WPS8245 Policy Research Working Paper 8245 Gender and Youth Employment in Sub-Saharan Africa A Review of Constraints and Effective Interventions Shubha Chakravarty Smita Das Julia Vaillant Gender Cross Cutting Solution Area November 2017 Policy Research Working Paper 8245 Abstract Although the ratio of female to male labor force participa- above and beyond the constraints faced by male youth. The tion rates is higher in Sub-Saharan Africa than in any other paper synthesizes the emerging lessons from a growing evi- region, these high rates of female labor force participation dence base on interventions that aim to support young mask underlying challenges for women. A large majority women’s employment, and identifies knowledge gaps and of employed women work in vulnerable employment. In priority research questions for the future. The objective is to addition, youth unemployment rates in Sub-Saharan Africa develop a gender-informed policy and research agenda on are double those of adult unemployment, and unemploy- youth employment that can guide practitioners, develop- ment rates for women are higher than rates faced by men. ment partners, and researchers who seek to advance young This paper discusses the specific barriers that youth face women’s empowerment and employment in the context in accessing employment in Sub-Saharan Africa, and the of youth employment programming and policy making. ways in which young women’s employment is constrained This paper is a product of the Gender Cross Cutting Solution Area. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at schakravarty@ worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Gender and Youth Employment in Sub-Saharan Africa: A Review of Constraints and Effective Interventions* Shubha Chakravarty, Smita Das, and Julia Vaillant Keywords: Gender, employment, labor market, skills. JEL: J16, J13, J24, O12 * This document is part of a series of technical review papers for the World Bank’s Africa Gender Innovation Lab. The views presented in this paper are those of the author and do not represent those of the World Bank. The authors thank Maria Cecilia Acevedo and Sarah Haddock for their valuable contributions and Markus Goldstein, Louise Fox, Indhira Santos, Elizabeth King, and colleagues in AFRGI, AFRCE, and GTGDR for detailed comments and suggestions. The authors gratefully acknowledge funding from the World Bank Group’s Umbrella Facility for Gender Equality. Contents 1. Introduction ............................................................................................................................................. 1 2. How and why are female youth constrained in the labor market? .................................................... 7 3. Evidence on interventions addressing constraints to young women’s employment ....................... 18 4. Research Agenda ................................................................................................................................... 30 5. References .............................................................................................................................................. 40 1. Introduction Africa is in the midst of massive changes in population, development, and growth. Mortality has declined substantially in the past 20 years,1 resulting in a pyramid-shaped age structure in which half of the population is under the age of 25 (see Error! Reference source not found.). Estimates indicate that 11 million youth will be entering the labor market in Sub-Saharan Africa each year for the coming decade, with the potential to dramatically promote development and reduce poverty. But to make the most of this historic development opportunity, young people need to engage in productive employment that fuels economic growth. Investing in youth also offers important spillovers: because youth have a longer working life ahead of them than older adults, improving labor prospects early can affect youth’s investments in their children and reduce the inter-generational transfer of poverty. Young women’s employment contributes additionally to economic development, through its potential effect on agency and Figure 1: The demographic structure of Africa is younger than other regions. Source: Filmer and Fox, 2014. fertility. This paper discusses the specific barriers that youth face in accessing employment, and further, the ways in which young women’s employment is constrained above and beyond those constraints faced by male youth.2 Youth unemployment rates in Sub-Saharan Africa are double those of adult unemployment for both men and women (ILO, 2013a), owing to two main challenges. First, the school-to-work transition is slow in Sub-Saharan Africa, meaning that after leaving school, young people struggle to enter the labor market. A school-to-work survey (SWTS) of eight countries3 in Africa found an average youth labor 1 Between 1993 and 2013 in Sub-Saharan Africa, the under-five mortality rate fell 49.4 percent, the adult female mortality rate fell 12.4 percent and the adult male mortality rate fell 14.4 percent (World Bank, 2015). 2 Although not the focus of this paper or GIL’s work, young men also experience constraints to employment (see Box 1). 3 The International Labour Organization conducted national School-to-Work Transition Surveys (SWTS) of men and women, ages 15 to 29, in nine countries in Sub-Saharan Africa between 2012 and 2015: Benin, Liberia, Madagascar, Malawi, Tanzania, Togo, Uganda, and Zambia. The Republic of Congo is included in aggregates for later papers (Elder and Kring, 2016). 1 underutilization rate of 61.1 percent for men and 72.8 percent for women (Elder & Kone, 2014).4 Second, much of the employment in which youth engage yields very limited earnings. This paper concentrates primarily on the first theme, by exploring the barriers to productive employment for young women, exploring individual and household constraints to labor market entry and earnings in informal and wage work. The World Bank’s Africa Gender Innovation Lab (GIL) white papers on entrepreneurship and agriculture expand on the barriers to productivity and earnings for those women (the majority) who end up in self-employment and agriculture, respectively. The importance of promoting youth employment has prompted massive investment in Sub-Saharan Africa in the past decade, but few programs have emphasized the importance of targeting young women. The GIL has led research to build the global evidence base on supporting young women’s employment for several years. This paper synthesizes the emerging lessons from this growing evidence base (from GIL and others) and places them in a larger framework of what we know about the constraints to young women’s employment and what we know about how to alleviate those constraints. The objective is to develop a gender-informed policy and research agenda on youth employment that can guide GIL’s work as well as policy makers, practitioners, development partners, and researchers who seek to advance young women’s empowerment and employment in the context of youth employment programming and policy making. Youth and Women’s Employment in Sub-Saharan Africa In Sub-Saharan Africa, employment in general and women’s employment in particular exhibit somewhat different patterns relative to other regions of the world. In this region, youth leave school earlier than in other regions and join the workforce at an earlier age. The ILO’s SWTS found that among those aged 15 to 17, 16.7 percent were out of school and working, and another 10.0 percent were not in education, employment, or training (ILO, 2015). In 2014, among those aged 15 to 24, 51.0 percent of women and 55.8 percent of men were part of the labor force (World Bank, 2015). While employment rates are among the highest in the world, at 46.8 percent in the region for those aged 15 to 24, they are not necessarily indicative of productivity. Because many countries in the region have yet to undergo a structural transformation into an industrialized economy, employment is concentrated in smallholder agriculture and informal self-employment, including household enterprises. Sub-Saharan Africa’s employment has the largest concentration in agriculture in the world at 62.0 percent, with South Asia following at 46.3 percent. Only 8.4 percent and 29.6 percent of Sub-Saharan Africa’s employment are in industry and services respectively (ILO, 2013a). Hence, in this paper, we define employment broadly to include any type of income-generating activity, including informal work. Wage employment, particularly in the formal sector, is the exception rather than the rule: while wage and salaried work accounts for close to half of global employment, it only accounts for 21.4 percent of employment in Sub-Saharan Africa (ILO, 2013b). In a survey of youth in eight countries in the region, 71.7 percent of working youth were engaged in self-employment. Moreover, women were more likely 4 The measure of underutilization includes those in irregular employment and those who are neither employed nor in school. 2 than men to be self-employed or family workers (61% for men and 74% for women), and less likely to be employees or employers (31.6% for men and 20.6% for women) (ILO, 2014). Employment by family members and self-employment is classified by the ILO as “vulnerable employment” due to the informal nature of these occupations. As depicted in Figure 2, a large majority of employed women work in vulnerable employment. Figure 2: Vulnerable Employment, Agriculture, and Unpaid Work by women in select countries 100 80 80 60 60 40 40 20 20 0 0 Kenya Nigeria Cameroon Madagascar Ghana Mozambique DRC Uganda Ethiopia Cote d'Ivoire Tanzania Nigeria Kenya Cameroon Madagascar Ghana DRC Mozambique Cote d'Ivoire Ethiopia Uganda Tanzania Employed by non-family member Percent of Women in Agriculture Self-employed Percent of Women Unpaid within Agricultural Work Employed by family member Percent of Women Unpaid within Non-Agr. Work Source: Demographic and Health Survey Data 2008-2014; includes data on working women ages 15 to 49. Note: The most populous countries with available data were selected. Why Young Women’s Employment Matters for Development In 2013, the ratio of female-to-male labor force participation rates was higher in Sub-Saharan Africa than in any other region at 0.84 (World Bank, 2015). However, these high rates of female labor force participation mask underlying challenges. In the same year, the unemployment rate for women aged 15 to 24 was 1.26 times the rate faced by men. Regional variation is quite large: the ratio of the female-to-male employment rate for young people (ages 15 to 24) varies widely across the region, from .44 in Mauritania to 1.24 in Sierra Leone (See Figure 3). While the following discussion is regional in nature, it is important to note that heterogeneity in this ratio varies with country, ethnicity, culture, and urbanization, indicating that policies that do not work under certain circumstances may prove successful in others. The differences in the constraints facing young women stem from differential access to key resources which enable employment, such as skills, time, and capital, as well as (and related to) underlying norms and institutions that govern men’s and women’s economic and household roles. Despite the relatively higher barriers faced by women, young men tend to dominate the policy dialogue on youth employment, motivated in part by the perceived threat of instability and socially disruptive behavior that could result from their unemployment. The resulting policies are often blind to young women’s needs and constraints. 