Carolyn Turk. Extensive and valuable contributions were made by Andre Croppenstedt, Malcolm Ehrenpreis, Rebekka Grun, Mary Hallward-Driemeier, Tazeen Hasan, Karla Hoff, Ghazala Mansuri, Claudio E. Montenegro, and Bob Rijkers. The World Development Report 2012 is co-sponsored by the Development Econom- ics Vice-Presidency (DEC) and the Poverty Reduction and Economic Management Vice- Presidency (PREM). The work was conducted under the joint guidance of Justin Yifu Lin in DEC and Otaviano Canuto dos Santos Filho in PREM. Ann E. Harrison and the DEC team and Mayra Buvinic and the PREM Gender (PRMGE) team provided valuable guidance and contributions at various stages of the production of this report. A panel of advisers comprising Bina Agarwal, Ragui Assad, Anne Case, Alison Evans, Raquel Fernández, Naila Kabeer, Ravi Kanbur, Santiago Levy, and Germano Mwabu provided excel- lent advice. Valuable comments and contributions were provided by Kathleen Beegle, Laura Chioda, Louise Cord, Maria Correia, Monica Das Gupta, Shantayanan Devarajan, Marianne Fay, Francisco H. G. Ferreira, Ariel Fiszbein, Indermit Gill, Alejandro Hoyos, Emmanuel Jime- nez, Elizabeth King, Andrew Mason, William Maloney, Ambar Narayan, Pierella Paci, Tara Vishwanath, and Michael Walton. Many others inside and outside the World Bank contrib- uted with valuable comments and input (their names are listed in the Bibliographical Note). World Bank President Robert B. Zoellick and Managing Directors Sri Mulyani Indrawati, Mahmoud Mohieldin, and Ngozi Okonjo-Iweala provided invaluable guidance and advice. The team benefited greatly from many consultations, meetings, and regional workshops held locally and in-country. These discussions included policy makers, civil society represen- tatives, academics, and development partners from Benin, Bolivia, Burkina Faso, Burundi, the Caribbean nations, the Central African Republic, Chile, Colombia, the Dominican Repub- lic, Georgia, Guatemala, India, Indonesia, Jordan, Kenya, Kuwait, Lebanon, Mali, Mexico, Morocco, Panama, Paraguay, Rwanda, Senegal, the Slovak Republic, South Africa, Sudan, Tanzania, Thailand, Togo, Turkey, Uganda, Uruguay, Vietnam, Zambia, and Zimbabwe. Consultations were also held at different stages of report preparation with representatives from multilateral and bilateral partners, including the Australian Agency for International Development (AUSAID), the Canadian International Development Agency (CIDA), the Inter-American Commission of Women-Organization of American States (CIM-OAS), the Danish International Development Agency (DANIDA), the U.K. Department for International Development (DFID), the Food and Agriculture Organization (FAO), the Japan International Cooperation Agency (JICA), the International Labour Organization (ILO), the Ministry for Foreign Affairs of Finland, MCC, NORAD, the Organisation for Economic Co-operation ix x AC K N OW L E D G M E N T S and Development-Development Assistance Committee (OECD-DAC) Gendernet, the Swiss Agency for Development and Cooperation (SDC), the Swedish International Development Cooperation Agency (SIDA), UN Women, the United Nations Children’s Fund (UNICEF), the United States Agency for International Development (USAID), and the United Nations Economic and Social Council (ECOSOC) 55th Commission on the Status of Women. The team would like to acknowledge the generous support of the Government of Nor- way through its Royal Ministry of Foreign Affairs, SDC, AUSAID, CIDA, the Government of Sweden through its Ministry for Foreign Affairs, the multi-donor Knowledge for Change Program (KCP), the Nike Foundation, the World Bank Nordic Trust Fund, and Fast Track Initiative Education Program Development Fund; as well as the in-kind support from JICA, DFID, and OECD. The team wishes to acknowledge the excellent support of the WDR production team com- prising Rebecca Sugui, Cecile Wodon, and Mihaela Stangu, and of the resource management team of Sonia Joseph and Evangeline Santo Domingo. We thank also Ivar Cederholm, Vivian Hon, Jimmy Olazo, and Irina Sergeyeva for their constant support. Other valuable assistance was provided by Gytis Kanchas and Nacer Mohamed Megherbi. Vamsee Krishna Kanchi, Swati P. Mishra, Merrell Tuck-Primdahl, and Roula Yazigi assisted the team with the website and communications. Bruce Ross-Larson was the principal editor. The Development Data Group contributed to the data appendix and was responsible for the Selected World Development Indicators. Design Symphony contributed to the design. World Development Report 2012: Gender Equality and Development Overview Baruani is reflecting on how women’s and men’s lives have changed over the past decade in Ijuhanyondo—a village in Tanzania. “Ten years back was terrible,” she recalls. “Women were very be- hind. They used to be only at home doing housework. But now, they are in businesses, they are in poli- tics.” Others hold similar views. “We do not depend a lot on men as it used to be,” says Agnetha. “We have some cash for ourselves, and this assists us in being free from men and to some extent controlling our lives.” In addition to managing their businesses, the women now make up half the members of the street committee that runs the village. Despite these positive changes, many challenges continue to weigh on women’s daily lives. Fewer than half the homes in the village have piped water. Even more difficult, Tungise and other women of the ­ village still fear violence by their partners: “When they are drunk, they can begin beating up women and children in the house. The worst bit of it is forcing sex with you.” Although legally women can inherit land or a house, tradition prevails. “Yes, women can inherit property,” says Flora, the executive secretary of the street committee. “In fact, in the will the father is supposed to give each son and daughter some- thing, and nowadays the law is strict, equally. But still, men give to their sons and argue that women have the property of where they are married.” Dodoma Rural Community Report, from “Defining Gender in the Century: Talking with Women and Men around the World: 21st ­ A Multi-Country Qualitative Study of Gender and Economic Choice” (World Bank 2011) Why does gender equality the law in such areas as property ownership, matter for development? inheritance, and marriage. In all, 136 countries now have explicit guarantees for the equality of The story of Ijuhanyondo village in Tanzania all citizens and nondiscrimination between men mirrors the evolution of gender equality across and women in their constitutions. the world over the past quarter century. Although Progress has not come easily. And it has not many women continue to struggle with gender- come evenly to all countries or to all women—or based disadvantages in their daily lives, things across all dimensions of gender equality. The have changed for the better—and at a pace that likelihood of women dying during childbirth would have been unthinkable even two decades in Sub-Saharan Africa and parts of South Asia ago. Women have made unprecedented gains in is still comparable to that in Northern Europe rights, in education and health, and in access to in the 19th century. A wealthy urban child in jobs and livelihoods. More countries than ever Nigeria—boy or girl—averages around 10 years guarantee women and men equal rights under of schooling, while poor rural Hausa girls aver- Overview 3 age fewer than six months. The rate at which advancement of women and has been ratified to women die relative to men is higher in low- and date by 187 countries. middle-income countries compared with their high-income counterparts, especially in the crit- Gender equality matters for ical years of infancy and early childhood and development—It is smart economics in the reproductive period. Divorce or widow- Gender equality matters also as an instrument hood causes many women to become landless for development. As this Report shows, gender and lose their assets. Women continue to clus- equality is smart economics: it can enhance eco- ter in sectors and occupations characterized as nomic efficiency and improve other develop- “female”—many of them lower paying. Women ment outcomes in three ways. First, removing are also more likely to be the victims of violence barriers that prevent women from having the at home and suffer more severe injuries. And same access as men to education, economic op- almost everywhere women’s representation in portunities, and productive inputs can generate politics and in senior managerial positions in broad productivity gains—gains all the more business remains far lower than men’s. important in a more competitive and globalized Do these patterns of gender inequality– world. Second, improving women’s absolute and in human and physical capital endowments, relative status feeds many other development in economic opportunities, and in the ability outcomes, including those for their children. to make choices to achieve desired outcomes Third, leveling the playing field—where women (agency)—matter, particularly those that per- and men have equal chances to become socially sist even as the development process unfolds? and politically active, make decisions, and shape This World Development Report (WDR) ar- policies—is likely to lead over time to more rep- gues that they do for two reasons. First, gender resentative, and more inclusive, institutions and equality matters intrinsically, because the abil- policy choices and thus to a better development ity to live the life of one’s own choosing and path. Consider each in turn. be spared from absolute deprivation is a basic human right and should be equal for everyone, Misallocating women’s skills and talent independent of whether one is male or female. comes at a high (and rising) economic cost Second, gender equality matters instrumentally, Gender equality can have large impacts on pro- because greater gender equality contributes to ductivity. Women now represent more than 40 economic efficiency and the achievement of percent of the global labor force, 43 percent of other key development outcomes. the agricultural workforce, and more than half of the world’s university students. For an econ- Gender equality matters in its own right omy to be functioning at its potential, women’s Following Amartya Sen, we see development skills and talents should be engaged in activities as a process of expanding freedoms equally for that make the best use of those abilities. But, all people.1 In this view of development, gen- as the stories of many women illustrate, this der equality is a core objective in itself (box is not always the case. When women’s labor is 1). So, just as development means less income underused or misallocated—because they face poverty or better access to justice, it should also discrimination in markets or societal institu- mean fewer gaps in well-being between males tions that prevents them from completing their and females. This viewpoint is also evident in education, entering certain occupations, and the international development community’s earning the same incomes as men—economic recognition that women’s empowerment and losses are the result. When women farmers gender equality are development objectives in lack security of land tenure, as they do in many their own right, as embodied in Millennium countries, especially in Africa, the result is lower Development Goals 3 and 5 (box 2). It is seen ­ access to credit and inputs and to inefficient as well in the adoption and widespread ratifica- land use, reducing yields. Discrimination in tion of the Convention on the Elimination of credit markets and other gender inequalities in All Forms of Discrimination against Women access to productive inputs also make it more (CEDAW). Adopted by the United Nations difficult for female-headed firms to be as pro- General Assembly in 1979, the convention es- ductive and profitable as male-headed ones. tablished a comprehensive framework for the And, when women are excluded from manage- 4 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Box 1 What do we mean by gender equality? Gender refers to the social, behavioral, and cultural attributes, that arise from circumstances beyond the control of individuals and expectations, and norms associated with being a woman or a man. those that stem from differences in preferences and choices. A sub- Gender equality refers to how these aspects determine how women stantial body of research documents such male-female differences in and men relate to each other and to the resulting differences in risk aversion, social preferences, and attitudes about competition. It power between them. follows that if men and women differ, on average, in attitudes, prefer- ­ This Report focuses on three key dimensions of gender equality ences, and choices, then not all observed differences in outcomes identified by men and women from Afghanistan to Poland to South can be attributed to differences in opportunities. Africa, as well as by researchers: the accumulation of endowments Those who argue for equality of outcomes argue that differ- (education, health, and physical assets); the use of those endow- ences in preferences and attitudes are largely “learned” and not ments to take up economic opportunities and generate incomes; inherent—that is, they are the result of culture and environment and the application of those endowments to take actions, or agency, that lead men and women to internalize social norms and expec­ affecting individual and household well-being. These are aspects of tations. Persistent differences in power and status between men equality where shortfalls of choice are reflected in shortfalls of wel- and women can become internalized in aspirations, behaviors, and fare. They matter in and of themselves. But they are also closely preferences that ­ perpetuate the inequalities. So, it is difficult to interlinked. define equality of opportunity without also considering how actual Gender inequality is both similar to and different from inequal- outcomes are distributed. Only by attempting to equalize out- ity based on other attributes such as race or ethnicity. Three differ- comes can one break the vicious circle of low aspirations and low ences are of particular relevance to the analysis of gender equality. opportunity. First, the welfare of women and men living in the same household Despite this debate, it is difficult in practice to measure oppor- is difficult to measure separately, a problem that is compounded by tunities separately from outcomes. Indeed, equality of opportuni- the paucity of data on outcomes in the household. Second, prefer- ties and equality of outcomes are tightly linked both in theory and ences, needs, and constraints can differ systematically between in measurement. For this reason, the Report takes a pragmatic men and women, reflecting both biological factors and “learned” approach, focusing on both outcomes and opportunities in relation social behaviors. Third, gender cuts across distinctions of income to endowments, agency, and access to economic activities. Follow- and class. These characteristics raise the question whether gender ing Sen, we also believe that while people may ­ disagree in what is equality should be measured as equality of outcomes or equality of just or fair, they will agree on eliminating what are “outrageously opportunity. The economic and philosophical literature on this unjust arrangements.” In other words, while it may be difficult issue is divided. to define whether gender equality is about outcomes or opportuni- Those who defend framing gender equality as equality of oppor- ties, most will agree that gross manifestations of gender inequality tunity argue that it allows one to distinguish between inequalities should be eliminated. Sources: Booth and Nolen 2009; Croson and Gneezy 2009; Gneezy, Leonard, and List 2009; Kabeer 1996; Sen 1999; World Bank 2011. The Box 2  Millennium Development Goals recognize the intrinsic and instrumental value of gender equality The 2010 Millennium Development Goal (MDG) Sum- equality and women’s empowerment are develop- mit concluded with the adoption of a global action ment objectives in their own right (MDG 3 and 5), as plan to achieve the eight goals by 2015. The summit well as serving as critical channels for achieving the also adopted a resolution calling for action to ensure other MDGs and reducing income and non-income gender parity in education and health, economic poverty. Gender equality and women’s empower- opportunities, and decision making through gender ment help to promote universal primary education mainstreaming in development policy making. The (MDG 2), reduce under-five mortality (MDG 4), resolution and the action plan reflect the belief of the improve maternal health (MDG 5), and reduce the international development community that gender likelihood of contracting HIV/AIDS (MDG 6). Source: WDR 2012 team. ment positions, managers are less skilled on function, is large: ensuring that women farmers average, reducing the pace of innovation and have the same access as men to fertilizer and other technology adoption.2 agricultural inputs would increase maize yields The direct payoff to correcting these failures, by 11 to 16 percent in Malawi and by 17 percent many rooted in how markets and institutions in Ghana.3 Improving women’s property rights Overview 5 in Burkina Faso would increase total household of countries (such as Bangladesh, Brazil, Côte agricultural production by about 6 percent, with d’Ivoire, Mexico, South Africa, and the United no additional resources—simply by reallocat- Kingdom) shows that increasing the share of ing resources (fertilizer and labor) from men to household income controlled by women, either women.4 The Food and Agriculture Organiza- through their own earnings or cash transfers, tion (FAO) estimates that equalizing access to changes spending in ways that benefit chil- productive resources between female and male dren.10 In Ghana, the share of assets and the farmers could increase agricultural output in share of land owned by women are positively developing countries by as much as 2.5 to 4 per- associated with higher food expenditures.11 In cent.5 Eliminating barriers that prevent women Brazil, women’s own nonlabor income has a from working in certain occupations or sectors positive impact on the height of their daugh- would have similar positive effects, reducing ters.12 In China, increasing adult female income the productivity gap between male and female by 10 percent of the average household income workers by one-third to one-half (chapter 5) and increased the fraction of surviving girls by 1 per- increasing output per worker by 3 to 25 percent centage point and increased years of schooling across a range of countries.6 But achieving these for both boys and girls. In contrast, a compa- gains will not occur automatically as countries rable increase in male income reduced survival get richer: multiple and sometimes reinforcing rates and educational attainment for girls, with barriers to gender equality can get in the way. no impact on boys.13 In India, a woman’s higher These productivity gains are likely to be even earned income increases her children’s years of larger in a more integrated world where efficiency schooling.14 in the use of resources is essential to a country’s Improvements in women’s own education competitiveness and growth. Indeed, recent and health also have positive impacts on these work shows that gender inequality has become and other outcomes for their children. Better more costly for most countries in a world of open nutritional status of mothers has been associ- trade.7 Gender inequality diminishes a country’s ated with better child health and survival.15 ability to compete internationally—particularly And women’s education has been positively if the country specializes in exporting goods and linked to a range of health benefits for chil- services for which men and women workers are dren—from higher immunization rates to bet- equally well suited. Industries that rely more on ter nutrition to lower child mortality. Mothers’ female labor expand more in countries where (and fathers’) schooling has been positively women are more equal.8 The relationship also linked to children’s educational attainment goes the other way: countries with an advantage across a broad set of countries; in Pakistan, in making products that rely more on women’s children whose mothers have even a single year labor also have become more gender equal.9 of education spend one extra hour studying at And in countries and regions with rapidly aging home every day and report higher test scores.16 populations, like China and Europe and Central Women’s lack of agency—as seen in domes- Asia, encouraging women to enter and remain tic violence—has consequences for their chil- in the labor force can help dampen the adverse dren’s cognitive behaviors and health as adults. impact of shrinking working-age populations. Medical research from developed countries has So, in a globalized world, countries that reduce established a link between exposure to domes- gender-based inequalities, especially in second- tic violence as a child and health problems as ary and tertiary education and in economic par- an adult—men and women who experienced ticipation, will have a clear advantage over those violence in the home as children are two to that delay action (chapter 6). three times more likely to suffer from cancer, a stroke, or cardiovascular problems, and five Women’s endowments, agency, and to ten times more likely to use alcohol or ille- opportunities shape those of the next gal drugs than those who did not.17 Numerous generation studies also document how experiencing vio- Greater control over household resources by lence between parents as a child is a risk factor women leads to more investment in children’s for women experiencing violence from their human capital, with dynamic positive effects own partners as adults, and for men perpetrat- on economic growth. Evidence from a range ing violence against their partners.18 6 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Increasing women’s individual and collective to systematically favor the interests of those with agency leads to better outcomes, institutions, more influence. Institutional constraints and and policy choices market failures that feed gender inequalities are Agency is about one’s ability to make choices— less likely to be addressed and corrected, leading and to transform them into desired actions and to their persistence. As highlighted in the World outcomes. Across all countries and cultures, Development Report 2006: Equity and Develop- there are differences between men’s and women’s ment, an “inequality trap” may thus emerge, ability to make these choices, usually to women’s preventing generations of women from getting disadvantage. These gendered differences mat- educated and taking up economic opportuni- ter for women’s well-being but also for a whole ties on a par with men, reducing their ability to set of outcomes for their families and for soci- make informed choices and to realize their po- ety in general. Women’s agency influences their tential as individuals.23 ability to build their human capital and take up economic opportunities. In Bangladesh, women with greater control over health care and house- What does this Report do? hold purchases have higher nutritional status. Women’s agency also matters for the welfare of This Report focuses on the economics of gender their children. In Mexico, the daughters (but equality and development. It uses economic the- not the sons) of women with more control ory to understand what drives differences in key over household decisions work fewer hours on aspects of welfare between men and women— household tasks. education and