3 Achieving gender equality in employment may require specialized interventions to overcome the constraints that bind more severely on young women than young men. Without tailoring these policies accordingly, their potential impact on economic development is weakened. The employment of young women can improve Figure 3: Ratio of female employment development directly through economic growth and ratio to male employment ratio (Ages 15- productivity, and indirectly via increased agency and lower 24) fertility. By engaging in economically productive activities, young women who have time and want to work contribute directly to the economy. Similarly, the misallocation of high-ability women into low-return occupations may reduce economic growth, as has been shown in the United States and India (Hsieh et al., 2013; Esteve-Volart, 2004), and inhibit investment in women’s human capital. One simulation estimates that reducing gender inequality in education and the labor market would increase annual GDP growth by 1.9 percentage points in Ethiopia (World Bank, 2009). However, the relationship between female labor force participation (FLFP) and economic growth is not straightforward. Cross-country evidence suggests that FLFP and growth exhibit a U-shaped pattern: FLFP is highest when (1) per capita GDP is low, and women are Source: World Bank, 2015. working due to necessity, and (2) per capita GDP is high, Note: Categories determined by quartiles. and women are working out of desire, but FLFP often drops at intermediate income levels (Goldin, 1994). While FLFP in Sub-Saharan Africa is currently high, some women may leave the labor force as household incomes improve, unless high-quality job opportunities are made available to women. Young women’s employment promotes development not only through economic growth but also by increasing women’s voice and agency,5 leading to positive spillovers at the individual, family, and societal levels (Klugman et al., 2014). Young women make decisions on marriage, childbearing, investment in education, joining the labor force, occupation, and industry, which have a persistent impact on their lifetime earnings trajectories. Increased income and control over assets can increase investments in a woman’s education, health, and business, which positively feed back into economic productivity (Baird et al., 2014). Economic theory suggests that increased income for women also positively benefits their children, primarily through higher human capital expenditures (Doepke and Tertilt, 2014), and serves as a mechanism to reduce the intergenerational transmission of poverty. Evidence from outside Africa shows that when women have more power relative to men (using relative education levels as a proxy), school enrollment increases among children (Gitter and Barham, 2008). Finally, women’s employment contributes to women’s empowerment in other realms, increasing their power and influence in society, with the long-term potential to accelerate the evolution of those same norms and institutions 5 We define agency as the increased capacity to make decisions and act on them, thereby exerting more control over one’s life. 4 which constrain young women’s employment today. Evidence from India shows that the presence of women in political leadership positions significantly improves the extent to which women constituents’ needs are represented, as well as young women’s career aspirations, time spent on household chores, and human capital investments (Chattopadhay and Duflo, 2004; Beaman et al., 2012). Another potential channel through which young women’s employment can indirectly promote development is by lowering fertility (Bloom et al., 2009). Africa has not yet completed its demographic transition, a phenomenon where declining mortality and fertility shift the population structure of a country or region. These demographic shifts can create a temporary population bulge (see Figure 1), reflecting a lower ratio of dependents to working age adults. Other regions that have experienced this transition have benefited from a subsequent boost in economic growth, known as the demographic dividend, resulting from this lowered dependency ratio. In Africa, while mortality rates have fallen, fertility rates remain persistently high in many countries (Guengant and May, 2013). The expansion of young women’s labor market opportunities can affect family formation, including fertility, through at least three theoretical channels. First, the ability to contribute economically expands the role of women in the household and society and can alleviate social and familial pressure for early marriage and fertility.6 Second, the loss in earnings associated with childrearing represents an opportunity cost that may increase young women’s desire to delay marriage and childbearing. Third, higher earnings may improve a woman’s bargaining power within the household and allow her to negotiate delays in sexual debut or marriage as well as the use of contraceptives.7 To the extent that these channels can be empirically demonstrated in Africa, promoting young women’s employment could deliver a double- barreled impact on growth: directly through their productive contribution to the economy and indirectly by lowering fertility and hence accelerating the demographic transition. 6 For example, Corno and Voena (2015) show that households in Tanzania have their daughters married early to obtain bride price payments in response to adverse income shocks. 7 Sivasankaran (2014) demonstrates empirical evidence for this channel in India, where increased periods of employment in textile factories led women to delay marriage and fertility, through increased household bargaining power. 5 Box 1. Addressing unemployment among young men In Africa, young women are more disadvantaged than young men in the labor market due to norms, institutions, and differences in resource endowments necessary for accessing income-generating opportunities. However, in fragile and conflict-affected states, governments often attempt to promote stability by targeting at-risk young men and addressing the drivers of conflict and violence.a Cross- country studies and within-country studies from outside Africa demonstrate the link between income shocks and crime, but no such link has been demonstrated for terrorism and the existence of conflict, and additional individual-level data are needed. Though limited evidence is available on which kinds of violence can be addressed by income-oriented programs, some information is available on which programs are most effective at changing attitudes towards violence and reducing time spent on illicit activities. Skills development and microfinance programs have generally been ineffective on their own (Blattman and Ralston, 2015), though one study suggests that agricultural skills programs may be more effective at changing incentives in fragile agrarian states (Blattman and Annan, 2015). However, when combined with the promise of future capital (suggesting the importance of periodic payments) conditional on the male beneficiary’s location, training programs shift hours away from illicit activities. Income-oriented programs and psychosocial programs may reinforce each other to produce more long-term effects: while one study from Liberia found that training, inputs, and counseling did not produce a change in attitudes towards violence and social integration (Blattman and Annan, 2015), another found that cognitive behavior therapy (focused on self-control and building a noncriminal self-image) changed attitudes and reduced violence, when combined with cash transfers (Blattman, Jamison, Sheridan, 2015). a Recognizing that engaging men is crucial in conflict and post-conflict settings, other teams within the World Bank have worked specifically on this topic. For example, the Learning on Gender and Conflict in Africa Program (LOGICA) was an initiative that aimed to generate knowledge and good practice on addressing gender and conflict issues, and funded several studies on at-risk young men in Sub-Saharan Africa. 6 2. How and why are female youth constrained in the labor market? This section discusses the constraints Figure 4: Framework for constraints to women’s economic faced by young women as they empowerment in Africa navigate their entry into the labor market in Sub-Saharan Africa. Figure 4 presents a framework for constraints to young women’s employment. In this framework, women’s choices and the resulting outcomes are constrained by differences in men’s and women’s endowments of key resources, such as skills and time, as well as different preferences with regard to time use, occupation and discounting the future. These endowments and preferences themselves are shaped by contextual factors including social norms and institutions that govern the roles of men and women in the household, economy, and society.8 The interaction between endowments, preferences and contextual factors is dynamic, in that changes in the distribution of endowments and preferences can feed back onto norms and institutions, resulting in a continuously evolving set of resources and constraints facing women. For example, increases in education and skills levels may improve a woman’s agency, and thus her ability and interest in pursuing traditionally male-dominated higher-earning occupations, which may alter expectations on returns to skills. The dynamic nature of these constraints implies that policies targeting one constraint may have low impact for women unless they account for the larger context of constraints. In applying this conceptual framework to youth employment, we focus our analysis on six main constraints that affect both women and youth: skills, capital, networks, time and family formation, occupational choice, and safety. A discussion of the surrounding social norms is interwoven into the analysis of each constraint. Our approach posits that being young and being female both serve as disadvantages in the labor market, and that these disadvantages are additive. For each constraint, we first describe whether and how youth are more constrained than adults and then focus on whether and how the constraint affects young women and men differently. (See Table 1 for a summary of constraints.) The constraints presented earlier in the section tend to be applicable to all youth, while those presented later are particularly applicable to young women. 8In this framework, gender norms and institutions are defined as the commonly held beliefs and expectations about the characteristics of men and women and how they are expected to behave, the value ascribed to these roles, and the formal and informal institutions that enshrine, codify and reinforce or challenge these expectations. These norms and institutions influence men and women’s voice, agency, and decision-making in the personal, household, economic, social, and political spheres. They apply to, inter alia: domestic production, mobility, reproductive control, safety and gender-based violence, and occupational choice. 7 2.1. Skills A low level of marketable skills among youth is widely recognized as a major barrier to youth employment in Africa today. After completing primary or even secondary school, youth lack the foundational and technical skills that employers need. According to the World Bank Enterprise Survey, 18.2% percent9 of firms in the region identify an inadequately educated workforce as a major constraint to operations (World Bank, 2016). The literacy rate among youth (ages 15-24) in Sub-Saharan Africa was 65.8 percent and 76.3 percent in 2010 for women and men, respectively, amid global literacy rates of 88.5 percent and 92.5 percent (World Bank, 2015). Young women attain less education on average than young men (Figure 5), putting them at a further disadvantage. In the ILO SWTS carried out across eight countries (see footnote 3), Figure 5: Women attain less education on average than men. young men and women cite economic reasons as the main 1.2 reasons for having no Total Years of Schooling (All) Schooling (Female)/Average 1 Average Total Years of schooling. However, 28.5 0.8 percent of young women cite 0.6 parental refusal compared to 0.4 17.2 percent of young men, and 0.2 2.1 percent of young women cite marriage compared to 0.4 0 percent of young men (Elder and Kone, 2014). As a result of these constraints to formal education, youth Source: World Bank Databank (Barro and Lee, 2010) enter the labor market without three types of critical skills. The first type is foundational skills, including basic literacy, numeracy, and soft skills, which improve employability directly and also prepare youth for subsequent skill acquisition. Soft skills, also called life skills or socio-emotional skills, are defined by the International Youth Foundation (2014) as “a comprehensive set of universal cognitive and non-cognitive skills and abilities, connecting behavior, attitudes, and knowledge.” They include workplace readiness, emotional regulation, and interpersonal skills and have been shown to be strongly correlated with labor market outcomes (Heckman et al., 2006; Heckman and Kautz, 2012). One systematic review (Lippman et al., 2015) identified five critical skills most likely to increase workplace success for the 15 to 29 age group: social skills, communication, higher-order thinking skills (e.g. problem solving and decision making), self- control (e.g. directing and focusing attention, managing emotions and behavior), and positive self- concept. Many of these skills are best developed starting in early childhood, but adolescence remains a crucial window of opportunity, as many of these skills are still malleable through young adulthood but are much harder to develop at later ages (Borghans et al., 2008). The optimal ages for soft skill acquisition 9 Based on data collected for different countries from 2006 to 2015. Values range from 3.1 percent in Namibia to 11.9 percent in Malawi to 29.7 percent in Kenya to 40.8 percent in Tanzania. According to a 2010 Gallup World Survey of unemployed and discouraged youth, 26 percent of those with primary schooling and 24 percent of those with secondary schooling reported “lacking employers’ requirements” as the main reason they were not working (AfDB, OECD., & UNDP 2012). 8 are further explored by Guerra, Modecki, and Cunningham (2014), who bridge the literature in psychology, economics, and education, to analyze and categorize socio-emotional skills important to employers. The authors found that it is important to begin development of resilience, control, teamwork, and initiative in early childhood (ages 0 to 5), and develop problem solving skills, achievement motivation, confidence, and ethics in middle childhood (ages 6 to 11). However, reinforcing these skills with intense practice during adolescence and emerging adulthood is essential to skill development for the workforce. While some of these socio-emotional skills can be taught in the classroom, they are also learned and practiced in everyday life, and influenced by social norms, culture, exposure to public spaces and the workforce, networks, and factors. These factors may limit young women’s development of these skills more than young men’s. Gender stereotypes and differential investments are often actualized in the household, which underscores the importance of involving families in foundational skill development. Because young people who do not gain these foundational skills through formal education or at home are disadvantaged when entering the labor market, there is growing attention to these skills in youth employment programs. More work is needed to disentangle the relative importance of specific types of socio-emotional skills needed for young men and women. In addition to these foundational skills, labor market success requires two other types of skills not normally delivered in primary and secondary education. These include technical and vocational skills that are in demand in the labor market as well as business and entrepreneurship skills, which are especially important in Sub-Saharan Africa where many youth are forced into self-employment due to the scarcity of wage jobs (see Figure 2). However, the market for business and vocational training is weak in most African countries. Many youth cannot afford and are unable to borrow to invest in these skills. Even when training is affordable, youth struggle to identify high quality training in market-relevant skills in a crowded and unregulated market. Youth are more constrained than adults by the weak market for skills acquisition because they have not had the chance to acquire skills, nor learn which specific skills to acquire, through work experience. The failure of education systems to adequately prepare youth, and the resulting difficulty in accessing employment, point to the need for government intervention to improve the public provision of education and training, develop standards and certifications for all training providers (public and private), actively promote both men’s and women’s acquisition of skills, and promote soft-skill development in and out of school. Further, while evidence indicates that technical, business, and soft skills are important to success, relatively less is known about the optimal methods and order of skill acquisition that are most effective in creating a productive workforce. 2.2 Capital Facilitating access to capital is an important way to address youth unemployment, either directly by enabling youth to start businesses to employ themselves or indirectly by facilitating existing firms to expand and hire youth. Yet access to finance remains an important barrier to business creation and growth in Sub-Saharan Africa. According to a World Bank survey of new entrepreneurs in Côte d’Ivoire, Kenya, 9 Nigeria, and Senegal, in the formal and informal sectors, 60 percent of female new entrepreneurs and 40 percent of male new entrepreneurs report finance as the main stumbling block when they set up a new business (Hallward-Driemeier, 2013). Being young makes it harder to access funds to start a business. Youth have lower rates of financial inclusion than adults, and they have had less time to accumulate savings or assets that can be used as collateral, a precondition for most formal loans. According to financial inclusion data from 2014 (Demirguc-Kunt and Klapper, 2015), young adults (age 15-24) in Sub-Saharan Africa are 14.7 percentage points less likely to have saved money in the past year and 6.1 percentage points less likely to have borrowed to start, operate, or expand their business than older adults (age 25+). Women (age 15+) were similarly 4.1 percentage points less likely to have saved and 1.3 percentage points less likely to borrow for business purposes than men. While these figures may reflect lower demand for funds, additional data suggest that access to funds is limited. When asked if individuals could come up with 1/20th of their local per capita GNI within a month in the case of an emergency, young adults were 9.0 percentage points less likely than older adults to say it was possible, and women (age 15+) were 11.9 percentage points less likely than men to say it was possible. The differential for young people may be partially attributed to constrained informal borrowing due to smaller and more capital-poor networks. Young adults are 6.6 percentage points less likely to have borrowed from family or friends than older adults, while there is no apparent differential for women and men (Demirguc-Kunt and Klapper, 2015). Women may face more difficulty than men in accessing capital when starting a business, to the extent that they have lower endowments in assets due to norms or laws that limit women’s asset ownership. We lack evidence on whether young women are more affected by these challenges than older women, since most of the empirical evidence cited does not distinguish between the two. Available evidence is mixed with regard to women’s ability to borrow to start or operate a business. Gajigo and Hallward-Driemeier (2010) find evidence of larger entry barriers, including access to start-up capital, for women to start a new business than for men in four African countries. Using data from nine African countries and 43,908 individuals, Aterido et al. (2013) find a gender gap in the use of formal financial services by both individuals and existing enterprises, although the difference is not significant once key characteristics of the individual and firm are controlled for. In their study, industry and firm size account for much of the unconditional gender gap for existing entrepreneurs, as women tend to operate businesses in sectors that require less external finance. There is some evidence that women may be self-selecting into sectors that require less start-up capital, which could contribute to occupational segregation. However, the direction of causality is unclear. Women may also have lower access to formal savings, as suggested in a study by Dupas and Robinson (2013), which found that female market vendors’ excess demand for savings in Kenya was so high that they were willing to pay for the ability to save, in contrast to male bicycle taxi drivers. Finally, there is also evidence that women who are already in business may be disadvantaged in their use of capital. Fafchamps et al. (2014) suggest that female entrepreneurs are less able to make effective use of capital and to separate business capital from household finances, either due to present bias or a greater risk of expropriation from family members. 