health, access to economic op­ Women’s collective agency can be transfor- portunities and productive resources, and the mative for society. It can shape the institutions, ability to make effective choices and take ac- markets, and social norms that limit their indi- tion. And it uses the same economic lens to vidual agency and opportunities. Empowering explore what policy interventions and broader women as political and social actors can change societal action can be taken to reduce these policy choices and make institutions more rep- gender differences and improve development resentative of a range of voices. Female suffrage outcomes generally. The Report does not limit in the United States led policy makers to turn itself to economic outcomes—indeed, it devotes their attention to child and maternal health and roughly equal attention to human endowments, helped lower infant mortality by 8 to 15 per- economic opportunities, and women’s agency, cent.19 In India, giving power to women at the signaling the importance of all three interrelated local level (through political quotas) led to in- aspects in human welfare. Nor does it ignore the creases in the provision of public goods (both central role of social and political institutions, female-preferred ones such as water and sani- whether formal or informal, in determining tation and male-preferred goods such as irri- gender outcomes. But in its framing of the issues gation and schools) and reduced corruption.20 and in the evidence it brings to the case for gen- Bribes paid by men and women in villages with der equality, it draws heavily on the economic a female leader were 2.7 to 3.2 percentage points literature on gender. less than in villages with a male leader.21 In India We adopt this approach for four reasons. First, and Nepal, giving women a bigger say in man- it provides valuable insights into how key gender aging forests significantly improved conserva- outcomes emerge and evolve as the development tion outcomes.22 Women’s greater public voice process unfolds, as well as how the role and ef- not only benefits women and children but can fectiveness of policy influence these outcomes. also benefit men. In many rich countries, greater Second, it builds on a tradition of World Bank female participation in economic activity has work on the economics of gender (most notably, combined with their increased representation the Engendering Development report24) and on in political leadership to reshape social views on the institution’s strongest areas of expertise and balancing work and family life in general and to specialization. Third, there are significant data pass more family-friendly labor legislation. and knowledge gaps that we can help fill in this Conversely, when women and men do not area. Fourth, while the Report often arrives at di- have equal chances to be socially and politically agnoses similar to those of other approaches, it active and to influence laws, politics, and policy provides different insights into the policy levers making, institutions and policies are more likely that can be used in support of gender equality. Overview 7 The Report focuses largely on inequalities affecting women, dwelling on ones likely to be How Box 3  women and men define gender in the reproduced and passed on to the next genera- 21st century tion. But it also focuses on inequalities affecting men, while recognizing that most of these male To inform this Report the World Bank conducted new field research in 19 coun- inequalities affect fewer realms of welfare. tries in all regions to gain a first-hand look at how men and women experience We adopt an empirical approach, prefer- gender in their everyday lives. ring rigorous and evidence-based analysis and Women and men from all age groups, incomes, and locations see educa- highlighting causality where feasible. For this, tion, the ownership of assets, access to economic opportunities, and opportunities to earn an income as the keys to improving their well-being and we draw on a large and growing body of quan- that of their families. In 500 focus groups, researchers identified women’s and titative gender research, complemented by new men’s roles and responsibilities in private and public spheres—with women’s analysis, particularly on time use, domestic tasks being largely associated with family care and home production, and violence, mortality risks, and inputs into agri- men’s with income generation and decision making. But differences across culture and entrepreneurship. We also draw on generations clearly show that these roles are being redefined in a world that new qualitative field research with more than offers new opportunities and demands for both men and women. 4,000 men and women in 98 communities from The findings also show that old problems persist in new settings even as new challenges are emerging. Many groups face pervasive disadvantages—for 19 developing countries, exploring how gender them, change remains an aspiration for future generations but not a reality in affects their everyday lives and their aspirations, their everyday lives. education, job choices, decision making, and other aspects of well-being (box 3).25 Source:  World Bank 2011. A global report like this one cannot provide Note:  The exercise was conducted with men and women of different age groups in 98 communities (about in-depth analysis of specific country circum- 4,000 individuals) in Latin America (Dominican Republic and Peru), Europe and Central Asia (Moldova, Po- land, and Serbia), Africa (Burkina Faso, Liberia, Sudan, South Africa, and Tanzania), South Asia (Afghanistan, stances. Nor can it cover all relevant dimensions Bhutan, and India), the Middle East (West Bank and Gaza and the Republic of Yemen), East Asia (Indonesia of gender equality. Instead, it proposes a con- and Vietnam), and the Pacific Islands (Fiji and Papua New Guinea). ceptual framework to explain gender inequal- ity and recommend public action, which can be adapted as necessary to specific countries, Drawing on past and recent work on gender issues, and sectors. It then illustrates the use of and development within the World Bank26 and this framework by focusing on aspects of gen- elsewhere, the Report posits that gender out- der equality where there has been most progress comes can be understood through the responses worldwide (education, fertility, life expectancy, of households to the functioning and structure labor force participation, and the extension of of markets and institutions, both formal and in- legal rights) and where there has been little or formal. Families decide how many children to very slow change (excess female mortality, seg- have and when, how much to spend on educa- regation in economic activity, gaps in earnings, tion and health for daughters and sons, how to responsibility for house and care work, asset allocate different tasks (inside and outside the ownership, and women’s agency in private and household), and other matters that influence public spheres). gender outcomes. “ I believe that a woman must be educated and must work in order to prove herself in society and to be a better mother. Young woman in Rafah city, West Bank and Gaza Women should work. Why should I stay at home if I can work outside? I should also earn income and my people and myself should enjoy the money I can make. Those days where our mothers were to ask for money from our fathers, even for simple things like underwear are gone: we need our own money and this means ” that we should work. Young woman in Bukoba municipality, Tanzania 8 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 (box 4). Voice and bargaining power of house- What Box 4  do we mean by markets, formal institutions, hold members are defined by a range of factors, and informal social institutions? including their ownership of and control over resources, their ability to leave the household (exit options), and social norms. In this way, Markets—a variety of arrangements that allow buyers and sellers to exchange household decision making, markets, formal (the rights over) any type of goods and services subject to a set of rules. Markets allow for any item that is exchanged to be evaluated and priced. Markets can be institutions, and informal institutions combine influenced and shaped by formal and informal institutions. and interact to determine gender-related out- comes (figure 1). Formal institutions—all aspects that pertain to the functioning of the state, The benefits of economic development (the including laws, regulatory frameworks, and mechanisms for the delivery of ser- vices that the state provides (such as judicial services, police services, basic infra- combination of higher incomes and better ser- structure, health, and education). vice delivery institutions) on gender outcomes can be seen clearly through this framework as Informal social institutions—the mechanisms, rules, and procedures that shape emerging from the workings of households, social interactions but do not pertain to the functioning of the state. In this Report, the focus is on gender roles, beliefs, social norms, and social networks. markets, and institutions and their interactions. Gender roles provide guides to normative behaviors for each sex within certain These impacts are illustrated in figure 1 by the social contexts. Roles gain power as they are learned through socialization, elab- “growth” arrow that turns the gears in the di- orated in cultural products, and enacted in daily life. The repeated experience of rection of greater gender equality. The impact of performing gender roles affects widely shared beliefs about men’s and women’s more gender equality on growth is in turn cap- attributes and one’s own sense of identity. Social norms refer to patterns of tured by the “gender equality” arrow that flows behavior that flow from socially shared beliefs and are enforced by informal social back into higher growth. sanctions. These can affect household bargaining in many ways: they set limits on what can be bargained about; they can be a determinant of or constraint to bargaining power; they can affect how bargaining is conducted; and they them- selves can be subject to bargaining and can change. Social networks refer to the Where has there been the most system of social relationships and bonds of cooperation for mutual benefit that progress in gender equality? shape one’s opportunities, information, social norms, and perceptions. For women and girls in developing countries, Sources: Agarwal 1994, 1997; Fehr, Fischbacher, and Gätcher 2002; Kabeer 1999; Sen 1990. much has changed for the better in the past quarter century. Take female life expectancy at They make these choices on the basis of the birth. It increased dramatically in developing preferences, incentives, and constraints of dif- countries (by 20 to 25 years in most regions in ferent family members, and in relation to their the past 50 years) to reach 71 years globally in relative voice and bargaining power. Preferences 2007 (compared with 67 for men), and women are shaped by gender roles, social norms, and now outlive men in every region of the world. social networks (which we group under the label The changes were much faster than when to- informal institutions). Incentives are largely in- day’s rich countries were poorer. It took more fluenced by markets (including the markets for than 100 years for the number of children born labor, credit, land, and goods), which determine to a woman in the United States to decline from the returns to household decisions and invest- 6 to 3; the same decline took just over 35 years ments. Constraints arise from the interplay of in India and less than 20 in Iran (figure 2). The formal institutions (comprising all that pertain same patterns can be seen in primary education. to the functioning of the state) and markets but It took the United States 40 years (from 1870 also reflect the influence of informal institutions until 1910) to increase enrollments among girls aged 6 to 12 years from 57 percent to 88 percent; Morocco achieved a similar increase for this age “ I think women should go out as well to look for a job because men are failing to get jobs; for women it is easier because they have more group in just over a decade (from 58 percent in 1997 to 88 percent in 2008). Girls’ education ” options. Progress in closing gender gaps in education has Young man in rural Ngonyameni, been steady and sustained at all levels—primary, South Africa secondary, and tertiary. In many countries, and especially for higher education, these gaps are Overview 9 Figure 1 Gender outcomes result from interactions between households, markets, and institutions R EQUALIT NDE Y po GE lic ies S UTION MAL INSTIT INFOR ECONOMIC OPPORTUNITIES TS HOUSEHOLDS RKE MA AGENCY ENDOWMENTS FOR MAL INST ITUT ION S GR OW TH Source: WDR 2012 team. now reversing, with boys and young men at a Across the world, women F i g u r e 2   relative disadvantage. Two-thirds of all coun- are having fewer children tries have reached gender parity in primary education enrollments, while in over one-third, How fast can fertility decline? girls significantly outnumber boys in secondary Iran, education (figure 3). Even in regions with the Islamic Rep. largest remaining gender gaps—South Asia and Bangladesh Sub-Saharan Africa (particularly West Africa)— there have been considerable gains. And in a Morocco striking reversal of historical patterns, more Zimbabwe women than men now attend universities, with women’s tertiary enrollment across the globe Colombia having risen more than sevenfold since 1970 India (fourfold for men). Yet while boy disadvantage United is slowly emerging in some places, girl disadvan- States tage where it exists tends to emerge earlier in life 0 20 40 60 80 100 and is deeper. number of years for the total fertility rate to fall from more than 6 children Women’s market work to less than 3 Women’s labor force participation has grown in the past 30 years as expanding economic op- Source: www.gapminder.org portunities have drawn many female workers 10 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2  ender parity in enrollments at lower levels has been achieved in much of Figure 3 G the world, but tertiary enrollments are very low and favor women Primary education Secondary education Tertiary education 100 100 100 men Some African Enrollment gaps dwarf disadvantaged school enrollment, women, gross, % countries lag behind gender gaps school enrollment, girls, net, % school enrollment, girls, net, % 80 80 80 60 60 60 40 40 40 Women are more likely 20 20 20 to participate girls girls than men disadvantaged disadvantaged 0 0 0 0 20 40 60 80 100 0 20 40 60 80 100 0 20 40 60 80 100 school enrollment, boys, net % school enrollment, boys, net % school enrollment, men, gross % East Asia and Paci c Europe and Central Asia Latin America and the Caribbean Middle East and North Africa South Asia Sub-Saharan Africa High-income countries Source: WDR 2012 team estimates based on World Development Indicators. Note: The 45° line in each figure above shows gender parity in enrollments. Any point above the 45° line implies that more women are enrolled relative to men. into the market. Between 1980 and 2008, the tures change so that activities in which men gender gap in participation narrowed from 32 no longer have an advantage become more percentage points to 26 percentage points. By prominent. This shift opens new opportunities 2008, women represented more than 40 percent for women’s employment, and households re- of the global labor force. Large increases in par- spond to these signals by educating daughters. ticipation in countries that started with very low Richer countries can also invest in more ac- rates (mainly in Latin America and the Carib- cessible education systems by building schools bean and to a lesser extent in the Middle East and hiring teachers. When combined with bet- and North Africa) combined with small declines ter incentive and accountability systems, these in countries that started with very high rates inputs help deliver better and cheaper services, (mainly in Eastern Europe and Central Asia) lowering the costs of access to households and mean that rates have converged across regions, increasing their use. Where all these factors although significant differences remain. Female have worked together, the gaps have closed rap- labor force participation is lowest in the Mid- idly, as in Morocco. dle East and Northern Africa (26 percent) and But even if bottlenecks appear in any one South Asia (35 percent) and highest in East Asia of these channels—pro-boy preferences within and the Pacific (64 percent) and Sub-Saharan households or inadequacies in the provision of Africa (61 percent). education or slow growth or limits on women’s employment opportunities—the other channels What explains progress? still have allowed progress in educating girls. Pol- Where gaps have closed quickly, it has been a icies targeted to getting children to school, such result of how markets and institutions have as the conditional cash transfers used in more functioned and evolved, how growth has played than 30 countries worldwide (many explicitly out, and how all these factors have interacted targeting girls, as in Bangladesh and Cambodia), through household decisions. For education, have also helped. These forces are illustrated in Primary Ed Secondary consider each Ed incomes in turn. Higher allow Edfigure 4 by the (green) gears representing house- Tertiary families that had previously only sent their sons holds, formal institutions, and markets all mov- to school to now send their daughters as well. ing in ways that narrow educational gender gaps As countries get richer, their economic struc- (“oiled” by supportive policies). Overview 11 Figure 4 Using the framework to explain progress in education R EQUALIT ND E Y po GE lic ies NS ITUTIO MA L INST INFOR ECONOMIC OPPORTUNITIES TS ns RKE r HOUSEHOLDS MA g retu a s i n ion More stable e t Incr educa incomes AGENCY ENDOWMENTS to FOR MA Low L INSTI er d TUT and irect, ION opp indi S ortu rect, cost nity s GR OW TH Source: WDR 2012 team. The interactions between households, mar- has increased sharply over time. So, at every level kets and institutions can also explain the pattern of per capita income, more women are now en- and pace of female labor force participation. A gaged in economic activity outside the home woman’s decision to work outside the home re- than ever before. sponds both to changes in her own wages and There are two main reasons why gains in to changes in her household income. As low- some domains of gender equality in many de- income countries grow richer, women partici- veloping countries came faster than they did for pate less in market work because their household today’s rich countries when they were at com- incomes also rise. Over time, their education lev- parable incomes. First, the incomes of many els also increase as formal institutions respond. developing countries have grown faster. Since Rising incomes also lead to later marriage and 1950, 13 developing countries have grown at childbearing and lower fertility. These factors an average of 7 percent a year for more than 25 all bring women back into the labor force. In 10 years or longer—a pace unprecedented before Latin American countries, almost two-thirds of the latter half of the 20th century.28 Second, the the increase in women’s labor force participa- various domains of gender outcomes are inter- tion in the past two decades can be attributed related. So, improvements in one have spurred to more education and to changes in family advances in others. The decline in fertility that formation (later marriage and lower fertility).27 has come with higher incomes has helped lower These different impacts of income growth and the number of deaths associated with maternal rising women’s wages lead to a U-shaped pat- mortality. And bearing fewer children has given tern of female labor force participation across women more time to invest in acquiring hu- countries (figure 5). But notably, since 1980, the man capital and to participate in the economy. female participation rate at each level of income Forward-looking parents have responded to the 12 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 from educating girls and from women working Figure 5  Female labor force participation has increased over become evident. Consider the notable advances time at all income levels in gender equality in two very different coun- tries: Bangladesh and Colombia. • In the four decades since Bangladesh gained female labor force participation rate, % 80 independence, the average number of chil- 70 2008 dren a woman will have during her lifetime fell from almost 7 to just over 2. School en- rollment among girls rose from a third in 60 1991 to 56 percent in 2005. And just in the latter part of the 1990s, labor force participa- 50 tion for young women more than doubled. 