10 2.3. Networks and Role Models Networks are critical to young adults’ decisions to join the labor force, invest in schooling or training, and choose an occupation. Once an individual has chosen a career path, networks are also an important source of soft skills and contacts central to success both in business (Feigenberg et al., 2013; Field et al., 2015) and in accessing wage employment. Some global evidence suggests that coupling skills interventions with social network-building for women can improve outcomes through social reinforcement (Austrian and Muthengi, 2014; Chong et al., 2014; Field et al., 2015). In the labor market, the process of matching employers to workers suffers from imperfect information on both sides. In the presence of ineffective judicial systems or restrictive regulations, hiring employees comes with an additional risk. Employers struggling to judge the quality of a prospective worker tend to rely on social networks and word of mouth to identify employees, and may discriminate against youth in the absence of reliable mechanisms to assess their abilities. Youth, particularly poor and vulnerable youth, have inadequate networks and do not hear about or are not referred to the few job opportunities that exist. These challenges are further complicated in Sub-Saharan Africa where wage employment is not the norm and young people often have little exposure to the world of work, exacerbating the disadvantage of poor networks. Young women may experience network constraints more acutely than young men if they experience less exposure to the workplace and employers, more discrimination, and have fewer role models and smaller networks due to limited mobility. Limited research has examined whether and how young women’s networks differ empirically from young men’s and adult women’s, and how those networks enhance or diminish job prospects.10 For example, Magruder (2010) demonstrates that fathers in South Africa serve as useful network connections for sons during the job search process, but not for daughters. A survey of employers in Zambia found that 15.1 percent of employers reported a preference for hiring men for professional positions, and 23.5 percent reported the same for production positions (Chigunta et al., 2013). Evidence from Malawi demonstrates that men are systematically less likely to refer women for jobs, despite the availability of qualified women in their networks (Beaman et al., 2013). Discrimination and network disparities may also partially explain why there is a larger percentage of working men “work for someone else” compared to working women (Filmer and Fox, 2014). Discrimination, whether overt or institutionalized, can diminish young women’s job prospects especially when they are unable to accurately signal their abilities in the formal labor market and employers rely on implicit judgments of their abilities. Finally, the gender of role models in women’s networks can also influence skill acquisition and occupational choice. A study from India finds that increasing the presence of women in leadership positions reduces the gender gap in education and improves young women’s career aspirations (Beaman et al., 2012). The type of female role model can also influence women’s interest in entering a high-return occupation, as detailed in Section 2.5 on occupational choice. 10 Family and kin networks in particular may exert more pressure on women to share resources than they do on men. While this constrains women’s ability to generate income and have control over earned income, there is no evidence on how this pressure influences labor market entry. Suggestive evidence comes from Grimm et al. (2013), who find that among entrepreneurs in West Africa, labor inputs to female enterprises were more affected by the proximity of family and kin, which they hypothesize was due to females’ responsibility to participate in non-business family activities. 11 2.4. Time, Family Formation, and Care Responsibilities Family formation and labor market entry are among the most critical, and most gendered, transitions for young people. Marriage and childbearing are intertwined with labor market participation and human capital investments for youth, and there are two reasons why men and women in Africa experience different long-term effects. First, women tend to form families earlier than men.11 Early marriage (defined as under 18) and teenage pregnancy are still commonplace in much of Sub-Saharan Africa, where the median age at first birth falls below 20 in 20 out of 35 countries with available data.12 By age 25, close to 80 percent of African women have given birth (Filmer and Fox, 2014). Thus, marriage and childbearing decisions and labor market entry decisions are often made simultaneously for young women. Second, prevailing social norms place the primary responsibility for domestic work and child care on women, and the primary responsibility of providing monetarily for the family on men. In the ILO SWTS (see footnote 3), among inactive youth, 16.5 percent of women and 7 percent of men in Malawi cite domestic responsibilities as the reason for inactivity in the workforce; similarly, rates are 17.4 percent of women and 2.2 percent of men in Uganda (Elder and Kring, 2016). These differences in social norms lead men and women to view family formation and job acquisition differently. The responsibility of being the primary family caretaker may reduce demand for female employees, as they are more “expensive” in the eyes of employers. For example, caretaking involves a risk of absence if a child is sick, time off from work in cases of pregnancy, and limited hours available for work or overtime. A lack of labor market opportunities may reinforce the preference of some young women and their families to pursue marriage, rather than employment, as a means to secure an economic future. Early ages of marriage and first birth also mean that during the critical period of labor market entry, women are more constrained by the time they have available for paid work and what kinds of work they can pursue, which may affect their long- term career trajectory. International evidence points to the close relationship between childbearing and labor force participation of women. Panel data from 97 countries indicates that a birth is associated with a two-year reduction in a woman’s labor supply (Bloom et al., 2009). Limited labor market opportunities may reduce the opportunity cost of marriage and childbearing, altering family formation decisions. There is significant causal evidence suggesting that family formation constrains labor decisions for women. Research from the US suggests that access to birth control altered women’s career plans and their age at first marriage (Goldin, 2002). In Latin America, access to contraception resulted in delayed childbearing, which, in turn, led to some increases in how much education they attained, as well as their employment rate (Miller, 2010). These studies show that women’s careers were constrained by their options for reproductive control, though there is little evidence as to whether this relationship holds for Africa. Canning et al. (2015) point out that Africa is unique in having both high fertility and high female labor market participation, perhaps because the informal nature of women’s employment in Africa does not preclude engaging in work and child care simultaneously. Additionally, the cost of children many impose an 11Cohort trends in fertility in Sub-Saharan Africa can be found in Sneeringer (2009). 12Source: Most recent DHS survey between 2004 and 2014 for each country. The median age at first birth is reported for women ages 25 to 29 at the time of the survey. 12 income effect that increases labor force participation. However, qualitative work in Liberia suggests that time spent on child care may be a binding constraint for women to work: Ruiz Abril (2008) found that 41 percent of young women, compared to 31 percent of adult women and just 11 percent of young males reported family responsibilities as a reason for inactivity. Even if family formation does not inhibit women’s labor force participation in Africa, it may affect employment on the intensive margin. Once faced with child care duties, women may work fewer hours than they otherwise would, or they may choose less-productive, low-return occupations that offer flexible schedules, part-time, or home-based work. A cross-sectional study in Guinea, for example, found that women, particularly in rural areas, were significantly more likely to experience time poverty13 due to domestic duties, spending on average 3 fewer hours per week than men on paid work, but 15 hours per week more on domestic activities. Severe time poverty was found to increase with the number of children under age 5 and the number of seniors in the home (Bardasi and Wodon, 2010). Another study found that South African and Tanzanian women were less likely to work, particularly in wage-employment, following the birth of a child, though results are imprecise for Tanzania (Canning et al. 2015). A study from Madagascar finds that when female-owned businesses are located at home, their efficiency is constrained by domestic tasks and sharing norms within the local network (Nordman and Vaillant, 2014). Finally, as indicated above, data from Ghana indicate that the percentage of household work completed by women, relative to men, does not vary with the percentage of household income that they generate (World Bank, 2012). Experience from developed countries suggests that while part-time work can be a good option for women seeking flexibility, it can often trap them in low-quality jobs even after flexibility is no longer needed (ibid). 2.5 Occupational Choice A fifth constraint that young people face when entering the labor market concerns the limitations on their occupational choice.14 Gender segregation in the labor market has potentially large effects on economic growth through sub-optimal allocation of labor across the sectors of the economy (Hsieh, 2013). Although the gendered nature of occupational choice affects both men and women, labor force surveys in Africa consistently find that it is women who are concentrated in low-wage, low productivity industries, whether in self-employment or in wage employment. For example, the 2010 Liberia labor force survey found that women are under-represented in the growth sectors of mining, manufacturing, construction, and services (LISGIS, 2010). The ILO’s Global Employment Trends for Women (2012) confirms the pervasive sectoral segregation by sex across Sub-Saharan Africa, noting that while women have made strides in entering the services sector in the past 20 years, they continue to lag in industry. Female entrepreneurs tend to work in less-productive, female-dominated sectors, and in agriculture, female farmers specialize in less-profitable staple crops (World Bank, 2012). The pattern of segregation also exists along the value chain, as for a given sector, they tend to cluster in less-profitable activities. 13 In this paper, “time poverty” is defined as the need to spend long hours working (in either the labor market or domestic work) because the alternative would be (even deeper) consumption poverty. 14 The fact that women and men undertake different economic activities is not exclusive to Africa (Hallward-Driemeier, 2013). Across the globe, women are more likely than men to engage in low productivity activities and to work in the informal sector (World Bank, 2012). 13 Occupational choice is strongly tied to the gender wage gap. The gender wage gap in the region appears to be heterogonous, but several studies have shed light on the source of the gender wage gap. In a study of 7 West African cities, Nordman et al. (2009) found that women earned 50 percent to 79 percent as much as men, and 39 to 61 percent of this gap was explained by occupation and sector (public, private formal, informal). The study emphasizes the importance of within-sector gender wage gaps. A study from Madagascar showed that informal employment was associated with a larger raw gender wage gap than formal employment. Women in Madagascar, unlike men, are penalized for working in informal wage employment, even after controlling for unobservable individual characteristics (Nordman et al., 2012). Finally, Nordman & Wolff (2009) used matched employer-employee data from Mauritius and Madagascar15 to demonstrate that the gender wage gap is attenuated when firm effects are considered. This suggests that men are more likely to be hired into firms paying higher wages than women. Evidence suggests that informational asymmetries, a lack of role models, social norms, and discrimination all contribute to this pattern of segregation (Hicks et al., 2015; Campos et al., 2015; Alibhai et al., 2015; Beaman et al., 2012; Beaman et al., 2013). Qualitative work led by GIL in Uganda (Campos et al., 2015) and Ethiopia (Alibhai et al., 2015) explored the determinants of female engagement in male-dominated trades, confirming that information was indeed a constraint: women in female-dominated trades were frequently mistaken regarding the earnings of women in male-dominated trades. The studies also found that a significant factor in women entering male-dominated trades was the presence and encouragement of male role models early in a young woman’s career, indicating that information is not enough to overcome occupational segregation. Without information and support, young women often opt for the types of work they see other women doing, even if it is not as profitable as other available opportunities. Vocational training, livelihoods projects, and apprenticeships often reinforce these social norms by offering training in stereotypically female (and low-return) trades such as arts and crafts, tailoring, and beauty salons (Fox et al., 2011). The continued demand from young women for these types of training indicates that overcoming these gender norms requires specialized interventions. Further research is needed to understand the relative importance of information, role models, and preferences as underlying drivers of occupational segregation. 2.6. Safety and Mobility The labor market participation of youth is additionally constrained by concerns about safety and mobility. Unsafe working conditions result from a lack of occupational safety and health laws, and weak enforcement of these laws, particularly for workers in the informal sector, who are usually not covered by existing labor protections. Street vendors and hawkers, for example, are vulnerable to eviction from the locations from which they operate and capricious enforcement of laws regarding licensing and taxation of informal enterprises (Filmer and Fox, 2014). We would hypothesize that youth are more vulnerable to this type of harassment by their concentration in vulnerable employment, limited knowledge of their rights, their relative inexperience in dealing with authorities, as well as the power inequities associated with being young. However, little data are available on harassment. 15 Data from Mauritius is confined to the manufacturing sector while data from Madagascar is from 14 different sectors. 14 While employment can substantially improve women’s welfare and empowerment, it can also increase the risk of gender-based violence, particularly from intimate partners, as a form of backlash against employment-induced changes in empowerment. While there is no quantitative measure of this risk in Africa to our knowledge, Vyas et al. (2015) conducted semi-structured interviews in Tanzania to shed light on the issue. Findings suggest that a woman’s access to money increased her confidence and reduced her frequency of negotiating money with men, a trigger of violence. In some settings, women are also restricted in their movements by families and intimate partners because of the risk and perceived risk of incurring physical harm on the way to work, which may constrain labor market entry. Demographic and Health Survey (DHS) data indicate that many women ages 15 to 49 report the belief that a man is justified in beating his wife if she goes out without telling him: prevalence of this belief ranges from 5.3 percent in Malawi, to 32.5 percent in Kenya, to 49.4 percent in Ethiopia, to 86.2 percent in Guinea. In some countries, younger women report this belief at higher rates than older women. In Zimbabwe for example, 26.1 percent of young women (ages 15-24) reported this belief, while 20.5 percent of women ages 25 to 34 and 18.7 of women ages 35 to 49 reported the same. Constraints on mobility may directly affect women’s preferences for seeking work outside the home, and indirectly affect livelihood development by increasing the time required for domestic responsibilities, and limiting access to education, markets, banks and credit, and social networks. For example, an apprenticeship program in Malawi experienced high rates of dropout. Young women, in contrast to men, often cited distance as an inhibiting factor in program participation (Cho et al., 2013). Women in rural and urban areas experience different burdens and risks associated with travel and transport. While mobility constraints are not well-documented or understood in the region, case studies and qualitative data indicate that women in rural areas contend with social norms that require them to walk long distances, consuming time and putting them at risk for violence and injury. In some cases, women’s use of bicycles is considered inappropriate; in others, women are responsible for the collection of firewood and water, which requires traveling long distances and carrying large loads (Porter, 2008; Porter et al. 2011). Though women in urban areas may have more transportation options, they are still more likely to walk and more likely to find work close to home. This may be a result of the inhibiting cost of transport and risks of harassment while using transport (Salon & Gulyani, 2009; Uteng, 2011). Women and youth in both rural and urban areas may also have lower access to a household’s transportation assets, which could have eased mobility constraints related to norms and safety. Additionally, young women are likely to contend with a much higher risk of sexual harassment in the workplace. Norms regarding sex and control over women’s bodies,16 weak enforcement or lack of laws, and limited capacity to prevent and respond to harassment, all contribute to an environment which limits female mobility and puts women at risk in the course of their employment. While nationally representative prevalence statistics are not available, smaller-scale surveys have found that experience with harassment and the perceived risk of harassment from bosses and other authority figures factors highly into a woman’s choice of type of work. For example, in surveys of young female apprentices in one city in Nigeria, two-thirds reported experiencing physical violence, and 39 percent of this group 16For example, only 64 percent of surveyed boys in a school in South Africa identified forced sex with rape (DeVries et al., 2014). In addition, among women in select African countries who have experienced sexual violence, DHS data suggest that women’s first experience often occurs between ages 15 and 22 (DHS, 2015). 15 reported their employer as the most recent perpetrator (Fawole et al., 2005). In qualitative assessments conducted in Liberia and Rwanda in advance of job training programs, young women reported a preference for self-employment or female-dominated industries because of the threat of sexual harassment. In Liberia, 20 percent of those who reported a preference for self-employment mentioned avoiding sexual harassment as a reason for their preference (Ruiz Abril, 2008). In fact, their fears often stemmed from negative experiences with teachers or classmates in school (e.g., teachers trading sex for grades) and lowered their willingness to participate in the training programs themselves. Surveys have found similar fears among female entrepreneurs in Tanzania (Holla et al., 2010) and cross-border traders in Rwanda (Brenton et al., 2011) and Liberia (UNWomen, 2012). These risks are particularly high for young women and girls who work in other people’s homes, interact with officials frequently, or need to purchase inputs from men. The risk of sexual harassment can affect the types of work women do directly, through their preferences, or indirectly, through pressure from families and communities to avoid professions perceived to be unsafe or unsuitable. Thus, concern surrounding safety may prevent labor market entry for young women or limit the locations that young women seek work in, and thus constrain their initial work experience to particular occupations or low-return jobs. 16 Table 1. Summary Table of Constraints to Employment for Youth and Young Women Ways constraint binds more on Ways constraint binds more on Constraint youth women Constraint  Lack of experience: youth have  Lower average educational Applies Skills: Education systems not had time to acquire skills attainment, possibly tied to To All and families do not through practice. social norms or lower expected Youth adequately build skills for  Youth cannot afford or are returns to human capital productive employment. unable to borrow and invest in investment. skills.  Lower endowments in assets and social capital, influenced by Capital: Asset norms and law. endowments and access  Less time to accumulate  Lower usage of formal financial to finance are a barrier to collateral and build networks. services. business creation and growth.  Competing priorities, greater risk of expropriation in use of capital.  Smaller social networks related Networks and Role to norms. Models: Poor signaling of  Small and inadequate networks.  Individuals in network are less skills and trust, and weak  Limited job search skills. likely to refer women. institutions for contract  Preference-based and statistical  Individuals in network may be enforcement leads discrimination. less likely to have connections employers to rely on to high quality jobs. word-of-mouth for hiring.  Fewer female role models.  High fertility rates and concurrent fertility and Time, Family Formation, employment decisions increase and Care need for flexibility. Responsibilities: Family  Choices made at a young age  Gender norms around child care formation and domestic may be persistent. responsibilities limit time and domestic duties limit women’s availability for paid available for work work and increases need for flexibility.  