1980 • In Colombia, the average number of children 40 a woman will bear dropped from 3.2 to 2.4 4 6 8 10 12 between the mid-1980s and 2005. Women log, GDP per capita (constant 2000 US$) also reversed the education gap and now have higher completion rates than men for primary, secondary, and even tertiary educa- Source: WDR 2012 team calculations based on International Labour Organization 2010 (130 countries). tion. And the country has the steepest in- crease in women’s labor force participation in the region, giving it one of the highest par- expanded employment opportunities by in- ticipation rates in Latin America. Women creasing their educational investments in their there are well represented in managerial po- daughters. These better educated girls are more sitions and in finance—the glass ceilings no- likely to work when they become older, have toriously hard to break through even in many fewer children, and exercise more voice in their rich countries. households—feeding the cycle of change. So, the progress in fertility, the gains in education, The problem of severely disadvantaged the gains in women’s agency, and the shifts of populations women to market work are not only related but The combined forces of markets, service deliv- also mutually reinforcing. Public policies have ery institutions, and income growth that have themselves played a role, because the big push contributed to closing gender gaps in education, for universal education of the past decade has fertility, and labor force participation for many helped get all children to school. women have not worked for everyone. For poor The main lesson: when market signals, for- women and for women in very poor places, mal institutions, and income growth all come sizable gender gaps remain. And these gaps are together to support investments in women, gen- even worse where poverty combines with other der equality can and does improve very quickly. factors of exclusion—such as ethnicity, caste, And these improvements can occur even when remoteness, race, disability, or sexual orienta- informal institutions, such as social norms tion. Even in education, where gaps have nar- about what is “appropriate” for girls and boys rowed in most countries, girls’ enrollment in or women and men, may themselves take time primary and secondary school has improved to adapt. This is not to say that social norms little in many Sub-Saharan countries and have not been important in determining these some parts of South Asia. School enrollments outcomes. The differences across countries and for girls in Mali are comparable to those in among regions within countries both in clos- the United States in 1810, and the situation ing gender gaps in educational attainment and in Ethiopia and Pakistan is not much better levels of women’s labor force participation high- (figure 6). And in many countries, gender dis- light their influence. But the fast pace of change parities remain large only for those who are in education and even in labor force participa- poor. In both India and Pakistan, while boys tion almost everywhere shows how these norms and girls from the top income quantile (fifth) adapt quite quickly as the economic returns participate in school at similar rates, there is a Overview 13 gender gap of almost five years in the bottom income quantile (figure 7). Figure 6 Low-income countries lag behind in realizing  Beyond the poor, gender gaps remain par- progress in female school enrollment ticularly large for groups for whom ethnicity, geographical distance, and other factors (such 100 Egypt, Arab Rep. as disability or sexual orientation) compound 90 gender inequality. Almost two-thirds of out-of- Bangladesh school girls globally belong to ethnic minority female school enrollment, ages 5–19, (%) 80 groups in their own countries.29 The illiteracy Mozambique Tajikistan rate among indigenous women in Guatemala 70 stands at 60 percent, 20 points above indige- Nigeria Niger Côte d’Ivoire nous men and twice the rate of nonindigenous 60 women.30 Ethiopia 50 For these severely disadvantaged groups— Burkina Faso Mali which can be pockets of disadvantage or en- 40 tire swaths of countries or regions—none of Pakistan the forces that favor educating girls and young 30 women are working. So, the growth in aggregate United States income may not be broad-based enough to ben- 20 in 1900 efit poor households. Market signals are muted 10 because economic opportunities for women do not expand much or because other barriers— 0 such as exclusion caused by ethnicity, race, or 17 0 17 0 18 0 18 0 18 0 18 0 18 0 18 0 18 0 18 0 18 0 19 0 19 0 19 0 19 0 19 0 19 0 19 0 19 0 19 0 19 0 20 0 20 0 05 7 8 9 0 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 17 caste—get in the way of accessing those oppor- year tunities. And service delivery is often riddled United States with problems because poverty, distance, and trend line (United States, 1850–2000) discrimination mean that these groups do not Source: WDR 2012 team estimates based on U.S. Census and the International Income Distribution see an expansion of schools and teachers. This Database (I2D2). does not mean that the channels that have fa- Note: Values between 1760 and 1840 are based on female school enrollment trending between 1850 vored girls’ education elsewhere will not work and 2000. for these groups. It means that efforts need to be redoubled to ensure that the essential build- ity for house and care work, gaps in asset owner- ing blocks for progress (broad-based income ship, and constraints to women’s agency in both growth, expanding employment opportunities the private and public spheres. Progress in these for women, and effective service delivery) are domains is difficult to see, despite greater pros- in place. And these efforts may need to be com- perity in many parts of the world. Indeed, many bined with complementary interventions that of these gender disparities remain salient even address specific disadvantages that compound among the richest countries. gender inequality (chapter 7). Gender disparities persist in these “sticky” domains for three main reasons. First, there may only be a single institutional or policy Where have gender inequalities “fix,” which can be difficult and easily blocked. persisted and why? We illustrate this problem with excess female mortality. Second, disparities persist when mul- By contrast to areas that have seen good prog- tiple reinforcing constraints combine to block ress, change has come slowly or not at all for progress. We use disparities in the economic many women and girls in many other dimen- sphere (the persistence of gender earnings gaps sions of gender equality. Health disadvantages and gender segregation in employment) and in that show up in the excess relative mortality of agency (differences in societal voice and house- girls and women fall into this category. So do hold decision making) to illustrate this prob- other persistent gender disparities, including lem. Third, gender differences are particularly segregation in economic activity, gender gaps in persistent when rooted in deeply entrenched earnings, male-female differences in responsibil- gender roles and social norms—such as those 14 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Figure 7 Female disadvantage within countries is more marked at low incomes Benin Congo, Dem. Rep. The Gambia 10 10 10 median grade attained, median grade attained, median grade attained, ages 15–19 ages 15–19 ages 15–19 5 5 5 0 0 0 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 income quintile income quintile income quintile India Pakistan Togo 10 10 10 median grade attained, median grade attained, median grade attained, ages 15–19 ages 15–19 ages 15–19 5 5 5 0 0 0 1 2 3 4 5 1 2 3 4 5 1 2 3 4 5 income quintile income quintile income quintile girls boys Source: WDR 2012 team estimates based on EdAttain. about who is responsible for care and house- below the age of 60. About two-fifths of them work in the home, and what is “acceptable” for are never born, one-fifth goes missing in infancy women and men to study, do, and aspire to. and childhood, and the remaining two-fifths do And these gaps tend to be reproduced across so between the ages of 15 and 59 (table 1). generations. Consider each in turn. Growth does not make the problem disap- pear. Between 1990 and 2008, the number of Higher mortality of girls and women missing girls at birth and excess female mortal- The rate at which girls and women die relative to ity after birth did not change much; declines in men is higher in low- and middle-income coun- infancy and childhood were offset by dramatic tries than in high-income countries. To quantify increases in Sub-Saharan Africa in the repro- this excess female mortality (“missing” girls and ductive ages. Part of the increase is because women) and identify the ages at which it occurs, populations increased. But, unlike Asia, where this Report estimated the number of excess fe- the population-adjusted missing women fell male deaths at every age and for every country in every country (dramatically in Bangladesh, in 1990, 2000, and 2008.31 Excess female deaths Indonesia, and Vietnam), most Sub-Saharan in a given year represent women who would not countries saw little change in the new millen- have died in the previous year if they had lived nium. And in the countries hardest hit by the in a high-income country, after accounting for HIV/AIDS epidemic, things got much worse. the overall health environment of the country The Report’s analysis helps explain these they live in. Globally, excess female mortality patterns. Depending on the period in the life after birth and “missing” girls at birth account cycle, girls and women are missing for dif- every year for an estimated 3.9 million women ferent reasons. Missing girls at birth reflect Overview 15 ta b l e 1 Almost 4 million missing women each year Excess female deaths in the world, by age and region, 1990 and 2008 (thousands) Total women girls at birth girls under 5 girls 5–14 women 15–49 women 50–59 under 60 1990 2008 1990 2008 1990 2008 1990 2008 1990 2008 1990 2008 China 890 1,092 259 71 21 5 208 56 92 30 1,470 1,254 India 265 257 428 251 94 45 388 228 81 75 1,255 856 Sub-Saharan Africa 42 53 183 203 61 77 302 751 50 99 639 1,182 High HIV-prevalence countries 0 0 6 39 5 18 38 328 4 31 53 416 Low HIV-prevalence countries 42 53 177 163 57 59 264 423 46 68 586 766 South Asia (excluding India) 0 1 99 72 32 20 176 161 37 51 346 305 East Asia and Pacific (excluding China) 3 4 14 7 14 9 137 113 48 46 216 179 Middle East and North Africa 5 6 13 7 4 1 43 24 15 15 80 52 Europe and Central Asia 7 14 3 1 0 0 12 4 4 3 27 23 Latin America and the Caribbean 0 0 11 5 3 1 20 10 17 17 51 33 Total 1,212 1,427 1,010 617 230 158 1,286 1,347 343 334 4,082 3,882 Source: WDR 2012 team estimates based on data from the World Health Organization 2010 and United Nations Department of Economic and Social Affairs 2009. Note: Totals do not necessarily add up due to rounding. overt discrimination in the household, result- South Asia. High maternal mortality is the main ing from the combination of strong preferences contributor to excess female mortality in the re- for sons combined with declining fertility and productive years. In Afghanistan, Chad, Guinea- the spread of technologies that allow parents to Bissau, Liberia, Mali, Niger, Sierra Leone, and know the sex before birth.32 This is a particular Somalia, at least 1 of every 25 women will die issue in China and North India (although now from complications of childbirth or pregnancy. spreading to other parts of India), but it is also And a much larger fraction will suffer long-term visible in parts of the Caucasus and the Western health consequences from giving birth.33 Balkans. Progress in reducing maternal mortality has Missing girls during infancy and early child- not been commensurate with income growth. hood cannot be explained by a preference for In India, despite stellar economic growth in re- sons alone, although discrimination against girls cent years, maternal mortality is almost six times may contribute to it. It is a result not so much of the rate in Sri Lanka. In the past two decades, discrimination as of poor institutions that force only 90 countries experienced a decline of 40 households to choose among many bad options, percent or more in the maternal mortality ra- particularly regarding water and sanitation. tio, while 23 countries showed an increase. The Markets and households cannot compensate for main problem is, again, that households are be- these poor services. ing asked to make many decisions in the face of Missing women in the reproductive ages re- bad options—a result of multiple service deliv- flect two main factors. First, stubbornly high ery failures. In many parts of the world, this sit- rates of maternal mortality persist, especially in uation is reinforced by social norms that influ- much of Sub-Saharan Africa and some parts of ence household behavior and make it difficult 16 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 for women to get maternal health care quickly in the quality of institutions—in the provi- enough even where it is available. And high fer- sion of clean water, sanitation, and maternal tility, partly reflecting low incomes, compounds health care. Because there is only a single point the problem in parts of Sub-Saharan Africa. of entry—through better institutions—for ad- Second, the impacts of the HIV/AIDS pan- dressing female mortality, solving the problem is demic on the mortality of women in many East- hard—much harder than getting girls to school. ern and Southern African countries have been But for any basic notions of human justice, the dramatic. The reason for the greater prevalence global development community must make ad- of HIV/AIDS among women relative to men is dressing this problem a priority. their greater susceptibility and the greater like- lihood that their sexual partners are older and Gender segregation in economic activity thus more likely than younger men to have HIV. and earnings gaps In addition, countries that have had a low-lying Although women have entered the labor force civil conflict (such as Democratic Republic of in large numbers across much of the developing Congo) have also seen an increase in the num- world in the past quarter century, this increased ber of “missing” women. This is in contrast to participation has not translated into equal em- other countries that have had outright wars— ployment opportunities or equal earnings for like Eritrea, where men who went “missing” in men and women. Women and men tend to the years of war increased. work in very different parts of the “economic An examination of the historical experience space,” with little change over time, even in of northern and western European countries high-income countries. In almost all countries, and the United States reveals that similar pat- women are more likely than men to engage in terns of excess female mortality in infancy and low-productivity activities. They are also more the reproductive years existed there but disap- likely to be in wage or unpaid family employ- peared between 1900 and 1950. These reductions ment or work in the informal wage sector. In occurred primarily because of improvements agriculture, especially in Africa, women operate smaller plots of land and farm less remunerative crops. As entrepreneurs, they tend to manage smaller firms and concentrate in less-profitable Figure 8  Women and men work in different sectors sectors. And in formal employment, they con- centrate in “female” occupations and sectors distribution of female / male employment across sectors (figure 8). These patterns of gender segregation 31% Communication Services 16% in economic activity change with economic de- velopment but do not disappear. 21% Retail, Hotels, and Restaurants 17% As a result of these differences in where 13% Manufacturing 12% women and men work, gender gaps in earn- ings and productivity persist across all forms of 4% Finance and Business 4% economic activity—in agriculture, in wage em- ployment, and in entrepreneurship (map 1). In 0.5% Electricity, Gas and Steam, and Water 1% almost all countries, women in manufacturing 0.5% Mining 2% earn less than men. In agriculture, farms oper- ated by women on average have lower yields 2% Transport and Telecommunications 7% than those operated by men, even for men and women in the same households and for men 27% Agriculture, Hunting, etc. 29% and women cultivating the same crops.34 Fe- 1% Construction 11% male entrepreneurs are also less productive than male entrepreneurs.35 In urban areas in Eastern 100% All Sectors / All Occupations 100% Europe and Central Asia, Latin America, and Sub-Saharan Africa, the value added per worker is lower in firms managed by women than in Source: WDR 2012 team estimates based on International Labour Organization 2010 (77 countries). those managed by men.36 For firms operating in Note: Totals do not necessarily add due to rounding. rural Bangladesh, Ethiopia, Indonesia, and Sri Overview 17 Map 1 Earnings gaps between women and men (female earnings relative to $1 of male earnings) Germany 62¢ Iceland 69¢ Georgia 60¢ India 64¢ A.R. of Egypt 82¢ Bangladesh 12¢ Mexico 80¢ Benin 80¢ Nigeria 60¢ Ethiopia 34¢ Sri Lanka 50¢ Malawi 90¢ Salaried Workers Farmers Entrepreneurs Sources: Data for Benin come from Kinkingninhoun-Mêdagbé and others 2010; for Malawi from Gilbert, Sakala, and Benson 2002; for Nigeria from Oladeebo and Fajuyigbe 2007; for Bangladesh, Ethiopia, and Sri Lanka from Costa and Rijkers 2011; and for Egypt, Georgia, Germany, Iceland, India, and Mexico from LABORSTA, International Labour Organization. Lanka, the differences in profitability are signifi- disproportionate responsibility for housework cant between female-owned and male-owned and care, while men are responsible mostly for businesses.37 market work (figure 10). When all activities are So, what explains this persistent gender seg- added up, women typically work more hours regation in economic activity and the resulting than men, with consequences for their leisure gaps in earnings? The Report argues that gender and well-being. And everywhere they devote differences in time use, in access to assets and more time each day to care and housework credit, and in treatment by markets and formal than their male partners: differences range institutions (including the legal and regula- from one to three hours more for housework, tory framework) all play a role in constraining two to ten times the time for care (of children, women’s opportunities. These constraints are elderly, and the sick), and one to four hours shown in figure 9 as wedges blocking progress less for market activities. Even as women take toward greater gender equality. Income growth up a bigger share of market work, they remain has some influence in shifting these patterns but largely responsible for care and housework. does not eliminate them. The mutually reinforc- And these patterns are only accentuated after ing interactions between these different factors marriage and childbearing. make the problem particularly difficult to break. A second factor driving segregation in em- Consider each in turn. ployment and earnings gaps is differences in The differing amounts of time that men and human and physical endowments (including women allocate to care and related household access to assets and credit). Despite increases in work are one factor driving segregation and women’s education, there are still differences in the consequent earnings gaps. In most coun- human capital between women and men. These tries, irrespective of income, women bear a include a gap in years of schooling among older 18 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Figure 9 Explaining persistent segregation and earnings gaps NS ITUTIO R M A L INST INFO on norms Social rket work a care/m ECONOMIC TS OPPORTUNITIES RKE s to MA cces HOUSEHOLDS ntia a /land l ere credit d Di erential Di r/ an allocation of labo rkets, s ma twork time/resources ne AGENCY ENDOWMENTS FOR MAL Bias INST ed l ITUT aws ION and /regula S l i t GRO infr m i t i ons astr e d , uctu re W H T Source: WDR 2012 team. cohorts as well as differences in what women large. Data for 16 countries in five developing and men choose to study in younger cohorts— regions indicate female-headed households are differences that affect employment segregation, less likely to own and less likely to farm land.38 especially in countries where most young people More generally, where evidence is available go to college. In agriculture and entrepreneur- for all farmers, women seldom own the land ship, large and significant gender disparities in they farm. For example, in Brazil, women own access to inputs (including land and credit) and as little as 11 percent of land. And their land- in asset ownership are at the root of the gender holdings are systematically smaller than those productivity gap. Indeed, yield differences for owned by men. In Kenya, women account for 5 female and male farmers disappear altogether percent of registered landholders nationally.39 when access to productive inputs is taken into And in Ghana, the mean value of men’s land- account (figure 11). Differences in access to in- holdings is three times that of women’s land- puts may be further compounded by differences holdings.40 Similarly large gaps are observed in in the availability of “market time,” as noted use of fertilizers and improved seed varieties in above, which can make the same investment agriculture, and in access to and use of credit less productive for women than for men. Jointly, among entrepreneurs. these constraints mean that women entrepre- Third, market failures and institutional con- neurs and farmers are often restricted to busi- straints also play a role. Labor markets often do nesses and activities that are less profitable and not work well for women, especially if their pres- less likely to expand. ence is limited in some sectors or occupations. How big are gender differences in access to When few women are employed, employers may assets (especially land), credit, and other in- hold discriminatory beliefs about women’s pro- puts? A variety of data sources suggests they are ductivity or suitability as workers—these beliefs Overview 19  cross the world, women spend more hours per day on care and Figure 10 A housework than men FOR SALE Market activities Housework Child care Pakistan 0.6 4.7 5.5 2.5 1.2 0.2 Cambodia 2.7 3.8 4.4 3.3 0.9 0.1 South Africa 2.1 3.8 4.2 1.8 0.5 0.0 Bulgaria 2.9 3.9 4.7 2.6 0.4 0.1 Sweden 3.2 4.6 3.2 2.3 0.6 0.3 Italy 2.1 4.8 4.9 1.4 0.6 0.2 women = 12 hours men Source: Berniell and Sanchez-Páramo 2011. can persist if there are no mechanisms in place measures, prevent women from entering some to correct them. Access to information about sectors or occupations. jobs, and support for promotions and advance- In sum, whether women are farmers, en- ment, often occur in gendered networks, hurt- trepreneurs, or workers, many are caught in a ing women trying to enter a male-dominated productivity trap: working hard on an uneven field (or equally hurting men trying to enter a playing field with unequal access to productive female-dominated one, such as nursing). And inputs. This trap imposes significant costs on sometimes, legal barriers, framed as protective women’s welfare and economic opportunities 20 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 of married women in India are not involved in Figure 11  Gender differences in agricultural productivity spending decisions, even about their own in- disappear when access to and use of productive ­ comes. Even in an upper-middle-income coun- inputs are taken into account try like Turkey, more than a quarter of married women in the lowest income quantile lack con- In Malawi, women’s agricultural trol over their earned income.41 Women’s ability productivity is 13.5% lower than men’s to own, control, and dispose of property still dif- this difference disappears when women and men have equal access to inputs –13.5% Malawi (National) 0.6% fers from that of men—sometimes legally, often in practice. And again, these patterns change –40% Nigeria (Osun State) 0% only slowly as countries grow richer. A clear manifestation of the lack of agency –21% Benin (Central) 0% is domestic violence. Violence is the opposite –17% of freedom—an extreme form of coercion that Ghana (National) 0% by definition negates agency. Women are at far –26% greater risk of violence by an intimate partner or Ethiopia (Central Highlands) –25% someone they know than from violence by other –7.7% people. And women are more likely than men to Kenya (Subnational) 12.5% be killed, seriously injured, or victims of sexual Kenya (Western), 2008 –19% violence by intimate partners.42 The prevalence 0% of domestic violence varies greatly across coun- Kenya (Western), 1976 –4% tries, with no clear relationship to incomes; 6.6% while incidence tends to rise with socioeconomic –60 –50 –40 –30 –20 –10 0 10 20 deprivation, violence knows no boundaries. In percent some middle-income nations, such as Brazil average gender gap (Sao Paolo and Pernambuco region) and Serbia gender gap with equal access to inputs (Belgrade), women report that the incidence of physical violence by intimate partners is as high as 25 percent.43 In Peru (Cusco), almost 50 Source: Alene and others 2008; Gilbert, Sakala, and Benson 2002; Kinkingninhoun-Mêdagbé and others 2010; Moock 1976; Oladeebo and Fajuyigbe 2007; Saito, Mekonnen, and Spurling 1994; Vargas Hill and percent of women are victims of severe physical Vigneri 2009. violence during their lifetime, and in Ethiopia (Butajira), 54 percent of women reported being today—and serious disincentives to invest in the subject to physical or sexual abuse by an inti- women of tomorrow. mate partner in the past 12 months. 