Women concentrated in low- Occupational choice: return industries.  Choices made at a young age leads to sub-optimal  Choice limited by informational may be persistent. allocations of labor asymmetries, lack of role models, social norms.  Higher risk of sexual harassment related to norms regarding sex and control over  Lack of experience and power women’s bodies. Constraint Safety and Mobility: inequities limit ability to  Mobility restricted by risk of Applies Lack of safe transport and identify and respond to safety intimate partner violence, Primarily working conditions leads issues. harassment on the way to work. To to sub-optimal allocations  Possibly, lack of access to  Limited access to motorized Young of labor. vehicles, even those owned by transport, and social norms Women the household. which prevent usage of transport options.  Limited capacity to prevent and respond to harassment. 17 3. Evidence on interventions addressing constraints to young women’s employment This section reviews the evidence on the effectiveness of programs to increase young women’s employment in Sub-Saharan Africa. The review aims to comprehensively cover the available rigorous (experimental or quasi-experimental) evidence in Sub-Saharan Africa, with selective inclusion of relevant studies from Latin America and South Asia. Interventions are organized by the constraint they try to alleviate and by the strength of the evidence in support of the intervention. Programs to develop skills (both vocational and business skills) have the largest number of evaluations, but the resulting evidence is weak at best. Relatively few evaluations have addressed the other constraints to young women’s employment, such as care responsibilities, limited occupational choice, and safety. Instead, we include interventions that were not designed to increase employment but rather may influence it indirectly, such as schooling interventions which delay family formation and child care interventions that free up women’s time. We did not find any stand-alone interventions to expand young women’s networks, although the safe space programs reviewed at the end of this section address networks as one of many constraints.17 To date, the evidence favors interventions that address multiple constraints simultaneously, but these comprehensive programs remain a “black box”, with limited insight as to which program elements are most effective at addressing which constraints. Relative to the rest of the world, the evidence base and effectiveness of various interventions differs for Sub-Saharan Africa. For an overview of findings from the global evidence base, see Box 2. 17One study of microfinance groups in India demonstrates the monetary returns to social interaction: Feigenberg et al. (2013) found that a higher frequency of social interaction with a self-help group increased individuals’ willingness to take risks and reduced their likelihood of default dramatically. 18 Box 2. What have global systematic reviews revealed about the effectiveness of Active Labor Market Programs (ALMPs)? Several papers have been written on the effectiveness of Active Labor Market Programs (ALMPs) on a global level. Some have focused on youth and others have focused on women’s employment. Todd’s (2012) review of youth employment programs finds that ALMPs (including wage subsidy, internship, placement, and public works programs) are effective at increasing employment rates for young women, often more so than for men, in the context of Latin America. However, these programs have a mixed record on their ability to increase earnings among young women. One explanation is that these programs reduce search times such that individuals do not wait for higher-wage opportunities. Other studies of ALMPs examined 107 youth employment programs in both developed and developing countries, and focused on entrepreneurship programs rather than public works programs (Kluve et al., 2016; S4YE, 2015). One-third of the programs were found to have statistically significant positive effects: effects were stronger in low- and middle-income countries, and tended to be higher in long-term evaluations than short-term evaluations. Much of the differences in performance appear to stem from design, implementation, and context. Several reviews extolled the merits of multi-service programs, and their effectiveness in addressing multiple constraints for women in particular (Buvenic, 2016; Kluve et al., 2016; S4YE, 2015). Skills training is the modal youth employment program in the world, and many reviews have found skills training programs to be effective if they are demand driven and located in areas where the formal labor market is developed (Buvenic, 2016; Kluve et al., 2016; S4YE, 2015; Buvinic and Furst-Nichols, 2014). However, Kluve et al. (2016) observe the largest effect for entrepreneurship programs and multi- service programs. Buvenic (2016) considered the impact of child care programs on female employment, and found their effectiveness to be proven, though results vary with context. Reviews suggest a mixed record for programs addressing constraints to capital. While women are more likely to benefit from microfinance programs if they are specifically targeted, targeting is associated with a cost that may not justify the limited benefits. Evidence from outside of Africa shows that combining microcredit with skills training and provision of peer group space significantly improves young women’s earnings (Dickson and Bangpan, 2012), as can longer repayment periods that allow women time to reduce their aversion to risk (Buvinic and Furst-Nichols, 2014). While small, one-time grants can prove effective for larger SMEs, in-kind grants may be more effective for smaller women- run businesses, and very poor women require more business development support for these grants to be effective at improving earnings. Young women have been shown to benefit from larger unconditional grants, even without support or training, and access to savings accounts significantly improves productivity for female entrepreneurs (Buvinic and Furst-Nichols, 2014). Provision of child care, through vouchers or community child care programs, has been found to be particularly effective in increasing female labor force participation, while also improving outcomes for children (Todd, 2012). 19 3.1 Skills Interventions to build employment-related skills usually focus on either technical skills or business skills, but results are mixed. One recent review of the global evidence on Technical and Vocational Education and Training (TVET) found a mixed record of success in securing employment for graduates and an even worse record on cost-effectiveness (Blattman and Ralston, 2015). In Africa, two randomized controlled trials of ‘pure’ TVET programs (not combined with other interventions) have been conducted. The first is a program in Malawi which placed male and female youth as apprentices to master craftspeople (MC). The program suffered from an unanticipated high rate of dropouts, particularly among women, and did not improve employment or earnings in the short run. The results were more positive for men, who had higher attendance rates and continued their investment in skills development after the training ended. Men were significantly more likely to receive help (food and/or money) from their MC and secure paid work from the MC after the training than women. Women, on the other hand, experienced a loss in savings, despite a stipend, and they were more likely to drop out due to distance, illness or injury, marriage, family obligations, and the expense of attending training. Men dropped out to take advantage of migration job opportunities (Cho et al. 2013). A randomized voucher program for TVET in public or private institutions in Kenya found weak but positive results for trainees (Hicks et al., 2015). Voucher-recipients were more likely to enroll in vocational training and obtain 0.55 additional year of education. Female voucher winners attended 8 percentage points less of their course than men. While rates of employment did not improve and dropout was still a concern, voucher winners were somewhat more likely to obtain wage employment, and obtained higher hourly earnings conditional on wage employment in the short term. Two additional programs in Kenya combined technical and life skills training with internships. First, in the Government of Kenya’s Youth Employment Program (KYEP), one group was offered a short life skills training class, and another randomly selected group was offered the same class and the opportunity to be selected by employers for an additional 3-month business training and a 3-month internship. Among those assigned to either treatment group, 15 months after the program ended, improvements in earnings were not robust, but employment rates were 6.9 percentage points higher for men and 4.5 percentage points higher for women relative to the control group. Among those who completed the program, employment rates were 14.2 and 8.7 percentage points higher, respectively. Both men and women were more likely to be wage workers, but results were only significant for men. Conversely, among wage workers, monthly earnings were higher for both men and women in the treatment group, but earnings were only significantly higher for women. Men and women were close to 45 percentage points more likely to receive training or join an internship, and more likely to accumulate savings and open a bank account, particularly women (Honorati, 2015). Second, the Kenyan Ninaweza program targeted young women living in informal settlements in Nairobi and offered Information and Communication Technology (ICT) training, internships, and job placement support, and life skills training. Participants who received the treatment were 14 percent more likely to obtain a job. They were also more likely to experience an increase in confidence, work full time, and receive benefits and higher earnings (Azevedo, 2012). This limited body of evidence for Sub-Saharan Africa is complemented by a large number of evaluations of TVET programs in other developing countries, such as the Jovenes programs in Latin America which 20 combined technical training with employment-related soft skills and internships starting in the late 1990s. Employment effects in Latin America were often low or not significant for men, but increases in employment rates among women ranged from 5 percentage points in Colombia to 16 percentage points in Panama. While earnings for men did not increase or increased by up to 10.7 percent, women’s earnings increased by up to 22 percent (Gonzalez-Velosa et al., 2012; Ibarran et al., 2015; Ibarraran and Shady, 2009; Attanasio et al., 2011; Attanasio et al., 2015).18 A large job training program for the unemployed in Turkey had modest positive effects on employment rates, which were stronger when training was offered by a private provider. However, effects dissipated quickly, and they were weaker for women and youth (Hirshleifer et al., 2014). In India, women aged 18 to 39 who were offered a training program in sewing were 6 percentage