44 Multiple factors are at work behind these Less voice in societal and household large gaps in women’s voice. In society, low decision making representation can be self-perpetuating, with In much of the world, women have less input women unable to convey their ability to lead. than men in decision making in their house- So, in politics, voters will not be able to judge holds, in their communities, and in their soci- accurately the capacity of a woman leader. And eties. Consider women’s underrepresentation in women’s entry may be limited by societal beliefs formal politics, especially in its upper reaches. that being engaged in politics is a masculine Fewer than one-fifth of all cabinet positions is activity or that women are less effective leaders held by women. And women’s lack of represen- than men—beliefs that are hard to break until tation extends to the judiciary and labor unions. a critical mass of women rises to political lead- These patterns do not change much as countries ership. Different responsibilities for care work get richer. The share of women parliamentari- also mean that women lack the flexibility or the ans increased only from 10 percent to 17 percent time to invest as heavily as men in participating between 1995 and 2009. in political institutions. The lack of networks for Whether and how much voice a woman has women also makes it more difficult for them to in household decision making over patterns of ascend to positions of authority in political par- spending, including spending on children, are ties or labor unions. important markers of her agency. As many as a In the household, two important determi- third of married women in Malawi and a fifth nants of a woman’s voice are her income and Overview 21 her control over household assets. Economic pro-male bias in India and Ethiopia but not in growth can improve the material conditions for Peru and Vietnam. exercising agency, with women generally having A growing body of research also suggests that more voice in wealthier households. But higher attitudes about women in the family and the household incomes alone are not enough to workplace are transmitted across generations. eliminate the lower capacity of women to ex- When women do not work outside the home, ercise agency. What matters are a woman’s own their daughters are also less likely to do so as income and assets as well as her ability to leave adults, and their sons are less likely to marry the household, all of which increase her bargain- women who work outside the home.48 Young ing power and ability to influence household men and young women also tend to study in choices. In India, owning property substantially very different fields—with women favoring enhances women’s voice in the household on education and the humanities, while men fa- various matters and reduces her risk of domes- vor engineering, agriculture, and sciences—in tic violence.45 Similarly, as women’s shares in ways that are unrelated to abilities (chapter 3) household earnings increase in Colombia and yet repeat themselves over generations and do South Africa, so does their control over key not go away as incomes rise. Evidence also sug- household decisions. There is also evidence of a gests that domestic violence witnessed as a child relationship between women’s assets, earnings, is repeated in adulthood.49 Women in Haiti who and shares of household income, and the inci- had witnessed domestic violence were more dence of domestic violence.46 likely to report being the victims of physical or But if women’s earnings are limited by mal- sexual violence.50 The pattern is similar in Cam- functioning markets or other gender-differenti- bodia and Mexico. 51 And men in South Africa ated barriers to economic opportunities and as- who reported witnessing violence between their set ownership, women’s voice in the household parents were significantly more likely to report will remain muted. Reinforcing these market perpetrating physical violence themselves.52 and legal influences are social norms dictating Norms may be learned in the household, but that men, not women, make the major decisions they are often reinforced by market signals and in households. institutions, which are gender biased in many aspects. For example, gender differences in the Reproducing gender inequality across responsibility for house and care work, as just generations discussed, are rooted in gender roles but strength- Perhaps the “stickiest” aspect of gender out- ened by discrimination in labor markets and by comes is the way patterns of gender inequal- a lack of child-care services. At the root of gen- ity are reproduced over time. Part of this dered patterns of what men and women study is persistence is rooted in slow-moving social a combination of factors that feed into household norms and how they affect what happens in decisions (norms about what is appropriate for the household. Women and men internalize girls and boys), institutions (gendered education social norms and expectations in ways that af- systems), and markets (gendered networks and fect not only their own aspirations, behaviors, occupational segregation). For domestic violence, and preferences but also those of their chil- empirical work finds significant explanatory dren. The Young Lives study looked at educa- power at the individual, household, and commu- tional aspirations and noncognitive skills of nity levels, reinforced by social perceptions and boys and girls at ages 8, 12, and 15 for 12,000 institutional failures (including a lack of protec- children in Ethiopia, Andhra Pradesh in India, tive laws and services or their poor enforcement Peru, and Vietnam.47 Parental aspirations for and delivery).53 the education of their children were biased to- ward boys in Ethiopia and India by the age of What can we learn from the persistence 12 and toward girls in Peru and Vietnam. By of all these gender gaps? the age of 15, these biases had been transmit- Markets and institutions (formal and informal) ted to children, with clearly higher educational can work against greater gender equality— aspirations among boys in Ethiopia and India, in ways that are often mutually reinforcing. and among girls in Vietnam. And by age 15, Sometimes service delivery institutions fail, as measures of agency or efficacy showed a strong for young girls and women during childbirth. 22 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Other times markets do not work well, with other domains of gender equality, such as occu- results that are worse for women, as illustrated pational segregation and many manifestations by evidence of discrimination in both labor and of women’s agency, income growth and better credit markets. Often reinforcing these mar- service delivery are far less effective in unlocking ket failures, however, are formal institutions the often multiple and reinforcing constraints that treat women and men differently. Laws that underlie persistent gender gaps. and regulations can constrain women’s agency The new forces of globalization can reduce and opportunities more than those of men, as many of these gaps. First, trade openness and when women and men have different owner- the diffusion of new information and com- ship rights, or when restrictions are placed on munication technologies have translated into hours and sectors of work for women but not more jobs and stronger connections to markets for men. Where credit and labor markets already for women, increasing their access to economic discriminate, such unequal laws and regulations opportunities and contributing to their eco- can accentuate the problem. Unequal treatment nomic empowerment. Second, urbanization may also manifest itself more indirectly through and greater access to information have allowed biased service delivery, as is the case for agricul- many in developing countries to learn about life ture extension services. Here, institutional bias and mores in other parts of the world, including and market structure (with women underrepre- the role of women, possibly affecting attitudes sented in nonfood crops that are often the target and behaviors. Third, the incentives for public of extension services) reinforce and even deepen action for gender equality are stronger than ever inequalities. because the rising global consensus on the in- All institutions (formal and informal) have trinsic importance of women’s economic, social, considerable inertia. They tend to reflect the in- and political empowerment means that gen- terests of those who wield more power and in- der inequality hurts a country’s international fluence, and they are difficult to change without standing. But this potential of globalization some form of collective agency or voice.54 Social will not be realized without effective domestic norms can be especially slow to change: norms public action to close remaining gender gaps in that may have served a purpose at one point in endowments, agency, and access to economic time, but are no longer useful, may endure sim- opportunities. ply because of custom or because a social pen- So, what should governments in developing countries do to foster greater gender equality? alty is associated with being the first to break the What areas of gender inequality should they norm, or because the norm benefits a dominant focus on? Should they start with interventions group in society (in this case, men). The norm’s in education and health, or should they focus persistence can perpetuate gender inequalities on access to economic opportunities or agency? long after its original rationale has disappeared. What combination of policies should they im- In sum, gender-differentiated market fail- plement, and in what sequence? At first blush, ures, institutional constraints, and persistent these questions can appear overwhelming be- social norms often combine to reinforce gender cause of the multiplicity of priority areas and inequalities and make improving gender equal- the number of available policy instruments. ity much more complex. When there are multi­ This Report shows how better analysis can help ple constraints, they all need to be addressed. reduce the complexity of policy choice and de- sign in several ways. The starting point is to determine which as- What is to be done? pects of gender inequality should be of highest priority for policy going forward. Three criteria Nothing is automatic about the growth and matter in this regard: development process that delivers greater gen- der equality on all fronts. Part of the reason is • First, which gender gaps are most significant that higher incomes and better delivery of ser- for enhancing welfare and sustaining devel- vices by the state help reduce gender gaps only opment? So, where are the likely payoffs for in some domains. And even in these domains, development from addressing gender dis- the improvements do not reach all women. In parities likely to be the largest? Overview 23 • Second, which of these gaps persist even as Policies to reduce gender gaps in countries get richer? So, where do higher human capital endowments incomes by themselves do little to reduce ­ (health and education) ­ disparities? Addressing gender gaps in human capital en- • Third, for which of these priority areas has dowments—excess female mortality at specific there been insufficient or misplaced atten- periods of the life cycle and pockets of gender tion? So, where would a reorientation of pol- disadvantage in education—requires fixing the icies yield the greatest benefits? institutions that deliver public services. Providing basic services in a timely manner to expectant Applying these criteria, we conclude that four mothers and improving the availability of clean areas should be of the highest priority for policy water and sanitation to households will go a long makers: way to closing the gender gaps in excess mortal- • Reducing gender gaps in human capital en- ity. Education services need to focus on improv- dowments (addressing excess female mortal- ing access for the significant population groups ity and eliminating pockets of gender disad- that are currently disadvantaged by poverty, eth- vantage in education where they persist) nicity, caste, race, or geography. Such a focus will help address the “gender inequality traps” that • Closing earnings and productivity gaps be- affect the poor and excluded in society. tween women and men These solutions can come from either the • Shrinking gender differences in voice demand or the supply side, but they cannot be gender blind. On the contrary, they must factor • Limiting the reproduction of gender inequality in explicitly, both for design and implementa- over time, whether it is through endowments, tion, the drivers of gender inequality that cause economic opportunities, or agency the gender gaps in health and education out- Obviously, not all these priorities apply to all comes to persist. And they must bring into the countries. And specific country characteristics process of policy design and implementation will dictate how corrective policies will need to the voices of those that the policy is trying to be customized. reach—excluded women and girls, and the men Our analysis also emphasizes that, in choos- and boys who live with them. ing and designing policies, it is necessary to tar- get the determinants of the gender gaps of con- Reducing excess female mortality cern, not the outcomes. The framework in Part The main determinants of excess female mor- 2 of the Report helps highlight these underlying tality in different periods of the life cycle have causes, and shows how they emerge from the little to do with how quickly countries grow. workings of markets and institutions and their They result from household preferences and interactions with each other and with house- from failures in the ways that markets and in- holds. In other words, the framework identifies stitutions function. The entry points for policy what problem needs to be solved and whether are dictated by which of these influences is most interventions should target markets, formal in- binding in each period. stitutions, informal institutions, or some com- Skewed sex ratios at birth is a problem in a bination of the three. few parts of the world, including China, parts of Having identified the underlying causes of India, and parts of the Caucasus and the Western the gender gap of concern, the Report draws on Balkans. The underlying cause is son preference the experience with policy interventions across among households, which has been exacerbated a broad range of countries to provide guid- in some of these places by rapid income growth. ance on specific interventions that could work Higher incomes have increased access to ultra- in different settings. It also looks at the political sound technologies that assist in sex selection at economy of reforms and emphasizes that policy birth. So, policies need to work on two fronts. design and implementation must be attuned First, laws need to be enacted and enforced to to countries’ institutional, social, political, and deal with the abuse of sex selection technologies, as cultural environments and to the societal actors has been done in China and India. But experi- involved. ence shows that enforcement is difficult, if not 24 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 impossible, to achieve without imposing dra- efficiently while ensuring that services remain conian restrictions that are not feasible in most accessible and affordable for poor people. societies and that raise other ethical concerns. The solution will depend on the setting, but a And the difficulty in enforcing such restrictions few elements are critical. is exacerbated because these problems are worse • Appropriate regulations that recognize the among the wealthy in these societies. rationale for government intervention. A second, and more promising, approach is to enhance household perceptions of the value of • An adequate structure of incentives for pro- daughters. Expanding economic opportunities viders to make them more accountable to for young women, including those in the labor policy makers. market, is one way of doing this, and it can work • Measures to strengthen accountability of in conjunction with the process of develop- both providers and policy makers to service ment to reverse son preference. Just look at the users. Republic of Korea, one of the few cases where excess female mortality at birth was reversed in In urban areas, providing clean water will a short period.55 And the process can be comple- require an emphasis on improving the struc- mented by providing financial incentives to par- ture of contracts and, in some circumstances, ents to have daughters (such as the “Apni Beti greater involvement of the private sector. In Apni Dhan” program in some Indian states) and Manila, such reforms yielded large impacts: wa- supporting media campaigns to change societal ter supply coverage expanded from 67 percent ideas about gender equality. in 1997 to 99 percent in 2009 and brought ef- In infancy and early childhood, excess ficiency gains through lower water losses and mortality of girls is not rooted in households operating costs. In low-income settings, where or markets—although both can worsen it. It financing options and the capacity of public in- is rooted in the failure of institutions to pro- stitutions may be more limited, charging small vide clean water, sanitation, waste disposal, and amounts for services, relying on independent drainage. Countries with high female mortality providers, and finding ways of making provid- in infancy are those where the burden of infec- ers more accountable to users can help, even in tious diseases remains high. Today’s rich coun- smaller urban areas—this was the path followed tries eliminated their excess female mortality of by Cambodia. In rural areas, local governments young girls by improving access to clean water can improve community systems, as Uganda has and sanitation in the early part of the 20th cen- done by collecting an additional small tax and tury. And developing countries that have ex- placing it in a fund administered by the district perienced large drops in excess girl mortality council to pay for major water repairs. during the past two decades, like Bangladesh, For sanitation in urban settings, there usu- China, and Vietnam, have done the same. So, ally is enough demand for improvements so for the “missing” girls in Sub-Saharan Africa long as individuals and communities can cap- to “reappear,” countries must invest in similar ture the benefits of investing in the facilities. So, systems and provide adequate water, sanita- the solution is to strengthen property rights and tion, and waste disposal services to their whole recognize informal settlements, thus stimulating population and not just to the better-off. While demand while ensuring that communities have these services will benefit all young children, access to independent providers. In rural areas young girls will benefit more due to the reduc- and less dense urban settings, the priority for tion in infectious diseases. improving sanitation is to change behavior, raise How exactly should countries do this? If the awareness, and boost demand, through commu- experience of today’s rich countries is a guide, nity peer pressure and information campaigns part of the solution is to provide clean water at as in some communities in Cambodia, Indone- the point of use through piped delivery. Other so- sia, and Vietnam by appealing to people’s sense lutions, like water treatment at the source, are of community responsibility. less effective in reducing diarrhea morbidity Increasing the coverage of piped water and because of the potential for recontamination.56 sanitation is expensive, so significant funding— The problem is then to design an institutional likely external—will be needed in poor coun- framework that expands access to clean water tries. A recent analysis of infrastructure funding Overview 25 needs for Sub-Saharan Africa concluded that vices need special attention. One way to help is additional spending on clean water and sani- to provide poor women with cash transfers con- tation would need to be about one and a half ditional on their seeking maternal care. An ex- times current levels—over $11 billion annu- ample is India’s Janani Suraksha Yojana, where ally—to improve access significantly.57 How- such transfers increased the uptake of assisted ever, as documented in chapter 3, the return to deliveries in the presence of a skilled attendant these investments taking account of mortality by around 36 percent.60 declines is very large. Fourth, efforts to reduce maternal mortality In the reproductive years, maternal mortal- need to go beyond improving health systems ity remains especially high in Sub-Saharan Af- and services and work across sectors. The suc- rica and parts of Asia. The main reason is a fail- cesses of Malaysia and Sri Lanka in addressing ure of the institutions that deliver medical care maternal mortality early in their development and services to expectant mothers. While norms illustrate this point (box 5). Fairly small invest- that delay women from getting prompt medi- ments in infrastructure (rural roads) and in cal help during childbirth and high fertility may women’s education, combined with training be contributing factors in some places, solving maternal health providers and building hospi- the problem, as with providing clean water and tals dramatically reduced maternal mortality.61 sanitation, requires fixing the institutions that Fifth, it is essential that the political profile of deliver these services. the problem be raised. Turkey illustrates what is This fix will require providing more resources possible in this regard. Turkey’s maternal mor- to frontline service providers and ensuring that the tality rate in 2000 was 70 per 100,000 live births. entire system of maternal care works: A new government capitalized on the political First, the quality of the people in deliver- ing the chain of services needs to be upgraded. While additional health workers, especially Reducing Box 5  maternal mortality—What works? skilled birth attendants, will be a continuing Look at Malaysia and Sri Lanka need, coverage can be increased in underserved areas by drawing in community-level providers Improving the delivery of maternal care is hard, but it can be done—even at and the private sector. relatively low incomes, as Sri Lanka and Malaysia show. From more than 2,000 Second, those providing maternal health ser- per 100,000 births in the 1930s, the maternal mortality ratio in Sri Lanka fell to vices have to be more responsive to expectant about 1,000 by 1947, and then halved to less than 500 in the next three years. By 1996, it had fallen to 24. In Malaysia, it halved from 534 over the seven years from mothers. One way is to make service providers 1950 to 1957. Then, with a halving every decade or so, it came down to 19 by more accountable to them. Getting information 1997. to users—for example, on service standards, To overcome the range of institutional obstacles that hampers the effective quality of services and policies to improve workings of health systems, Sri Lanka and Malaysia adopted integrated and them—can help but needs to be combined with phased approaches. And they did this with modest total public expenditures some way for users to act on that information. on health—1.8 percent of gross domestic product, on average, since the 1950s. In Uganda, community-based monitoring im- Health programs in both countries exploited synergistic interactions of health care with basic education, water and sanitation, malaria control, and integrated proved both the quality and quantity of primary rural development—including building rural roads, which helped deal with health care services.58 Another route to account- obstetric emergencies. Financial, geographic, and cultural barriers to maternal ability is to ensure that citizens are able to hold care were addressed by ensuring a front line of competent, professional mid- their political representatives responsible for the wives widely available in rural areas, providing them with a steady supply of failures; the politicians, in turn, need to exercise drugs and equipment, linking them to back-up services, and improving com- more effective control over the service provid- munication and transportation. Simultaneously, facilities were strengthened to ers. The power of this mechanism is evident in provide obstetric care and deal with complications. Better organizational man- agement improved the supervision and accountability of providers. Area- Peru, where improving maternal health required specific mortality data were provided through monitoring systems so that extending coverage, giving the right incentives empowered communities could hold political leaders accountable, and to service providers, and having citizens’ voices national and subnational actors were forced to recognize the unacceptability loud enough to be heard by policy makers.59 of every maternal death. Finally, both countries were strongly committed to Professional attention for deliveries rose from 58 improving the status of women: women gained voting rights before or soon percent of births in 2000 to 71 percent in 2004. after national independence, and female education received special attention. Third, the financial constraints that poor women face in accessing maternal health ser- Source:  Pathmanathan and others 2003. 26 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 support that brought it to power and, in 2003, concentrated in lower productivity activities, undertook a Health Transformation Program, self-employment, and the informal sector. Even emphasizing institutional reform, client respon- in the formal wage sector, they cluster in certain siveness, and a focus on underserved areas. The occupations and industries, usually lower pay- budget allocated to primary health care and pre- ing. These differences remain even as countries vention in underserved areas rose by 58 percent, get richer. air ambulances were put in service for remote Three factors drive these patterns. First, populations, the health workforce was redis- women and men have very different responsi- tributed for better coverage in poor areas, and bilities for care and housework, and as a result conditional cash transfers encouraged pregnant very different patterns of time use, which im- mothers to use prebirth hostels and deliver in pinge directly on choices of employment and public hospitals. By 2009, the maternal mortal- economic activity. Second, women and men ity rate had fallen to 19.8.62 face differential access to productive inputs and often differential treatment by markets and in- Providing education to severely stitutions. Third, these mutually reinforcing disadvantaged populations constraints can generate a “female productivity Even as gender gaps in educational enroll- trap.” Policies thus need to target these underly- ment shrink nationally, they remain for poor ing factors. Because multiple factors may be at people and for those disadvantaged by other play more often than not, effective policy inter- circumstances—remoteness, ethnicity, caste, ventions may need to target several of them— race, or disability. To reach these groups, pol- either simultaneously or sequentially. icy makers can build on experience and evi- dence from Cambodia, Colombia, Honduras, Releasing women’s time Mexico, Nicaragua, Pakistan, and Turkey. A Gender differences in access to economic op- range of options includes remedies for both sup- portunities are driven in part by differences in ply (such as building more schools in remote ar- time use that result from deep-rooted norms for eas and recruiting local teachers) and demand the distribution of responsibility for care and (such as cash transfers conditioned on girls’ school housework. Addressing these binding norms attendance). A key to designing cost-effective and releasing women’s time means paying more interventions is the availability and the cost of attention to three types of policies: child care and collecting information about local character- parental leave policies; improvements in infra- istics and conditions. Where relatively little is structure services; and policies that reduce trans- known, less locally customized policies, such as action costs associated with accessing markets. cash transfers conditioned on sending daugh- Policies such as subsidies to or public provi- ters to school, may be more effective in reducing sion of child care can compensate women for gender disparities. The transfers have had posi- the costs they incur within the home from tive impacts on enrollments in both middle- engaging in market work. Child care can be income and lower-income settings, especially provided either directly by the state (including for increasing enrollments among groups with local governments) or through the private sec- low enrollments to start (such as pockets of tor, possibly with public subsidies and regula- severely disadvantaged populations).63 Having tion. Among developing countries, child-care gained wide political acceptance because of their policies have been used in some middle-income efficacy, such policies are being implemented in Latin American countries. Examples include more than 30 countries. publicly provided or subsidized day care such as Estancias Infantiles in Mexico, Hogares Co- Policies to improve women’s economic munitarios in Colombia, and similar programs opportunities in Argentina and Brazil. The evidence from Across the world, women and men access eco- these countries as well as from rich countries nomic opportunities—whether in wage em- (mainly in northern and western Europe) that ployment, agriculture, or in entrepreneurship— have similar schemes is that they increase the in fundamentally different ways. Women tend number of hours worked by women as well as to occupy very different parts of the economic lead them to work more in formal employment. space from men and are disproportionately In lower-income countries, child-care solutions Overview 27 are particularly needed for women employed working outside the home, making it easier to in the informal sector and rural women. In In- manage the multiple burdens of house, care, dia, the nongovernmental organization (NGO) and market work. And information and com- Mobile Creches is experimenting with differ- munication technologies can help reduce both ent models for providing child-care services for the time and mobility constraints that women women employed in the rural informal sector face in accessing markets and participating in and on public works programs. Similar efforts market work. Mobile banking programs, such have been undertaken in the Indian state of as M-PESA in Kenya, are allowing women to ­ Gujarat by the Self Employed Women’s Asso­ process small financial and banking transac- ciation, which has set up day-care centers for tions more effectively and promoting savings, the 0–6-year-old children of its members. Other which is especially beneficial to small women options for publicly provided day care are ei- entrepreneurs. In India, a program run by an ther to lengthen the school day (particularly NGO, the Foundation of Occupational Devel- at grades where attendance is only for half of a opment, organized groups of women to focus day) or to lower the age at which children enter on marketing, provided them with access to the school system. cell phones and the Internet, thus helping them Parental leave policies have been tried mainly market their products directly and increase their in rich countries—and these typically take the profit margins.64 form of maternity leave. While these policies have increased women’s labor force participa- Closing gaps in access to assets and inputs tion in these countries, their applicability in de- Female farmers and entrepreneurs have less ac- veloping countries may be more limited. First, cess to land than their male counterparts. Simi- they can be used only in the formal sector, which larly, both the demand for and use of credit are typically represents only a fraction of employ- lower among female farmers and entrepreneurs ment in emerging and low-income economies. than among their male counterparts. These dif- Second, they can actually make it less attractive ferences are rooted in failures of markets and in- for employers to hire women of child-bearing stitutions and in their interactions with house- age unless the maternity leave is publicly fi- hold responses. For example, accessing credit nanced. Policies that provide both paternity and often requires collateral, preferably land or im- maternity leave and make the former mandatory mobile assets. Women are thus at a disadvantage (as in Iceland, Norway, and Sweden) have the because they have lower or less secure access to advantage of not being biased against women land and are disproportionately employed in the while also helping to shift the underlying norms service sector where capitalization is lower and about care. But this approach may be financially output is often intangible. These forces may be beyond the fiscal capacity of many developing further reinforced by ­ gender-based preferences economies. in the households that can lead to unequal re- Improvements in infrastructure services— source allocations (of land, for example) to male especially water and electricity—can help free and female members. up women’s time spent on domestic and care Policies need to focus on these underlying work. Electrification in rural South Africa, for determinants of differential access—leveling instance, has increased women’s labor force par- the institutional playing field by strengthening ticipation by about 9 percent; in Bangladesh, it women’s ownership rights, correcting biases in has led to more leisure time for women. In Paki- service delivery institutions, and improving the stan, putting water sources closer to the home functioning of credit markets. was associated with increased time allocated to Strengthening women’s land and ownership market work. Other studies show no impact rights can help female farmers and entrepreneurs. on market work but noticeable impacts on lei- The main constraint that needs to be addressed sure time, which also increase women’s welfare is the restriction on women’s ability to own and (chapter 7). inherit assets and to control resources. Experi- Interventions can also focus on reducing the ence from India and Mexico shows that equal- (time) transaction costs associated with access to izing provisions of inheritance laws between markets. Better and more effective transport op- women and men increases asset ownership by tions can reduce the time costs associated with women. Discriminatory land laws, which lie 28 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 at the root of agricultural productivity gaps in such as Grameen Bank in Bangladesh and many countries, also need to be reformed to FINCA in Peru. Microcredit has now evolved provide, at least, for joint ownership in marriage, beyond group lending to such schemes as Banco increasing women’s ability to use land in access- Sol in Bolivia and Bank Rakyat Indonesia that ing economic opportunities. An even better way offer larger individual loans and rely on repay- to secure married women’s land rights (espe- ment incentives rather than peer monitoring. cially in the case of divorce or death of a spouse) Lack of access to formal credit can also be sur- is mandatory joint land titling. In two regions mounted through financial innovation and by in Ethiopia where land certification involved the adapting a credit model that addresses the needs issuance of joint titles to both spouses, women’s of small businesses, as Access Bank in Nigeria, names appeared on more than 80 percent of all DFCU in Uganda, and Sero Lease and Finance titles, four times the 20 percent in the region in Tanzania have done. Recognizing that women where the certificate was issued only in the name are less likely to have established credit records of the household head.65 than men, and lower asset bases on which to Correcting biases in service delivery institu- draw for collateral, these large commercial tions such as the workings of government land banks partnered with the International Finance distribution and registration schemes and agri- Corporation to develop new instruments to cultural extension agencies can improve wom- support and extend credit services to female- en’s access to economic opportunity in many owned businesses and female entrepreneurs. In- countries. Redressing these biases requires ac- terventions included developing new products tions on several fronts. First, service providers such as loans that are collateralized with equip- need to target women explicitly and addition- ment or based on cash flow as well as training ally. For example, land redistribution programs for the staff of financial institutions and strate- that target the head of the household will not gic support to help banks increase their number serve women well. Instead, governments can put of female clients. Initial experience with these in place mandatory joint titling on redistributed interventions shows an increase in the shares of land that is coupled with gender sensitization female clients using financial services and taking policies and more female representation on local out larger loans with better-than-average repay- land boards. Second, women can be given some ment (chapter 7). power within the service delivery organization, including in setting priorities. For agricultural Addressing discrimination in labor markets extension, for example, women could be put In wage employment, the underrepresentation in decision-making positions at the Ministry of women in certain sectors or occupations can of Agriculture. Third, technology can be used feed discriminatory beliefs among employers to expand the reach of services, as was done in (or reinforce preexisting beliefs) that women are Kenya for agricultural extension through the not suitable workers or good candidates for em- use of call centers. Fourth, improved monitor- ployment. The importance of networks (often ing can make the problem visible. Finally, the gendered) in job search and professional pro- female users of the service should be provided motion can further reinforce women’s exclusion information on the level of service they are due. from certain jobs, positions, sectors, or occupa- This step can be aided by building the collective tions. Breaking this information problem and element of demand—for example, supporting expanding networks can be addressed through women’s farmer organizations. three main types of policies: active labor market Improving the functioning of credit markets by policies; affirmative action programs; and group addressing the information problems caused by formation and mentoring interventions. lack of experience with women borrowers can Active labor market policies combine training, help address productivity gaps between women placement, and other support to enable women and men in agriculture and entrepreneurship. to enter or reenter the workforce. Although Microcredit schemes have been the most com- these policies are not typically motivated by the mon way of addressing these problems, by help- goal of narrowing gender wage gaps, evidence ing women access small-scale credit and build a from Argentina, Colombia, and Peru indicates track record of borrower performance. Typically, that they can increase women’s employment these take the form of group lending schemes and earnings in the formal sector by allowing Overview 29 Box 6  Catalyzing female employment in Jordan Despite growing education levels, labor force par- service, re­ sume writing, interviewing, and positive ticipation rates for women in the Middle East and thinking. North Africa remain very low. In Jordan, only 17 There appears to be strong demand for these percent of 20- to 45-year-old women work, com- policies. Despite low employment rates, the major- pared with 77 percent of men. This labor force ity of recent female graduates want to work: 93 per- participation gap also holds among the more ­ cent say they plan to work after they graduate, and educated; among community college graduates ­ 91 percent say they would like to work outside the it starts immediately on graduation. These low house after they are married. Of those invited to employment rates make it difficult for new gradu- attend the training courses, 62 percent completed ates seeking to enter the labor market. With few them, with unmarried women more likely to attend. women employed, young women lack role models Those who began the courses gave them positive to follow into employment as well as the network reviews, claiming the courses had given them much connections to help them find jobs. Employers, more confidence to begin searching for jobs. Four lacking experience with working women, may be months into the wage subsidy program, about a reluctant to hire women if they believe women are third of those using vouchers had found a job. less committed to staying employed. Early results from a midline evaluation suggest The Jordan New Opportunities for Women that job vouchers have significant employment (Jordan NOW) program is a pilot to rigorously eval- effects: employment rates among graduates who uate the effectiveness of two policies: short-term received vouchers alone or vouchers plus training wage subsidies, and employability skills training. are between 55 to 57 percent compared with 17 to Short-term wage subsidies give firms an incen- 19 percent among those who received training tive to take a chance on hiring young female grad- alone or received neither training nor vouchers. In uates and an opportunity to overcome stereotypes all groups, employment effects are higher for by directly observing young women working for unmarried women. Financial empowerment (mea- them. The subsidies can also give young women sured as the proportion of women who have their more confidence to search for work and to own money and can decide how to use it) also approach employers. In the pilot, each voucher increased significantly for all who received either has a value equal to the minimum wage for six vouchers or training or both. Follow-up surveys months. will determine whether these employment effects Employability skills training augments the of job vouchers are sustained in the longer term technical skills that graduates learn in community and will also focus on other measures of empower- college with the practical skills to find and suc- ment and changes in attitudes. The surveys will ceed in employment. Many employers say recent also allow further investigation of the link between graduates lack these interpersonal skills and other marriage and work, given the early findings that basic job skills. In the pilot, students received 45 married women are less likely to attend the train- hours of instruction in team building, communica- ing, less likely to use the vouchers, and less likely to tions, presentations, business writing, customer be employed. Source: WDR 2012 team. participants to better communicate their abili- rules, careful monitoring of impacts, and cred- ties to employers. A similar program currently ible sanctions for noncompliance are essential.66 under implementation and evaluation in Jordan Where such programs have been implemented, shows promising signs of success (box 6). the effect has clearly been to redistribute wage Affirmative action is another way to over- employment from men to women. And, while come information failures. The goal is to push the economic efficiency of such policies is still de- women’s participation in wage employment to bated, the most comprehensive evidence (from a “critical threshold” (often argued to be about long experience in the United States) points to 30 percent), where information failures and little or no adverse efficiency effects.67 This ex- networks no longer bind. Experience (mainly perience and that in other countries also show from rich countries) shows that affirmative ac- that any potential negative efficiency impacts tion works best if it is mandatory. Affirmative can be addressed by ensuring that the affirma- action also can be implemented through public tive action programs are temporary and are re- sector employment and contracting, but clear moved once women’s representation reaches the 30 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 needed critical mass. In the absence of explicit Because these constraints resemble the ones affirmative action policies, female employment that limit women’s prospects in labor markets, in the public sector in fairly large numbers can the policy solutions are similar. Quotas and have a demonstration effect. In rich countries, other types of affirmative action have promoted public sector growth has been important in in- women’s political representation at various lev- tegrating women into the labor markets.68 els of politics. Such measures range from vol- Supporting the creation of women’s networks untary commitments by political parties to in- can be effective where gendered networks hinder clude women candidates on their electoral lists women workers, farmers, or entrepreneurs. Such to specifying shares of legislative seats reserved interventions work best when they combine the for women. Which option is best for a country building of social capital and networks with the depends on its political system. For example, provision of training, information, and mentor- reserving individual seats for women will not ing. One example is the Jordan NOW program work in proportional representation systems, described in box 6. Another, more established whereas voluntary party quotas may work when example, is the Self Employed Women’s Asso- parties have strong leadership and internal dis- cipline. Whatever the system, its design and en- ciation in India, which has evolved into an ef- forcement are critical. In Spain, where positions fective organization representing the interests of on the ballot for the Senate were in alphabetical a large number of informal sector workers and order, parties tended to choose women with last entrepreneurs, providing extensive information, names that landed lower in the ballot and who support, and training services to its members. thus were less likely to win a seat. 70 Removing discriminatory treatment in la- Broader tensions also need to be acknowl- bor laws and regulations can promote women’s edged and taken into account if quotas are used economic opportunities. Among these laws and to increase women’s political representation. regulations, the priority should be to revisit Mandatory quotas involve the state circum- the limits (including outright bans) on part- scribing part of the democratic process, so this time work in many countries. Because women distortion has to be balanced against the need to provide a disproportionate share of household redress persistent inequalities. One option, taken and care work, such restrictions end up limit- by local governments in India, is to implement ing work options for women much more than quotas on a rolling basis—with a different set of for men. Relaxing these prohibitions would seats chosen for reservation in different elections give women more opportunities for paid em- over time. And as with all affirmative action, it ployment. In Argentina, removing the ban on helps to specify a clear goal or time period up part-time contracts in the formal sector led to front. The structure of the reservation also mat- a significant shift of women with children from ters. Designating particular seats for women part-time work in the informal sector to part- runs the risk of creating “token” women’s seats. time contracts in the formal sector.69 Quotas have increased female representa- tion. In Mexico, candidate quotas increased the Policies to shrink differences in voice share of women in parliament from 16 percent to more than 22 percent. Reserved seats in Mo- Increasing women’s societal voice rocco increased the proportion of women in Women generally have less voice than men parliament from less than 1 percent to almost 11 both in society and in households. At the soci- percent. Quotas in local governments in India etal level, income growth does little to reduce also showed that such measures can