points more likely to be employed, and 4 percentage points more likely to be self- employed. They also worked 2.5 additional hours per week and earned 150 percent more per month than women in the control group (Maitra and Mani, 2012). The evidence on the impact of business skills training19 in developing countries is reviewed in McKenzie and Woodruff (2013), concluding that although the impact of these training programs on employment and business profits for existing entrepreneurs is minimal, they do seem to have success in generating business start-ups. Again, the evidence for Sub-Saharan Africa is lacking, and few studies have a sufficient sample size to disaggregate the effects by sex. Premand et al. (2016) introduce an entrepreneurship track to undergraduate students in Tunisia, and find improvements in business knowledge and optimism. The study finds a 3 percentage point increase in start-ups among women and 6 percentage point increase among men, though the difference is not significant, pointing to the need for studies that are sufficiently powered to detect differential impacts by gender (McKenzie and Woodruff, 2013). A meta-analysis of entrepreneurship training programs by Cho and Honorati (2013) finds that most programs improve business practices, particularly for youth, but do not improve income or the likelihood that an individual will create a start-up. The study finds that financial support is more effective than training at increasing labor market activity and improving business practices for women, and business training is more effective at improving business practices for existing entrepreneurs, both male and female. One recent study in India found that peer effects improved the impact of business skills training: Field et al. (2015) offered business training to women and invited a random subsample of women to attend training with a friend. The study found that four months later, women who attended business counseling with a friend reported higher business activity and household income, in contrast to those who did not attend and those who attended without a friend. Women who attended with a friend were also less likely to report their occupation as housewife. 18 Ibarran et al. (2015) find no statistically significant impact on employment for the Juventud y Empleo program in the Dominican Republic which included classroom training and a private sector internship. They find improvements in long term labor market trajectories, as they relate to job quality and earnings (including benefits), particularly for men and urban women. They do find a marginally significant earnings effect of about 10 percent. An evaluation of a similar program in Colombia found that 10 years later, female beneficiaries experienced a 5 percentage point increase in formal sector employment, and a 17.5 percent increase in formal sector earnings. While men did not experience an increase in formal sector employment, they experienced a 10.7 percent increase in formal sector earnings (Attanasio et al., 2015). 19 We focus on business skills training for aspiring entrepreneurs. The evidence on business skills for existing entrepreneurs is reviewed in GIL’s white paper on women’s entrepreneurship. 21 Finally, interventions that build non-cognitive/socio-emotional skills for children and youth can have long-term implications for earnings and employment. In Jamaica, community health workers visited mothers of 64 stunted toddlers living in poverty. Over a two-year period, health workers encouraged play and psychosocial stimulation. A long-term 20-year follow-up of this study found that participants’ earnings were 42 percent higher than the control group. Participants’ earnings caught up to the earnings of a comparison group of non-stunted individuals of the same age from the same neighborhoods (Gertler et al., 2013). 3.2 Capital Several programs with rigorous evaluations have attempted to improve financial inclusion among women in Africa. One study in Kampala, Uganda, found that providing adolescent girls with savings accounts increased their likelihood of budgeting and saving close to two-fold. Effect sizes increased if girls met regularly with a group and were given financial and reproductive education (Austrian and Muthengi, 2014). Most programs that address the direct link between capital and employment attempt to overcome capital constraints to new businesses and supplement loans or grants with business training. However, microcredit often involves high borrowing costs and short repayment periods that are not conducive to agriculture and self-employment (Blattman and Ralston, 2015). While microfinance programs can result in higher investments, profits, and durable goods expenditure, effects dissipate quickly and health, education, and female empowerment are often unchanged (Banerjee et al., 2015). However, effects are highly dependent on variation in targeting and supplemental training (Todd, 2012; Dickson and Bangpan, 2012). The closest evaluation of pure cash grants for self-employment examines the Youth Opportunities Program (YOP) in Uganda (Blattman et al., 2013). The program invited youth to form groups and apply for grants to obtain training and start a skilled trade. Close to 17 percent of groups were single sex, according to the discretion of participants, and the rest were mixed with one-third female on average. The cash transfers had large impacts as they led to substantial and persistent increases in individual investment, work hours and income. The gains for women who benefited from the YOP project in Uganda were greater than for men, despite their initial human capital and experience being lower. The authors find that after four years, incomes of treated women are 84 percent greater than for female controls (compared to a 31 percent relative gain for men). Another program in Uganda, the WINGS project, provided basic five-day training and cash grants to poor women ages 20 to 35. The results indicate that the program doubled business ownership and income (Blattman et al., 2014). These results indicate that the provision of capital, perhaps in combination with skills, is a promising intervention to stimulate self-employment. 22 3.3 Time, Family Formation, and Care Responsibilities Interventions to overcome constraints related to young women’s care responsibilities can be classified into two types: those that seek to delay marriage and reduce fertility through supply-side or demand-side interventions and those that alleviate time constraints for women who already have children. While a comprehensive review of supply-side interventions to increase access to family planning is beyond the scope of this paper, studies from outside Africa show that increasing reproductive control reduces and delays fertility, particularly among less educated women, and increases educational investments (e.g., Pop-Eleches, 2010; Gershoni and Low, 2015). However, supply-side interventions have weaker results in Africa,20 which suggests that modifying preferences towards desired fertility may be more promising than increasing the supply of family planning. Demand-side interventions to delay marriage and reproduction often focus on improving enrollment and retention of girls in school, based on the positive association between secondary education and delayed childbearing.21,22 Several of these programs have been rigorously evaluated and their results are promising. Perhaps the best known example is the Zomba Cash Transfer program in Malawi, which provided initially never-married females aged 13-22, and their parents, with financial support in the form of monthly cash transfers for two academic years. A randomized evaluation of the Zomba program found that cash transfers conditioned on school attendance had a positive impact on girls who had dropped out of school at baseline. One year later, these girls were 45 percentage points more likely to be enrolled in school, 12 percentage points less likely to have ever been married, and 5 percentage points less likely to be pregnant (Baird et al. 2009, Baird et al. 2011, Baird et al. 2014). Another randomized evaluation in Kenya (Duflo et al., 2014) found that providing two free school uniforms over the last three years of primary school reduced the school dropout rate after three years from 19 percent to 16 percent for girls and the teen pregnancy rate from 16 percent to 13 percent. This reduction came entirely through a reduction in the number of pregnancies within marriage, and there was no change in the out-of-wedlock pregnancy rate. By year 7, there was still a 7 percent gap in the childbearing rate between girls exposed to the education subsidy program and those in the control group (46 percent vs. 49 percent). The evaluation of the Kenya Cash Transfer for Orphans and Vulnerable Children, a national unconditional cash transfer program, found a reduction in the likelihood of pregnancy by five percentage points among females aged 12 to 24, but no reduction in early marriage, after four years of the program. The authors posit that the impacts were the result of increased school enrollment and financial stability of the household as well as delayed age at first sex (Handa et al., 2015). 20 In Africa, there are either experimental or quasi-experimental studies that analyze the impact of contraceptive access on fertility in Ethiopia (Desai and Tarozzi, 2011), Ghana (Debpuur et al., 2002), Tanzania (Angeles et al., 1998), Uganda (Lutalo et al., 2010) and Zambia (Ashraf et al., 2014b). Significant reductions in fertility are found only in Ghana and Tanzania. No impact is evident in Ethiopia, Uganda, and Zambia, even when larger access translated into larger demand for contraception. Results are mainly explained by the fact that women increased demand for better contraceptive technology and therefore substitute away from other methods of contraception (Ashraf et al., 2014). 21 Murphy and Carr (2007) and the World Bank (2012). In Africa, a strong causal link between increases in education and reduction in fertility has been found in Nigeria (Osilii and Long, 2008), Kenya (Ozier, 2011) and Uganda (Keats, 2014). 22 These interventions, including school feeding, conditional cash transfers, latrine construction, sanitary pad distribution, and provision of uniforms and books, often seek to reduce early marriage and pregnancy as a primary or secondary outcome (see Lloyd and Young (2009) for a review). 23 Rather less evidence is available on the second type of intervention: alleviating the time constraints for women who already have children through child care services. To our knowledge, the only experimental evidence in Africa is a randomized evaluation of a preschool intervention in Mozambique, which relieved caregivers of 15 hours of child care per week and yielded a 26 percent increase in labor force participation of preschoolers’ caregivers (Martinez et al., 2012). While a large proportion of primary caregivers are mothers, the sample also included fathers, older siblings, and grandparents, and effect sizes were not isolated for young women. These results are consistent with positive results from Latin America (Attanasio and Vera-Hernandez, 2004; Berlinski and Galiani, 2007; Berlinski et al., 2009). 