change un- these gaps. Norms that dictate that politics is for derlying beliefs among voters about the efficacy men; beliefs that women make worse leaders, of women politicians, even in a short period, which are fed in part by the low participation and increase the proportion of women elected of women in politics; norms around care and to these positions even after the quotas are no housework, which limit the time available to longer in place.71 women to participate in formal political institu- Affirmative action in the political realm tions; and the gendered networks within politics needs to be complemented by measures that that work against women all matter more than increase women’s voice in other societal institu- income growth. tions, such as trade unions, corporations, the ju- Overview 31 diciary, and professional associations. This can work, and jobs requiring higher skills.72 Mo- be done through quotas as well as by mentoring rocco also reformed its family laws in the 1990s, schemes, women’s networks, and skills develop- and in 2004 the new Family Code completely ment in these realms targeted to women. Collec- eliminated references to the husband as the head tive action by women’s groups can be particu- of the household. larly effective in this, as with the Self Employed Greater effort is also needed to make these Women’s Association in India. More generally, rights more effective and justice systems more because women tend to be better represented responsive to women’s needs. Interventions are in less formal organizations, laws and regula- needed on both the supply side and the de- tions should ensure a level playing field for such mand side. Greater capacity of the institutions ­ organizations. that apply the laws, more accountability in the justice system to promote predictable outcomes Increasing women’s voice within households in line with the law, and procedures to promote The muted voice of women within their house- women’s access to justice and women’s repre- holds reflects the combined influence of their sentation in judicial institutions are critical on more limited access to economic opportunities, the supply side. Also important are mechanisms the nature of social norms, the legal framework, for the implementation of laws. Evidence from and the enforcement of laws. Key determinants Ethiopia illustrates how the procedures around of control over household resources are ac- mandatory joint land titling helped promote cess to economic opportunities and the legal women’s rights over land. Women’s demand for framework—particularly rights over property enforcement of their rights can be promoted by and those that determine access to assets. For broadening literacy, increasing the accessibility domestic violence, social norms and the content of legal aid services, and reducing costs of legal and enforcement of laws are important. And for procedures. And data have to be collected and fertility, norms, bargaining power, and service made public so that the problems of women’s delivery are critical. access to justice are made more visible. Increasing control over household resources Reducing domestic violence Thus, the most promising policies to increase Reducing domestic violence requires action women’s voice in households center on reform- on multiple fronts. The goal is to prevent vio- ing the legal framework so that women are not lence before it happens. The first step is to en- disadvantaged in controlling household assets act laws that define different types of violence and expanding their economic opportunities. against women, prescribe mandates and duties For the legal framework, land laws and aspects for enforcement and investigation, raise societal of family law that govern marriage, divorce, awareness, and signal a government’s commit- and disposal of property are particularly im- ment. These laws must be put in place in coun- portant. A cross-cutting issue applies to the tries that lack them, especially in the Middle many countries where multiple legal systems East and North Africa, South Asia, and Sub- exist. Reconciling these systems, which may in- Saharan Africa. And countries with such laws on clude customary and religious law, is a priority, the books need to make them more specific and especially to ensure that all laws are consistent actionable. with a country’s constitution. Kenya made such A second step is to shift norms and behaviors changes in its recent constitutional reforms. around domestic violence to emphasize preven- Although reforms in these areas are po- tion. Education and awareness programs, such litically and socially complex and depend very as Soul City in South Africa, can change norms much on country context, experience shows that about domestic violence among both men and change is possible. Ethiopia reformed its family women. Increasing women’s bargaining power law in 2000 by eliminating the husband’s ability in their households—by improving women’s to deny permission for his wife to work outside economic opportunities and enhancing their the home and requiring both spouses to agree in control over resources and their ability to leave administering family property. The first phase marriages—also can change behavior. But in- of these changes shifted women’s economic ac- creasing women’s bargaining power can also tivities toward work outside the home, full-time risk increasing the likelihood of violence in the 32 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 short term. So, specific mitigation measures may contraception. Contraception uptake is higher be needed. when husbands are included in family planning Third, when violence does occur, victims education, as was shown in Bangladesh74 and in need timely and effective assistance ranging Ethiopia. 75 from the police and judiciary to health and The second is increasing the quality of fam- social services, as with support integrated in ily planning services. Improvements in this re- Malaysia’s one-stop crisis centers at govern- gard need to focus on three areas. First, a suf- ment hospitals. Service providers—police and ficient range of contraceptive options needs to judiciary, health, and social services—need to be available. Second, adequate information on target women explicitly and additionally. Tar- the available options, their side effects, and the geting women also requires bringing services advantages and limitations of different methods closer to women to deal with time and mobility needs to be given to women so they can make constraints—for example, by providing com- an informed decision. Third, services need to munity paralegals and mobile legal aid clinics be provided in a manner that protects the in- that enable women to use the justice system. dividual’s or the couple’s privacy. This will re- In many contexts, bringing services closer to quire training health care providers in protocols individuals (demand) can be combined with designed specifically for family planning. Recent increasing the awareness of service delivery or- experience in Zambia shows that very different ganizations, particularly management, about outcomes regarding fertility and contraception gender issues (supply). PEKKA Women’s Le- can result depending on whether women are ap- gal Empowerment in Indonesia trains village proached individually and in private or together paralegals, with a focus on domestic violence with their partners. and family law. Another way of improving women’s access to justice is to increase the share Policies to prevent the reproduction of of women in the judicial and police forces re- gender inequality across generations sponsible for addressing domestic violence. The reproduction of specific gender inequali- The Indian state of Tamil Nadu introduced ties across generations gives rise to “gender in- 188 all-women police units to cover both rural equality traps,” which are likely to most affect and urban areas and to focus on crimes against the poor and excluded in society. Women’s lack women. These units increased women’s com- of political voice means that the market and in- fort in approaching the police, including mak- stitutional failures feeding gender inequality are ing reports of domestic abuse.73 unlikely to be corrected. Income growth alone does little to address the processes that underlie Increasing control over fertility these persistent gaps. The previous sections dealt Increasing women’s control over their fertility with policies to address three of these gaps that requires actions in several areas. Availability of reproduce over generations—reaching pockets family planning services is still limited in parts of remaining disadvantage in education, in- of the world. In some cases the underserved creasing women’s voice and participation in so- population covers entire countries, but more cietal institutions, and increasing women’s voice often these women live in specific geographic within households. Here, we address measures areas within countries —usually rural areas—or to address the gender inequalities in human are poor. For these groups, improved delivery of capital, opportunities, and aspirations that are family planning services is a priority. set early in life. Control over fertility decisions—the num- Decisions in adolescence can shape skill ac- ber and spacing of children—goes beyond is- quisition, health outcomes, and economic op- sues related to provision of reproductive health portunities. Adolescence is also a period when services, so two other policy areas need to be one’s lifelong aspirations are molded, and when addressed. The first is to boost women’s ability social norms and perceptions start to bind for within the household to voice their preferences re- boys and girls. Horizons for girls often shrink, garding number and spacing of children. As dis- especially for poor girls or girls in rural areas cussed earlier, access to economic opportunities, where distance and norms for mobility can control over assets, and appropriate laws help. So be a significant constraint. Empowering ado- does educating men on the benefits and use of lescents to make better choices for themselves Overview 33 can make a big difference to their lives, to their and reported sexual behavior in boys.83 For families, to their communities, and, as future adolescents, the promotion of contraception, workers and citizens, to society more broadly. when combined with education interventions Interventions need to build human and social and skill building, and appropriately targeted capital; facilitate the transition from school to to cultural and social settings, can be effective work; and increase their aspirations and agency. in reducing unintended pregnancies.84 Such a Efforts to influence and reduce risky behavior program for adolescent girls in Uganda resulted are also important. in a significant increase in condom use and a Scholarships and conditional cash trans- lower number of children among participants.85 fers can increase school attendance and reduce Sometimes, economic empowerment alone can dropout rates for adolescents, especially girls. have a marked impact. A recent evaluation of a These positive impacts are well documented youth job training program in the Dominican in Latin America in countries such as Colom- Republic that included life skills training plus an bia, Ecuador, Mexico, and Nicaragua.76 More apprenticeship showed a significant reduction recently, evidence from Africa is beginning to in pregnancies among participants.86 show similar results. In Malawi, fairly small Exposure to female role models whose posi- cash transfers to girls increased enrollment and tions of leadership or power contradict stereo- reduced dropout rates.77 Moreover, while these types of women’s role can reduce the intergener- transfers were aimed at education, they had ational transmission of gender norms. A study of benefits in other realms, such as reducing HIV political reservation for women in India showed infections. Other tools can also be brought to that teenage girls who have repeated exposure to bear to help girls stay in school. Providing them women leaders are more likely to express aspira- with information on the returns to schooling tions that challenge traditional norms, such as is one such tool: for example, in Madagascar, a desire to marry later, have fewer children, and information on earnings for primary school obtain jobs requiring higher education.87 In- completers provided to boys and girls as well as creased economic opportunities for young girls their parents increased attendance rates by 3.5 can also change their own and their communi- percentage points.78 In the Dominican Republic, ties’ perceptions of gender roles for adolescent a similar effort to provide accurate information girls. A study of a program in Delhi that linked on actual returns to education to boys also had communities to recruiters for high-paying tele- a positive impact.79 Other evidence suggests that phone work found that these communities were incentives (prospects for winning a scholarship more likely to have lower expectations of dowry or direct payments for performance) can affect and to find it acceptable for women to live alone children’s own perception of their abilities and before marriage and to work before and after can improve test scores.80 marriage or childbirth.88 Vocational training targeted specifically to youth increased both the likelihood of employ- Making other policies “gender-smart” ment and the earnings for young women in Understanding how gender factors into the Colombia and Peru.81 In Kenya, providing in­ workings of households, markets, and institu- formation to young girls about the relative re- tions can matter for policies even when improv- turns to vocational training in male or female- ing gender equality is not the main objective. dominated industries increased girls’ enrollment Why? Because gender-differentiated failures in in trade school courses that prepared them for markets, gender biases in institutions, and the typically male-dominated trades that yielded way gender relations play out in the household higher returns.82 The Adolescent Girls Initiative all affect (and sometimes constrain) the behav- aims to evaluate a range of these interventions, ior of both men and women. These changes in including skills training and mentoring, in a behavior can affect how men and women re- number of low- and middle-income countries spond to policies. Failing to take them into ac- (box 7). count can thus mean the policy will have unin- Health education programs have proved to be tended consequences or simply not work. effective in reducing risky behaviors. A program Take relations within the household. They in rural Tanzania substantially improved sexual clearly affect how the household responds to knowledge, attitudes, condom use in both sexes, policy—sometimes with unintended conse- 34 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Intervening Box 7  early to overcome future labor market failures—The Adolescent Girls Initiative The Adolescent Girls Initiative, a public-private trades attractive to women, with some trained as partnership, promotes the transition of adoles- electricians, masons, and mobile phone repair cent girls from school to productive employment technicians. through innovative interventions that are tested Lessons from implementation highlight girls’ and then scaled up or replicated if successful.a distinctive social capital needs, which must be Under way in Afghanistan, Jordan, Lao People’s addressed to facilitate the uptake of economic Democratic Republic, Liberia, Nepal, and South opportunities among often vulnerable and iso- Sudan (and soon in Rwanda), the initiative targets lated young females. Trainers in the Liberia pilot about 20,000 adolescent girls and young women formed girls into teams of three or four, who made ages 16–24. public commitments to support one another, both Interventions range from business develop- inside and outside the classroom, throughout the ment skills training and services to technical and training. The positive peer pressure helped keep vocational training, targeting skills in high demand. attendance rates high, with almost 95 percent In all projects, girls receive life-skills training to completing the training, and addressed the vari- address the most important barriers to their eco- ety of educational levels among participants. nomic independence. Each country intervention is Another promising innovation from the Libe- tailored to the local context and specific needs of ria pilot was a formal savings account at a local adolescent girls. Because the evidence on what bank for all participating girls, with an initial works is thin, rigorous impact evaluation is a part deposit of $5. The savings accounts not only of the initiative. enabled the girls to practice their financial literacy The skills training aims to equip girls with tech- skills beyond the classroom but built trust with nical skills with proven demand in the local labor formal financial institutions, and girls expressed market. In all pilots, training providers are asked to satisfaction with being connected to the modern conduct market assessments before selecting economy for the first time. trades for which training is developed and offered. Job fairs were organized to market the pro- While the focus is on matching skills to the mar- gram to potential employers interested in placing ket, the results in many cases challenge norms for girls in internships or jobs. Private sector human ­ gender-appropriate occupations. resource and career development specialists met In Liberia, participating young women are with trainees individually to impart their knowl- offered six months of training for jobs in house edge about the industry, coach them on profes- painting, professional driving, and security guard sionalism in the workplace, and give constructive services. In Nepal, they are offered three months feedback on the skills demonstrated. These one- of occupational skills training followed by a man- on-one meetings offered girls the opportunity to datory skills test and three-month job placement. build networks and to tap into industry-specific The focus is on identifying nonstereotypical information crucial for new job entrants. a.  Current donors to the Adolescent Girls Initiative include the Nike Foundation, a main partner of the initiative, as well as the governments of Australia, Denmark, Norway, Sweden, and the United Kingdom. The Bank’s Gender Action Plan also supports the initiative’s country projects. Pledges to the initiative stand at $22 million. quences. For example, many conditional cash of these issues for mothers and families (Co- transfer programs initially targeted women be- lombia, Peru), or even dedicated social workers cause—cognizant of how women were likely to (Chile). spend money differently from men—it seemed So, how can considerations of gender in- a way to get more of the transfer spent on chil- equalities and their underlying determinants dren’s endowments. But the transfers changed be integrated into broader policy and program bargaining power within the household and, in design? The analytical framework in this Report some cases such as Mexico, resulted in short- provides a guide. First, what happens inside the term increases in domestic violence. While this household shapes the impact of policies. An ex- effect can disappear or change in nature in the ample comes from Papua New Guinea. Gender longer term, a number of later transfer pro- roles in harvesting oil palm have men climbing grams included conditions to discourage do- the trees and harvesting the fruits, while women mestic violence (Brazil), training and awareness collect the fruit that has fallen on the ground. Overview 35 The oil palm industry came to the realization The political economy of that 60–70 percent of the fruit on the ground reforms for gender equality was not being collected. They tried multiple initiatives designed to deal with constraints Well-designed public policies to address specific women faced, including giving the women spe- market failures and institutional or normative cial nets to use, and timing the collection to constraints can support significant advances in deal with women’s care duties. None of these gender equality. But the choice of policies and worked. Finally, the Mama Lus Frut scheme was their implementation does not occur in a vac- introduced whereby women received their own uum. Policies must be attuned to countries’ in- harvest record cards and were paid directly into stitutional, social, and political environment and their personal bank accounts. Yields increased to the societal actors involved. It is important to significantly, as did female participation in oil understand how reforms actually take place and palm harvesting. what factors allow them to be sustained so that Second, many non-gender-focused policies they produce change. and programs can benefit from taking into ac- Two characteristics of gender reform pro- count women’s underrepresentation in markets, cesses are worth noting. First, as with all reforms, sectors, or occupations, a situation that can cause they redistribute resources and power between information problems not only for women but groups in society, including between men and also for those who seek to employ them, lend to women. Even when policies to advance gender them, or provide them with services. One exam- equality are well chosen and enhance economic ple of how to take women’s underrepresentation efficiency, some groups may lose as a result. into account is Ecuador’s program to expand Second, such reforms often confront power- credit bureau databases to include microfinance. ful societal norms and beliefs regarding gender This intervention will help microfinance insti- roles. Each of these features means there will tutions make better lending decisions, indepen- likely be opposition, and managing this push- dent of to whom they are lending. And because back is the key to successful reforms. microfinance clients are predominantly women, A range of countries—rich and develop- it will also help them access a broader range of ing—show that several aspects of the political financial services. economy of reforms are especially relevant to Third, policy design should seek to level the gender equality. First, reforms are most likely to playing field for women and men, especially succeed when support for them is broad-based. where laws and regulations treat them dif- It is essential, then, to build coalitions that mo- ferently and where systems enforce laws and bilize around the reforms. These coalitions can regulations differentially, even when nomi- include nonstate actors such as political parties, nally equal. Looking for and fixing this type of trade unions, civic organizations and associa- discrimination when revising laws and regula- tions, and the private sector. Women’s groups in tions or enforcement mechanisms can provide particular have been a driving force for greater an opportunity to improve gender equality as a gender equality in labor legislation and family secondary benefit. Take the case of taxes, which law. For instance, women workers in the infor- can explicitly discriminate against women mal sector have challenged their employers and when women face different rates for the same sometimes the state through such organiza- income as men. For example, in Morocco, tions as the Self Employed Women’s Association the tax allowance for children is allocated to in India and Nijera Kori in Bangladesh. These men, so men face a lower tax burden. Women groups have provided voice for women and cre- receive this allowance only if they prove that ated space for public action to counter the resis- their husband and children depend financially tance to reform. on them. This design is neither efficient not Women and men are partners in improving gender equitable. gender equality. While most initiatives that call When these considerations are factored into on men to support gender equality are still small, general policy design, the policies are more signs point to broader engagement in many areas likely to reach their intended objectives, and it and growing male support for women’s rights becomes easier for policy makers to tweak them in many developing countries. For instance, the to improve gender equality in the process. Rwanda Men’s Resource Center engages men 36 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 and boys in combating gender-based violence. such “incremental” reforms may not be enough In surveys of male attitudes toward gender to overcome the path dependence and institu- equality in Brazil, Chile, Croatia, and Mexico, tional rigidities that result in persistent gender adult men overwhelmingly express the view that inequality. Bold government action with “trans- “men do not lose out when women’s rights are formative” reforms may be necessary to alter promoted.”89 Even in India, where men in this social dynamics and move to a more equitable survey were less supportive of gender equality equilibrium. In choosing between incremen- overall, their support for some policies, such as tal and transformative policies as part of gen- quotas in universities and places in government, der reforms, the challenge for policy makers is was strong. to balance the pace of change with the risks of Second, firms—big and small—can articulate reversal. Incremental policies will bring about a compelling business case for gender equality. In change only slowly. But transformative policies a fast-changing global economy, the demand for can risk a backlash. A way forward is to be se- skills has swelled, encouraging firms to expand lective in implementing transformative policies their talent pool. Businesses have sought not and ensure that adequate attention is paid to only to attract and recruit female talent but also their implementation. to retain it through measures to facilitate work- life balance. Firms know that a diversity of opin- ions can enrich decision making and stimulate A global agenda for greater ingenuity. And gender equality per se has grown gender equality to be a desirable trait that customers and inves- Domestic action is central to reducing inequali- tors look for. Corporate social responsibility is ties. Global action—by governments, people an avenue for firms to enhance competitiveness and organizations in developed and developing through product differentiation and capture the countries, and by international institutions— loyalty of women’s growing market power. cannot substitute for equitable and efficient Third, shocks and exogenous changes can domestic policies and institutions. But it can present policy makers with windows of oppor- enhance the scope for and impact of domestic tunity to launch reforms that can improve gen- policies. And it can influence whether global der outcomes. Such windows sometimes stem integration and the opportunities it brings— from unpredictable circumstances, such as a through information, mobility, and technolo- national disaster. In 1998 the disaster that fol- gy—lead to greater gender equality and better lowed Hurricane Mitch in Nicaragua facilitated lives for all women, or just for some. a dialogue on domestic violence. A national Global action should focus on areas where campaign and the enactment of legislation fol- gender gaps are most significant both intrinsi- lowed. Other windows come from shifts in the cally and in terms of their potential develop- political or economic landscape. The changes in ment payoff—and where growth alone cannot Spain during the democratic transition in the solve the issues. In other words, international late 1970s were particularly dramatic for fam- action should focus on complementing country ily law and reproductive rights. And still oth- efforts along the four priority areas identified in ers emerge from the advocacy of transnational this Report: agencies and role modeling in the global agenda. For example, the monitoring and advocacy by • Reducing excess female mortality and closing Colombian local women’s groups in CEDAW education gaps where they remain informed the expansion of reproductive health • Improving access to economic opportunities guarantees in the Colombian Constitution, ad- for women opted in 1990, and facilitated greater access to contraception. • Increasing women’s voice and agency in the Finally, there are multiple paths to reform. household and in society Often governments follow societal cues in pacing • Limiting the reproduction of gender inequal- and pushing reforms. When policy formulation ity across generations and implementation follow cues from ongoing shifts in markets and social norms, convergence In addition, there is one cross-cutting prior- and alignment can fuel sustainable change. But ity: supporting evidence-based public action Overview 37 through better data, better knowledge genera- and across borders. Such partnerships could tion and sharing, and better learning. be built among those in the international de- The motivation for an agenda for global ac- velopment community around funding issues, tion is threefold. First, progress on some fronts with academia and think tanks for the purpose requires channeling more resources from rich of experimentation and learning, and, more to developing countries (for example, to cre- broadly, with the private sector to promote ac- ate greater equity in human endowments or to cess to economic opportunities. Together, these tackle the root causes of excess female mortal- partnerships could support countries in lever- ity around the world). Second, effective action aging the resources and information needed to sometimes hinges on the production of a pub- successfully promote gender equality in today’s lic good, such as the generation of new (global) globalized world. information or knowledge. And third, when the The relative importance of these activi- impact of a particular policy cuts across borders, ties will obviously vary across countries. Table coordination among a large number of coun- 2 provides a bird’s eye view of the proposed tries and institutions can enhance its effective- agenda for global action (described in more de- ness, not least by building momentum and pres- tail in chapter 9). Areas marked with a check are sure for action at the domestic level. those where new or additional action is needed Based on these criteria, initiatives included or where a refocus of existing initiatives is called in the proposed agenda for global action can be for. Of course, there are also important ongoing grouped into three types of activities: efforts in the areas not marked with checks—for instance, innovation around the delivery models Providing financial support. Improvements for the prevention of HIV/AIDS, or partnerships in the delivery of clean water and sanitation or focused on adolescents. In these latter areas, the better health services, such as the ones needed focus should be on sustaining ongoing efforts and to bring down excess female mortality among partnerships, and meeting prior commitments. girls and mothers in the developing world, will Finally, the framework and analysis presented require significant resources—often beyond the in the Report provide four general principles for means of individual governments, particularly policy and program design, which can enhance those of relatively poor countries. The interna- the impact and effectiveness of global action tional development community can financially across all priority areas. These principles are: support countries willing and able to undertake • Comprehensive gender diagnostics as a pre- such reforms in a coordinated manner through condition for policy and program design. Gen- specific initiatives or funding facilities to ensure der disparities persist for multiple reasons: maximum impact and minimize duplication. there may be a single institutional or policy “fix” that is difficult and easily blocked; there Fostering innovation and learning. While a may be multiple reinforcing constraints in great deal has been learned about what works markets, formal institutions, and households and what does not when it comes to promot- that combine to block progress; or they may ing greater gender equality, the truth remains be deeply rooted in gender roles or social that progress is often held back by the lack of norms that evolve only slowly. Effective pol- data or adequate solutions to the most “sticky” icy design requires a good understanding of problems. That is the case, for instance, regard- which of these situations prevails in a partic- ing gender differences in time use patterns and ular context, and of where and what the the norms around care they stem from. The de- binding constraints are. To be useful, this di- velopment community could promote innova- agnostic must drill down into what happens tion and learning through experimentation and in households, markets, and formal institu- evaluation in ways that pay attention to results tions, their interactions, and how they are and process, as well as to context, and thus facili- shaped by social norms. tate a scaling-up of successful experiences. • Targeting determinants versus targeting out- Leveraging effective partnerships. As chapter comes. In choosing and designing policies, 8 makes clear, successful reform often requires it  is necessary to target the market and in­ coalitions or partnerships that can act within stitutional constraints that generate existing 38 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Ta b l e 2  The agenda for global action at a glance Directions for the global development community New/additional initiatives Providing financial Fostering innovation Priority area Leveraging partnerships that need support support and learning Increasing access to education √ √ among disadvantaged groups Increasing access to √ √ clean water Closing gender gaps in human endowments Increasing access to √ √ √ specialized maternal services Strengthening support for prevention and treatment of √ √ HIV/AIDS Increasing access to child Promoting women’s access to care and early childhood √ √ economic opportunities development Investing in rural women √ √ Increasing women’s access to √ Closing gender gaps in voice the justice system and agency Shifting norms regarding √ √ violence against women Preventing intergenerational Investing in adolescent girls reproduction of gender √ and boys inequality Generating new information √ √ Supporting evidence-based public action Facilitating knowledge sharing √ and learning Source: WDR 2012 team. gender gaps, rather than the outcomes them- tures, and as a result so do household and selves. These constraints may be multiple and individual behaviors. This implies that the even outside the immediate domain where same policy can have very different results the outcome is observed. depending on the contexts—or, as the dis- cussion in chapter 8 makes clear, that there • “Upstreaming” and strategic mainstreaming. are multiple paths to reform. Because gender gaps are often the result of multiple and mutually reinforcing constraints, effective action may require coordinated Notes multisectoral interventions, or sequential in- terventions. And in many instances, such   1.  Sen 1999.   2.  Esteve-Volart and Bagues 2010. interventions can take the form of general ­   3.  Gilbert, Sakala, and Benson 2002; Vargas Hill and policies that are made “gender smart” by in- Vigneri 2009. corporating gender-related issues into their   4.  Udry 1996. design and implementation. To maximize   5.  FAO, IFAD, and ILO 2010. impact, it is thus necessary for gender issues   6.  Cuberes and Teignier Baqué 2011; Hurst and oth- to be upstreamed from specific sector prod- ers 2011. ucts and projects to country and sector pro-   7.  Do, Levchenko, and Raddatz 2011. grams. That will allow for more strategic   8.  Do, Levchenko, and Raddatz 2011. gender mainstreaming.   9.  Do, Levchenko, and Raddatz 2011. 10.  Haddad, Hoddinott, and Alderman 1997; Katz • No one size fits all. The nature, structure, and and Chamorro 2003; Duflo 2003; Thomas 1990; functioning of markets and institutions vary Hoddinott and Haddad 1995; Lundberg, Pol- widely across countries, as do norms and cul- lak, and Wales 1997; Quisumbing and Maluccio Overview 39 2000; Attanasio and Lechene 2002; Rubalcava, 48.  Fernández and Fogli 2009; Fogli and Veldkamp, Teruel, and Thomas 2009; Doss 2006; Schady and forthcoming; Farré and Vella 2007. Rosero 2008. 49.  Agarwal and Panda 2007. 11.  Doss 2006. 50.  Gage 2005. 12.  Thomas 1990. 51.  Yount and Carrera 2006; Castro, Casique, and 13.  Qian 2008. Brindis 2008. 14.  Luke and Munshi 2011. 52.  Abrahams and others 2009. 15.  Thomas, Strauss, and Henriques 1990; Allendorf 53.  Kishor and Johnson 2004. 2007. 54.  World Bank 2005. 16.  Andrabi, Das, and Khwaja 2011; Dumas and 55.  Chung and Das Gupta 2007. Lambert 2011. 56.  Waddington and others 2009. 17.  Felitti and others 1998; McEwen 1999. 57.  Chioda, Garcia-Verdú, and Muñoz Boudet 2011. 18.  Kishor and Johnson 2004; Jeyaseelan and others 58.  Björkman and Svensson 2009. 2007; Hindin, Kishor, and Ansara 2008; Koenig 59.  Cotlear 2006. and others 2006; Martin and others 2002. 60.  Lim and others 2010. Janani Suraksha Yojana 19.  Miller 2008. also had significant impacts on perinatal and 20.  Beaman and others, forthcoming; Chattopad- neonatal deaths, which declined by 3.7 deaths per hyay and Duflo 2004. 1,000 pregnancies and by 2.5 deaths per 1,000 21.  Beaman and others, forthcoming. live births, respectively. The study was unable to 22.  Agarwal 2010a; Agarwal 2010b. detect an ­effect on maternal mortality; however, 23.  World Bank 2005. perhaps because maternal death is a relatively 24 World Bank 2001. rare event and the sample size of the study was 25.  World Bank 2011. big enough only to detect very large effects. 26.  See World Bank (2001) and also World Bank 61.  Prata and others 2010; WHO and others 2010. (2011b), from which we draw for the conceptual 62.  Baris, Mollahaliloglu, and Sabahattin 2011. framework. 63.  Fiszbein and others 2009. 27.  Chioda, Garcia-Verdú, and Muñoz Boudet 2011. 64.  FAO 2003. 28.  World Bank 2008. 65.  Deininger, Ali, and Zevenbergen 2008. 29.  Lewis and Lockheed 2006. 66.  Leonard 1989, Holzer and Neumark 2000. 30.  Chioda, Garcia-Verdú, and Muñoz Boudet 2011. 67.  Holzer and Neumark 2000. 31.  For a detailed description of the methodology, 68.  Gornick and Jacobs 1998; OECD 1993; Schmidt which builds on Anderson and Ray (2010), see 1993. chapter 3. 69.  Bosch and Maloney 2010. 32.  The problem of many missing girls was first doc- 70.  Esteve-Volart and Bagues 2010. umented by Sen (1992), Coale (1984), and Das 71.  Quotas are implemented on a rotating basis Gupta (1987). across localities. 33.  WHO, UNICEF, UNFPA, and World Bank 2010. 72.  Gajigo and Hallward-Driemeier 2011. 34.  FAO 2011. 73.  Natarajan 2005. 35.  For the purpose of the discussion in this chap- 74.  Barker and Ricardo 2005. ter, the term “entrepreneur” refers to individuals 75.  Terefe and Larson 1993. who are self-employed with no employees own 76.  See Rawlings and Rubio (2003) for Mexico and account workers and with employees employers. Nicaragua, Barrera-Osorio and Linden (2009) 36.  Sabarwal, Terrell, and Bardasi 2009; Bruhn 2009; for Colombia, and Schady and Araujo (2006) for Hallward-Driemeier 2011. Ecuador. 37.  Costa and Rijkers 2011. 77.  Baird and others 2009. 38.  FAO 2011. 78.  Nguyen 2008. 39.  Nyamu-Musembi 2002. 79.  Jensen 2010. 40.  Deere and Doss 2006. 80.  Angrist and Lavy 2009; Kremer, Miguel, and 41.  Team estimates based on ICF Macro 2010. Thornton 2009. 42.  Reed and others 2010. 81.  Attanasio, Kugler, and Meghir 2008; Ñopo, Robles, 43.  WHO 2005. and Saavedra 2007; Hjort and ­ others 2010. 44.  United Nations Department of Economic and 82.  Hjort and others 2010. Social Affairs 2010. 83.  Ross and others 2007. 45.  Agarwal and Panda 2007. 84.  Gilliam 2010; Bearinger and others 2007. 46.  Pronyk and others 2006; ICRW 2006; Swamina- 85.  Bandiera and others 2011. than, Walker, and Rugadya 2008. 86.  Martinez and others 2011. 47.  Dercon and Singh 2011. 87.  Beaman and others 2009. 40 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 88.  Jensen 2010. Working Paper Series 13931, National Bureau of 89.  Barker and others 2011. Economic Research, Cambridge, MA. Attanasio, Orazio, and Valérie Lechene. 2002. “Tests of Income Pooling in Household Decisions.” Re- References view of Economic Dynamics 5 (4): 720–48. Baird, Sarah, Ephraim Chirwa, Craig McIntosh, and The word processed describes informally repro- Berk Özler. 2009. “The Short-term Impacts of a duced works that may not be commonly avail- Schooling Conditional Cash Transfer Program able through libraries. on the Sexual Behavior of Young Women.” Policy Research Working Paper Series 5089, World Bank, Abrahams, Naeemah, Rachel Jewkes, Lorna J. Martin, Washington, DC. Shanaaz Mathews, Lisa Vetten, and Carl Lombard. Bandiera, Oriana, Niklas Buehren, Robin Burguess, 2009. “Mortality of Women from Intimate Partner Markus Goldstein, Selim Gulesci, Imran Rasul, Violence in South Africa: A National Epidemiolog- and Munshi Sulaiman. 2011. “Economic Em- ical Study.” Violence and Victims 24 (4): 546–56. powerment of Female Adolescents: Evidence Agarwal, Bina. 1994. A Field of One’s Own: Gender from Uganda.” Presentation to the American Agri- and Land Rights in South Asia. Cambridge, U.K.: cultural Association, Pittsburgh, PA. Cambridge University Press. Baris, Enis, Salih Mollahaliloglu, and Aydin Sabahat- ———. 1997. “ ‘Bargaining’ and Gender Relations: tin. 2011. “Healthcare in Turkey: From Laggard Within and Beyond the Household.” Feminist Eco- to Leader.” British Medical Journal 342 (c7456): nomics 3 (1): 1–51. 579–82. ———. 2010a. “Does Women’s Proportional Barker, Gary, Manuel Contreras, Brian Heilman, Ajay Strength Affect Their Participation? Governing Singh, Ravi Verman, and Marcos Nascimento. Local Forests in South Asia.” World Development 2011. “Evolving Men: Initial Results from the 38 (1): 98–112. International Men and Gender Equality Survey ———. 2010b. Gender and Green Governance: The (IMAGES).” International Center for Research on Political Economy of Women’s Presence Within and Women and Instituto Promundo, Washington, Beyond Community Forestry. New York: Oxford DC. University Press. Barker, Gary, and Christine Ricardo. 2005. “Young Agarwal, Bina, and Pradeep Panda. 2007. “Toward Men and the Construction of Masculinity in Sub- Freedom from Domestic Violence: The Neglected Saharan Africa: Implications for HIV/AIDS, Con- Obvious.” Journal of Human Development and Ca- flict, and Violence.” Social Development Papers, pabilities 8 (3): 359–88. World Bank, Washington, DC. Alene, Arega D., Victor M. Manyong, Gospel O. Barrera-Osorio, Felipe, and Leigh L. Linden. 2009. Omanya, Hodeba D. Mignouna, Mpoko Bokanga, “The Use and Misuse of Computers in Educa- and George D. Odhiambo. 2008. “Economic Effi- tion: Evidence from a Randomized Experiment in ciency and Supply Response of Women as Farm Colombia.” Policy Research Working Paper Series Managers: Comparative Evidence from Western 4836, World Bank, Washington, DC. Kenya.” World Development 36 (7): 1247–60. Beaman, Lori, Raghabendra Chattopadhyay, Esther Allendorf, Keera. 2007. “Do Women’s Land Rights Duflo, Rohini Pande, and Petia Topalova. 2009. Promote Empowerment and Child Health in Ne- “Powerful Women: Does Exposure Reduce Bias?” pal?” World Development 35 (11): 1975–88. Quarterly Journal of Economics 124 (4): 1497–540. Anderson, Siwan, and Debraj Ray. 2010. “Missing Beaman, Lori, Esther Duflo, Rohini Pande, and Petia Women: Age and Disease.” Review of Economic Topalova. Forthcoming. “Political Reservation and Studies 77 (4): 1262–300. Substantive Representation: Evidence from Indian Andrabi, Tahir, Jishnu Das, and Asim Ijaz Khwaja. Village Councils.” In India Policy Forum, 2010, ed. 2011. “Students Today, Teachers Tomorrow. Iden- Suman Bery, Barry Bosworth, and Arvind Pana- tifying Constraints on the Provision of Educa- gariya. Brookings Institution Press and the Na- tion.” Policy Research Working Paper Series 5674, tional Council of Applied Economic Research, World Bank, Washington, DC. Washington, DC, and New Delhi. Angrist, Joshua, and Victor Lavy. 2009. “The Effects of Bearinger, Linda H., Renee E. Sieving, Jane Ferguson, High Stakes High School Achievements Awards: and Vinit Sharma. 2007. “Global Perspectives on Evidence from a Randomized Trial.” American the Sexual and Reproductive Health of Adoles- Economic Review 99 (4): 1384–414. cents: Patterns, Prevention, and Potential.” Lancet Attanasio, Orazio, Adriana Kugler, and Costas Meghir. 369 (9568): 1220–31. 2008. “Training Disadvantaged Youth in Latin Berniell, Maria Inés, and Carolina Sánchez-Páramo. America: Evidence from a Randomized Trial.” 2011. “Overview of Time Use Data Used for the Overview 41 Analysis of Gender Differences in Time Use Pat- Deere, Carmen Diana, and Cheryl R. Doss. 2006. terns.” Background paper for the WDR 2012. “Gender and the Distribution of Wealth in Devel- Björkman, Martina, and Jacob Svensson. 2009. “Power oping Countries.” Research Paper Series 2006/115, to the People: Evidence from a Randomized Field United Nations University and World Institute for Experiment on Community-Based Monitoring in Development Economics Research, Helsinki. Uganda.” Quarterly Journal of Economics 124 (2): Deininger, Klaus, Daniel Ayalew Ali, Holden T. Stein, 735–69. and Jaap Zevenbergen. 2008. “Rural Land Certi- Booth, Alison L., and Patrick J. Nolen. 2009. “Gen- fication in Ethiopia: Process, Initial Impact, and der Differences in Risk Behaviour: Does Nurture Implications for Other African Countries.” World Matter?” Centre for Economy Policy Research, Development 36 (10): 1786–812. London. Dercon, Stefan, and Abhijeet Singh. 2011. “From Nu- Bosch, Mariano, and William F. Maloney. 2010. trition to Aspirations and Self-Efficacy: Gender “Comparative Analysis of Labor Market Dynam- Bias over Time among Children in Four Coun- ics Using Markov Processes: An Application to tries.” Oxford University, Oxford, U.K. Processed. ­ Informality.” Labour Economics 17 (4): 621–31. Do, Quy-Toan, Andrei Levchenko, and Claudio Bruhn, Miriam. 2009. “Female-Owned Firms in Latin Raddatz. 2011. “Engendering Trade.” Background America. Characteristics, Performance, and Ob- paper for the WDR 2012. stacles to Growth.” Policy Research Working Paper Doss, Cheryl R. 2006. “The Effects of Intrahousehold Series 5122, World Bank, Washington, DC. Property Ownership on Expenditure Patterns Castro, Roberto, Irene. Casique, and Claire D. Brin- in Ghana.” Journal of African Economies 15 (1): dis. 2008. “Empowerment and Physical Violence 149–80. throughout Women’s Reproductive Life in Mex- Duflo, Esther. 2003. “Grandmothers and Grand- ico.” Violence Against Women 14 (6): 655–77. daughters: Old-Age Pensions and Intrahousehold Chattopadhyay, Raghabendra, and Esther Duflo. Allocation in South Africa.” World Bank Economic 2004. “Women as Policy Makers: Evidence from a Review 17 (1): 1–25. Randomized Policy Experiment in India.” Econo- Dumas, Christelle, and Sylvie Lambert. 2011. “Edu- metrica 72 (5): 1409–43. cational Achievement and Socio-Economic Back- Chioda, Laura, with Rodrigo Garcia-Verdú, and Ana ground: Causality and Mechanisms in Senegal.” María Muñoz Boudet. 2011. Work and Family: Journal of African Economies 20 (1): 1–26. Latin American Women in Search of a New Balance. Esteve-Volart, Berta, and Manuel F. Bagues. 2010. “Are Office of the Chief Economist and Poverty Gender Women Pawns in the Political Game? Evidence Group, LAC. Washington, DC: World Bank. from Elections to the Spanish Senate.” Working Chung, Woojin, and Monica Das Gupta. 2007. “The Paper Series 2009–30, Fundación de Estudios de Decline of Son Preference in South Korea: The Economía Aplicada, Madrid. Roles of Development and Public Policy.” Popula- Farré, Lídia, and Francis Vella. 2007. “The Intergen- tion and Development Review 33 (4): 757–83. erational Transmission of Gender Role Attitudes Coale, Ansley J. 1984. “Rapid Population Change in and Its Implications for Female Labor Force Par- China, 1952–1982.” Report 27, Committee on ticipation.” Discussion Paper Series 2802, Institute Population and Demography, National Academies for the Study of Labor, Bonn. Press, Washington, DC. Fehr, Ernst, Urs Fischbacher, and Simon Gächter. Costa, Rita, and Bob Rijkers. 2011. “Gender and Rural 2002. “Strong Reciprocity, Human Cooperation Non-Farm Entrepreneurship.” Background paper and the Enforcement of Social Norms.” Human for the WDR 2012. Nature 13 (2002): 1–25. Cotlear, Daniel, ed. 2006. “A New Social Contract for Felitti, Vincent D., Robert F. Anda, Dale D. Norden- Peru: An Agenda for Improving Education, Health berg, David F. Williamson, Alison M. Spitz, Valerie Care, and the Social Safety Net.” Country Study, Edwards, Mary P. Koss, and James S. Marks. 1998. World Bank, Washington, DC. “Relationship of Childhood Abuse and House- Croson, Rachel, and Uri Gneezy. 2009. “Gender Dif- hold Dysfunction to Many of the Leading Causes ferences in Preferences.” Journal of Economic Lit- of Death in Adults: The Adverse Childhood Expe- erature 47 (2): 448–74. riences (ACE) Study.” American Journal of Preven- Cuberes, David, and Marc Teignier Baqué. 2011. tive Medicine 14 (4): 245–58. “Gender Inequality and Economic Growth.” Fernández, Raquel, and Alessandra Fogli. 2009. “Cul- Background paper for the WDR 2012. ture: An Empirical Investigation of Beliefs, Work, Das Gupta, Monica. 1987. “Selective Discrimination and Fertility.” American Economic Journal: Macro- against Female Children in Rural Punjab, In- economics 1 (1): 146–77. dia.” Population and Development Review 13 (1): Fiszbein, Ariel, Norbert Schady, Francisco H. G. 77–100. Ferreira, Margaret Grosch, Nial Kelleher, Pedro 42 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Olinto, and Emmanuel Skoufias. 2009. World Bank Hjort, Jonas, Michael Kremer, Isaac Mbiti, and Ed- Policy Research Report: Conditional Cash Transfers: ward Miguel. 2010. “Vocational Education Vouch- Reducing Present and Future Poverty. Washington, ers and Labor Market Returns: A Randomized DC: World Bank. Evaluation among Kenyan Youth.” Harvard Uni- Fogli, Alessandra, and Laura Veldkamp. Forth- versity and Southern Methodist University, Berke- coming. “Nature or Nurture? Learning and the ley, CA. Processed. Geography of Female Labor Force Participation.” Hoddinott, John, and Lawrence Haddad. 1995. “Does Econometrica. Female Income Share Influence Household Ex- FAO (Food and Agriculture Organization). 2003. penditures? Evidence from Côte D’Ivoire.” Oxford “Revisiting the ‘Magic Box.’ ” Case Studies in Lo- Bulletin of Economics and Statistics 57 (1): 77–96. cal Appropriation of Information and Communi- Holzer, Harry J., and David Neumark. 2000. “Assess- cation Technologies, FAO, Rome. ing Affirmative Action.” Journal of Economic Lit- ———. 2011. “The State of Food and Agriculture erature 38 (3): 483–568. 2010–11. Women in Agriculture: Closing the Gen- Hurst, Erik, Chang-Tai Hsieh, Charles Jones, and der Gap for Development.” FAO, Rome. Peter Klenow. 2011. “The Allocation of Talent FAO (Food and Agriculture Organization), IFAD and Economic Growth.” Chicago Booth, Chicago. (International Fund for Agricultural Develop- Processed. ment), and ILO (International Labour Office). ICF Macro. 2010. “Demographic and Health Surveys.” 2010. “Gender Dimensions of Agricultural and Measure DHS, ICF Macro, Calverton, MD. Rural Employment: Differentiated Pathways out ILO (International Labour Organization). 2010. “Key of Poverty. Status, Trends and Gaps.” FAO, IFAD, Indicators of the Labour Market.” ILO, Geneva. and ILO, Rome. ICRW (International Center for Research on Gage, Anastasia J. 2005. “Women’s Experience of In- Women). 2006. Property Ownership & Inheritance timate Partner Violence in Haiti.” Social Science & Rights of Women for Social Protection: The South Medicine 61 (2): 343–64. Asia Experience. Washington, DC: ICRW. Gajigo, Ousman, and Mary Hallward-Driemeier. Iyer, Lakshmi, Anandi Mani, Prachi Mishra, and 2011. “Constraints and Opportunities for New Petia Topalova. 2010. “Political Representation Entrepreneurs in Africa.” World Bank, Washing- and Crime: Evidence from India’s Panchayati Raj.” ton, DC. Processed. International Monetary Fund, Washington, DC. Gilbert, Robert A., Webster D. Sakala, and Todd D. Processed. Benson. 2002. “Gender Analysis of a Nationwide Jensen, Robert. 2010. “Economic Opportunities and Cropping System Trial Survey in Malawi.” African Gender Differences in Human Capital: Experi- Studies Quarterly 6 (1). Gilliam, Melissa L. 2010. “Interventions for Prevent- mental Evidence for India.” Working Paper Series ing Unintended Pregnancies among Adolescents.” 16021, National Bureau of Economic Research, Obstetrics and Gynecology 115 (1): 171–72. Cambridge, MA. Gneezy, Uri, Kenneth L. Leonard, and John A. List. Jeyaseelan, L., Shuba Kumar, Nithya Neelakantan, 2009. “Gender Differences in Competition: Evi- Abraham Peedicayil, Rajamohanam Pillai, and dence from a Matrilineal and a Patriarchal Soci- Nata Duvvury. 2007. “Physical Spousal Violence ety.” Econometrica 77 (5): 1637–64. against Women in India: Some Risk Factors.” Jour- Gornick, Janet C., and Jerry A. Jacobs. 1998. “Gen- nal of Biosocial Science 39 (5): 657–70. der, the Welfare State, and Public Employment: Kabeer, Nalia. 1996. “Agency, Well-Being & Inequal- A Comparative Study of Seven Industrialized ity: Reflections on the Gender Dimensions of Pov- Countries.” American Sociological Review 63 (5): erty.” IDS Bulletin 27 (1): 11–21. 688–710. ———. 1999. “Resources, Agency, Achievements: Haddad, Lawrence, John Hoddinott, and Harold Al- Reflections on the Measurement of Women’s derman. 1997. Intrahousehold Resource Allocation Empowerment.” Development and Change 30 (3): in Developing Countries: Models, Methods, and 35–64. Policy. Baltimore: Johns Hopkins University. Katz, Elizabeth, and Juan Sebastian Chamorro. 2003. Hallward-Driemeier, Mary. 2011. “Improving the “Gender, Land Rights, and the Household Econ- Legal Investment Climate for Women in Sub- omy in Rural Nicaragua and Honduras.” Paper Saharan Africa.” World Bank, Washington, DC. presented at the Annual Conference of the Latin Hindin, Michelle J., Sunita Kishor, and Donna L. American and Caribbean Economics Association, Ansara. 2008. “Intimate Partner Violence among Puebla, Mexico, October 9. Couples in 10 DHS Countries: Predictors and Kinkingninhoun-Mêdagbé, Florent M., Aliou Diagne, Health Outcomes.” DHS Analytical Studies 18, Franklin Simtowe, Afiavi R. Agboh-Noameshie, U.S. Agency for Iinternational Development, and Patrice Y. Adégbola. 2010. “Gender Discrimi- Washington, DC. nation and Its Impact on Income, Productivity, Overview 43 and Technical Efficiency: Evidence from Benin.” Nguyen, Trang. 2008. “Information, Role Models Agriculture and Human Values 27 (1): 57–69. and Perceived Returns to Education: Experimen- Kishor, Sunita, and Kiersten Johnson. 2004. Profiling tal Evidence from Madagascar.” Working Paper, Domestic Violence: A Multi-Country Study. Calver- Massachusetts Institute of Technology, Cam- ton, MD: ORC Macro. bridge, MA. Koenig, Michael A., Rob Stephenson, Saifuddin Ñopo, Hugo, Miguel Robles, and Jaime Saavedra. Ahmed, Shireen J. Jejeebhoy, and Jacquelyn 2007. “Occupational Training to Reduce Gender Campbell. 2006. “Individual and Contextual De- Segregation: The Impacts of ProJoven.” Working terminants of Domestic Violence in Northern Paper Series 623, Inter-American Development India.” American Journal of Public Health 96 (1): Bank Research Department, Washington, DC. 132–38. Nyamu-Musembi, Celestine. 2002. “Are Local Norms Kremer, Michael, Edward Miguel, and Rebecca and Processes Fences or Pathways? The Example Thornton. 2009. “Incentives to Learn.” Review of of Women’s Property Rights.” In Cultural Trans- Economics and Statistics 91 (3): 437–56. formations and Human Rights in Africa, ed. Ab- Leonard, Jonathan S. 1989. “Women and Affirmative dullahi A. An-Na’im. London: Zed Books. Action.” Journal of Economic Perspectives 3 (1): Oladeebo, J. O., and A. A. Fajuyigbe. 2007. “Technical 61–75. Efficiency of Men and Women Upland Rice Farm- Lewis, Maureen A., and Marlaine E. Lockheed. 2006. ers in Osun State, Nigeria.” Journal of Human Ecol- Inexcusable Absence: Why 60 Million Girls Aren’t in ogy 22 (2): 93–100. School and What to Do About It. Washington, DC: OECD (Organisation for Economic Cooperation and Center for Global Development. Development). 1993. Private Pay for Public Work. Lim, Stephen S., Lalit Dandona, Joseph A. Hoisington, Performance-Related Pay for Public Sector Manag- Spencer L. James, Margaret C. Hogan, and Em- ers. Paris: OECD. manuela Gakidou. 2010. “India’s Janani Suraksha Pathmanathan, Indra, Jerker Liljestrand, Jo M. Mar- Yojana, A Conditional Cash Transfer Programme tins, Lalini C. Rajapaksa, Craig Lissner, Amalia to Increase Births in Health Facilities: An Impact de Silva, Swarna Selvaraju, and Prabha Joginder Evaluation.” Lancet 375 (9730): 2009–23. Singh. 2003. “Investing in Maternal Health Learn- Luke, Nancy, and Kaivan Munshi. 2011. “Women as ing from Malaysia and Sri Lanka.” World Bank, Agents of Change: Female Income and Mobility Washington, DC. in India.” Journal of Development Economics 94 Prata, Ndola, Paige Passano, Amita Sreenivas, and (1): 1–17. Caitlin Elisabeth Gerdts. 2010. “Maternal Mortal- Lundberg, Shelly J., Robert A. Pollak, and Terence J. ity in Developing Countries: Challenges in Scaling Wales. 1997. “Do Husbands and Wives Pool Their Resources? Evidence from the United Kingdom Up Priority Interventions.” Women’s Health 6 (2): Child Benefit.” Journal of Human Resources 32 (3): 311–27. 463–80. Pronyk, Paul M., James R. Hargreaves, Julia C. Kim, Martin, Sandra L., Kathryn E. Moracco, Julian Garro, Linda A. Morison, Godfrey Phetla, Charlotte Amy Ong Tsui, Lawrence L. Kupper, Jennifer L. Watts, Joanna Busza, and John D. H. Porter. 2006. Chase, and Jacquelyn C. Campbell. 2002. “Do- “Effect of a Structural Intervention for the Pre- mestic Violence across Generations: Findings vention of Intimate-partner Violence and HIV in from Northern India.” International Journal of Rural South Africa: A Cluster Randomized Trial.” Epidemiology 31 (3): 560–72. Lancet 2368 (9551): 1973–83. Martinez, Sebastian, and others. 2011. “Hard Skills Qian, Nancy. 2008. “Missing Women and the Price of or Soft Skills.” Presentation to the World Bank, Tea in China: The Effect of Sex-Specific Earnings Washington, DC. on Sex Imbalance.” Quarterly Journal of Economics McEwen, Bruce S. 1999. “Stress and Hippocampal 123 (3): 1251–85. Plasticity.” Annual Review of Neuroscience 22 (1): Quisumbing, Agnes R., and John A. Maluccio. 2000. 105–22. “Intrahousehold Allocation and Gender Rela- Miller, Grant. 2008. “Women’s Suffrage, Political tions: New Empirical Evidence from Four De- Responsiveness, and Child Survival in American veloping Countries.” Discussion Paper 84, Food History.” Quarterly Journal of Economics 123 (3): Consumption and Nutrition Division, Interna- 1287–327. tional Food Policy Research Institute, Washing- Moock, Peter R. 1976. “The Efficiency of Women as ton, DC. Farm Managers: Kenya.” American Journal of Agri- Rawlings, Laura, B., and Gloria M. Rubio. 2003. “Eval- cultural Economics 58 (5): 831–5. uating the Impact of Conditional Cash Transfer Natarajan, Mangai. 2005. “Status of Women Police in Programs: Lessons from Latin America.” Policy Asia: An Agenda for Future Research.” Journal for Research Working Paper Series 3119, World Bank, Women and Policing 17: 45–47. Washington, DC. 44 WO R L D D E V E LO P M E N T R E P O RT 2 0 1 2 Reed, Elizabeth, Anita Raj, Elizabeth Miller, and Husbands Make a Difference?” American Journal Jay G. Silverman. 2010. “Losing the ‘Gender’ in of Public Health 83 (11): 1567–71. Gender-Based Violence: The Missteps of Research Thomas, Duncan. 1990. “Intra-Household Resource on Dating and Intimate Partner Violence.” Vio- Allocation: An Inferential Approach.” Journal of lence Against Women 16 (3): 348–54. Human Resources 25 (4): 635–64. Ross, David A., John Changalucha, Angela I. N. Obasi, Thomas, Duncan, John Strauss, and Maria-Helena Jim Todd, Mary L. Plummer, Bernadette Cleophas- Henriques. 1990. “Child Survival, Height for Age, Mazige, Alessandra Anemona, Dean Everett, and Household Characteristics in Brazil.” Journal Helen A. Weiss, David C. Mabey, Heiner of Development Economics 33 (2): 197–234. Grosskurth, and Richard J. Hayes. 2007. “Biologi- Udry, Christopher. 1996. “Gender, Agricultural Pro- cal and Behavioural Impact of an Adolescent Sex- duction, and the Theory of the Household.” Jour- ual Health Intervention in Tanzania: A Commu- nal of Political Economy 104 (5): 1010–46. nity-Randomized Trial.” AIDS 21 (14): 1943–55. United Nations Department of Economic and Social Rubalcava, Luis, Graciela Teruel, and Duncan Affairs. 2009, “World Population Prospects 2009.” Thomas. 2009. “Investments, Time Preferences, United Nations, New York. and Public Transfers Paid to Women.” Economic ———. 2010. “The World’s Women 2010: Trends Development and Cultural Change 57 (3): 507–38. and Statistics.” United Nations, New York. Sabarwal, Shwetlena, Katherine Terrell, and Elena Vargas Hill, Ruth, and Marcella Vigneri. 2009. “Main- Bardasi. 2009. “How Do Female Entrepreneurs streaming Gender Sensitivity in Cash Crop Perform? Evidence from Three Developing Re- Markets Supply Chains.” International Food Pol- gions.” World Bank, Washington, DC. Processed. icy Research Institute, Washington, DC. Saito, Katrine A., Hailu Mekonnen, and Daphne Waddington, Hugh, Birte Snilstveit, Howard White, Spurling. 1994. “Raising the Productivity of and Lorna Fewtrell. 2009. “Water, Sanitation Women Farmers in Sub-Saharan Africa.” Africa and Hygiene Interventions to Combat Childhood Technical Department Discussion Paper Series Diarrhoea in Developing Countries.” Interna- 230, World Bank, Washington, DC. tional Initiative for Impact Evaluation 31E, Syn- Schady, Norbert, and Maria Caridad Araujo. 2006. thetic Review, New Delhi, London, and Washing- “Cash Transfers, Conditions, School Enrollment, ton, DC. and Child Work: Evidence from a Randomized WHO (World Health Organization). 2005. “WHO Experiment in Ecuador.” Policy Research Working Multi-country Study on Women’s Health and Paper Series 3930, World Bank, Washington, DC. ­ Domestic Violence against Women: Initial Results Schady, Norbert, and José Rosero. 2008. “Are Cash on Prevalence, Health Outcomes and Women’s Transfers Made to Women Spent Like Other Responses.” WHO, Geneva. Sources of Income?” Economics Letters 101 (3): WHO, UNICEF (United Nations Children Fund), 246–48. UNFPA (United Nations Population Fund), and Schmidt, Manfred G. 1993. “Gendered Labour Force World Bank. 2010. “Trends in Maternal Mortality: Participation.” In Families of Nations: Patterns of 1990 to 2008.” WHO, Washington, DC. Public Policy in Western Democracies, ed. Frances World Bank. 2001. “Engendering Development— G. Castles. Dartmouth Publishing Company, Al- Through Gender Equality, Resources, and Voice. dershot, U.K., and Brookfield, VT. Policy Research Report, World Bank, Washington, Sen, Amartya. 1990. “Gender and Cooperative DC.” Conflict.” In Persistent Inequalities: Women and ­ ———-. 2005. World Development Report 2006: Eq- Development, ed. Irene Tinker. Oxford, U.K.: Ox- uity and Development. New York: Oxford Univer- ford University Press. sity Press. ———. 1992. “Missing Women.” British Medical ———. 2008. Growth Report. Strategies for Sustained Journal 304: (6827): 587–8. Growth and Inclusive Development. Washington, ———. 1999. Development as Freedom. New York: DC: World Bank. Knopf. ———. 2011. “Defining Gender in the 21st Century: Swaminathan, Hema, Cherryl Walker, and Margaret Talking with Women and Men around the World, A. Rugadya, eds. 2008. Women’s Property Rights, A Multi-Country Qualitative Study of Gender HIV and AIDS, and Domestic Violence: Research and Economic Choice.” World Bank, Washington, Findings from Two Rural Districts in South Africa DC. and Uganda. Cape Town: HSRC Press. Yount, Kathryn M., and Jennifer S. Carrera. 2006. Terefe, Almas, and Charles P. Larson. 1993. “Modern “Domestic Violence against Married Women in Contraception Use in Ethiopia: Does Involving Cambodia.” Social Forces 85 (1): 355–87. Contents of the World Development Report 2012 Foreword Acknowledgments Abbreviations and data notes Overview Introduction Part I  Taking stock of gender equality 1 A wave of progress 2 The persistence of gender inequality Spread 1  Women’s pathways to empowerment: Do all roads lead to Rome? Part II  What has driven progress? What impedes it? 3  Education and health: Where do gender differences really matter? 4 Promoting women’s agency Spread 2  The decline of the breadwinner: Men in the 21st century 5 Gender differences in employment and why they matter 6 Globalization’s impact on gender equality: What’s happened and what’s needed Spread 3  Changing ages, changing bodies, changing times—Adolescent boys and girls Part III  The role of and potential for public action 7 Public action for gender equality 8 The political economy of gender reform 9 A global agenda for greater gender equality Bibliographical Note   Background Papers and Notes Selected Indicators   Selected World Development Indicators   Index   46 Four easy ways to order PHONE: MAIL: ONLINE: FAX: +1-703-661-1580 or P.O. Box 960 www.worldbank.org/publications +1-703-661-1501 1-800-645-7247 Herndon, VA 20172-0960, USA World Development Report 2012 PRICE QTY TOTAL Gender Equality and Development US$26.00 Paperback: (ISBN: 978-0-8213-8810-5) SKU 18810 Hardcover: (ISBN: 978-0-8213-8825-9) SKU 18825 US$50.00 Prices vary by country as World Bank Publications offers geographical Subtotal discounts on its titles. Please visit publications.worldbank.org/discounts Within the US (prepaid orders): $8 per order + $1 per item. Geographic discount* Outside of the US: • Nontrackable airmail delivery (US$7 per order + US$6 per item). Delivery time: 4-6 weeks Shipping and Handling** • Trackable couriered airmail delivery (US$20 per order + US$8 per item). Delivery time: 2 weeks. Total US$ MAILING ADDRESS METHOD OF PAYMENT Name Charge my Organization Visa Mastercard American Express Address Credit card number City Expiration date State Zip Name Country Signature Phone Enclosed is my check in US$ drawn on a U.S. bank and Fax made payable to the World Bank Email Customers outside the United States Contact your local distributor for information on prices in local currency and payment terms http://publications.worldbank.org/booksellers THANK YOU FOR YOUR ORDER! ECO-AUDIT Environmental Benefits Statement The World Bank is committed to preserving Saved: endangered forests and natural resources. • 49 trees The Office of the Publisher has chosen to •20 million Btu of total print World Development Report 2012: energy Overview on recycled paper with 50 per- •5,052 pounds of net cent postconsumer fiber in accordance greenhouse gases with the recommended standards for paper • 22,779 gallons of usage set by the Green Press Initiative, a waste water nonprofit program supporting publishers • 1,444 pounds of in using fiber that is not sourced from solid waste dangered forests. For more information, en­ visit www.greenpressinitiative.org. The lives of women around the world have improved dramatically, at a pace and scope difficult to imagine even 25 years ago. Women have made unprecedented gains in rights, education, health, and access to jobs and livelihoods. More countries than ever guarantee equal rights in property, marriage, and other domains. Gender gaps in primary schooling have closed in many countries, while in a third of all countries girls now outnumber boys in secondary school. And more young women than men attend universities in 60 countries. Women are using their education to participate more in the labor force: they now make up for 40 percent of the global labor force and 43 percent of its farmers. Moreover, women now live longer than men in every region of the world. Despite the progress, gaps remain in many areas. Women are more likely to die—relative to males—in many low- and middle-income countries than their counterparts in rich countries—especially in childhood and during their reproductive years. Primary and secondary school enrollments for girls remain much lower than for boys in many Sub-Saharan African countries and some parts of South Asia, as well as among disadvantaged populations. Women are more likely than men to work as unpaid family laborers or in the informal sector, to farm smaller plots and grow less profitable crops, operate in smaller firms and less profitable sectors, and generally earn less. Women—especially poor women—have less say over decisions and less control over household resources. And in most countries, fewer women participate in formal politics than men and are underrepresented in the upper echelons. The World Development Report 2012: Gender Equality and Development argues that closing these persistent gender gaps matters. It matters because gender equality is a core development objective in its own right. But it is also smart economics. Greater gender equality can enhance productivity, improve development outcomes for the next generation, and make institutions more representative. Building on a growing body of knowledge on the economics of gender equality and development, the Report identifies the areas where gender gaps are most significant—both intrinsically and in terms of their potential development payoff—and where growth alone cannot solve the issues. It then sets forth four priorities for public action: • Reducing excess female mortality and closing education gaps where they remain • Improving access to economic opportunities for women • Increasing women’s voice and agency in the household and in society • Limiting the reproduction of gender inequality across generations. Policies need to focus on the underlying determinants of gender gaps in each priority area. In some priority areas—such as excess female mortality in infancy and early childhood as well as in the reproductive years— improving service delivery (especially of clean water and sanitation, and maternal care) is of primary importance. For other priority areas—such as gender gaps in earnings and productivity—policies need to tackle the multiple constraints that originate in the workings of markets and institutions to limit progress. Policy makers will need to prioritize these constraints and address them simultaneously or sequentially. While domestic policies are central to reducing gender inequalities, development partners should focus on complementing these efforts in each of the four priority areas, and on supporting evidence-based public action through better data, evaluation, and learning. This will require a mix of more funding, efforts to foster innovation and learning, and broader partnerships. The funding should be directed particularly to supporting the poorest countries in reducing excess female mortality and gender gaps in education. Investments are needed to improve the availability of better gender-disaggregated data and support more experimentation and systematic evaluation. And the partnerships should involve the private sector, development agencies, and civil society organizations. SKU 32604