3.4 Occupational Choice Few interventions have addressed the occupational choice of young women directly, and the only experimental evidence from Africa relates to labor market information, under the hypothesis that young women and their families have little knowledge on the availability and profitability of jobs. Hicks et al. (2015) study vocational training vouchers in Kenya and found that, prior to enrollment, students were frequently mistaken about which trades had the highest average earnings in their region. In addition, preferences for which training courses to pursue were divided sharply along gender lines. A randomized controlled trial providing accurate information about the returns to various male-dominated and female- dominated trades was successful in getting young women to sign up for training in male-dominated trades. However, further follow-up found that these women were not more likely to complete training or pursue work in those trades than women who were not provided with such information. Information in the form of media and “edutainment” may be particularly effective in affecting occupational choice. In Ethiopia, poor farmers were randomly invited to view a one-hour documentary on individuals from similar communities who had succeeded in agriculture or small business. After six months, these individuals exhibited more forward-looking behavior. They scored higher on tests of aspirations, saved more, and were more likely to enroll their children in school. When a larger density of individuals had viewed the documentary in a village, effects were stronger, confirming that peer networks affect individual aspirations (Bernard et al., 2014). Suggestive evidence collected by GIL in Uganda and Ethiopia demonstrates that peer groups and mentorship could be important complements to information interventions for occupational segregation in particular, though this has not been rigorously tested (Alibhai, 2015; Campos, 2014). Evidence from South Asia confirms the potential of interventions to tackle social norms surrounding occupational segregation and aspirations through two interventions: (i) exposing girls to industries where no clear gender assignment has been made, and (ii) placing women in leadership positions. A program in India evaluated by Jensen (2012) provided three years of recruiting services to help young women in randomly selected rural Indian villages get jobs in the business process outsourcing (BPO) industry. By connecting the villages to experienced recruiters, the intervention aimed to increase awareness of and access to BPO jobs, and increase employment opportunities for women. The intervention led to employment gains and increased enrollment in post-school training courses for working women aged 18 to 24 and corresponding delays in marriage and fertility. Women aged 15 to 21 at baseline from villages exposed to the recruiting intervention were 4.6 percentage points more likely to work in a BPO job than 24 women in control villages and 2.4 percentage points more likely to work for pay outside the home. They were also 5.1 percentage points less likely to get married and 5.7 percentage points less likely to have given birth during the three-year period of the study, consistent with increased aspirations for a career. Beaman et al. (2012) also found that local female leadership improved girls’ aspirations relative to boys and parents’ aspirations for their daughters relative to those for their sons. The proportion of parents who believed that their daughter’s occupation should be determined by her in-laws fell 11 percentage points when the community had a female leader. 3.5 Safety and Mobility While the changes in social norms needed to reduce workplace sexual harassment may take a generation or more, interventions to promote gender awareness and safety in workplaces can make a difference in the short term. Unfortunately, there is close to no rigorous evidence on women’s safety in the workplace or on the way to work, nor on interventions that could improve safety and facilitate female labor force participation. The ILO’s Better Work Program has demonstrated success in improving worker conditions and productivity for the garment industry in Cambodia and expanded to seven other countries including 25,000 workers in Lesotho, though it does not target safety issues in particular. The program uses a multi- stakeholder approach and offers training and advisory services to promote behavioral change. Its primary function is ensuring compliance with national and international labor standards addressing safety, compensation, and discrimination. In Lesotho, both male and female workers report better welfare, lower stress, and improved occupational safety due to better factory policies, financial literacy classes, greater awareness, and improved communication (Pike and Godfrey, 2015; Kotikula et al., 2015). However, experimental methods have not been utilized, and the studies have not evaluated effects on a woman’s decision to work. A recent World Bank survey of traders on the Rwanda-DRC border documented high reported rates of threats, harassment and violence such as beatings, stripping, and even rape among the predominantly female traders (Brenton et al., 2012). In response, the World Bank is facilitating a series of trainings for both border officials and adult female cross-border traders about their rights. Recent preliminary evidence from the project’s associated impact evaluation reveals reductions in sexual and gender-based violence with no increase in “unofficial” fees. Other interventions have formed associations of female traders to travel together when transporting goods (UN Women, 2012) or provided workshops for female workers on their rights as well as how to seek help in cases of violence (Fawole et al., 2003; Fawole et al., 2005), though rigorous evidence on the impact of these measures is lacking. In order to address mobility constraints on the way to work, a number of programs in Mexico City, Cairo, Jakarta, Japan, New Delhi, Rio de Janeiro, and Moscow have attempted to address harassment on public transportation with sex-segregated transportation. However, the programs have not been evaluated (UN Habitat, 2013). Another intervention provided unemployed youth, who were willing to take a new job within the following two weeks, with non-fungible subsidies for transportation to the center of Addis Ababa for 12 weeks. A sample of more educated, active job searchers was far more likely to find secure employment, while a sample of those withdrawn from the formal labor market was more likely to be 25 employed and have higher quality jobs. However, the results were not disaggregated by gender (Franklin, 2015). 3.6 Comprehensive Programs The set of interventions with the most promising results at increasing young women’s employment are those that relieve multiple constraints simultaneously. First, center-based TVET programs (those that are based in classrooms with accredited teachers) can be enhanced with girl-friendly design elements to address an array of constraints that go beyond skills, including the inadequacy of networks, and care and capital constraints.23 Under the World Bank’s Adolescent Girls Initiative, a randomized controlled trial (RCT) of such a program in Liberia showed promising results. The Empowerment of Adolescent Girls and Young Women (EPAG) project provided socio-emotional and employment skills training to 2,500 young women in two randomly assigned sequential rounds from 2010 to 2012. Key design features included stipends contingent upon attendance, free child care provision, small mentorship groups, savings accounts, and performance-based contracts for the service providers. The program included six months of classroom training and six months of follow-up support, and allowed participants to choose between a business skills track (65 percent of participants) and job skills track (35 percent). The results were impressive (Adoho et al., 2014): EPAG increased employment by 47 percent and earnings by about 80 percent (or USD 32 per month). The impacts were much larger for the business skills track than job skills, which is not surprising given the limited job opportunities in and around Monrovia. The impact evaluation also found evidence of greater empowerment, as measured by a 10 percent increase in control over earned income. EPAG graduates also report that they are more confident about future outcomes and were more satisfied with their jobs or occupations. There is no evidence of any impacts on fertility preferences or sexual behaviors, such as contraceptive use or number of partners. This, too, was not surprising given that the life skills curriculum focused mostly on employment-related soft skills (including sexual harassment) and to a lesser extent on sexual and reproductive health. The lack of impact on fertility and sexual behavior in this study points to a need for further research on the conditions under which economic empowerment may lead to reduced fertility, the mechanisms of change, and the length of time required to affect change. The second class of comprehensive interventions are community-based ‘safe space’ programs. This increasingly popular model to promote young women’s empowerment and employment involves the formation of girls’ groups who meet regularly over several weeks or months, and aims to address the safety, capital, networks, care, and skills constraints concurrently. There are critical distinctions between this type of intervention and center-based TVET: (i) groups meet within walking distance of girls’ houses, minimizing travel costs; (ii) female peer- or mentor-led training minimizes the cost professional trainers; (iii) there is an emphasis on life skills including sexual and reproductive health; (iv) the girls-only spaces; and (v) the focus on younger girls (often as young as 10) before the onset of marriage and sex. These groups serve two critical functions: first, the ‘safe spaces’ are a platform in which girls can be reached with a variety of interventions and educational topics. For example, these programs often include complementary interventions to build financial assets or teach vocational skills. Second, peer groups build 23 Limited female-friendly enhancements, such as child care, have been included in TVET programs elsewhere (e.g., some of the Jovenes programs), but, to our knowledge, not in sub-Saharan Africa. 26 social capital, including friendships, trusting relationships, and self-esteem, which can have a positive influence on different dimensions of girls’ lives, including their livelihoods, health, and perceptions of social norms. The only experimental evidence of a community-based safe space program in Africa to date comes from the Empowerment and Livelihoods of Adolescents (ELA) program in Uganda.24 The program offered community-based girls’ clubs, life skills training, vocational training, financial literacy, savings and microfinance to young women aged 12 to 25. In partnership with BRAC, an international NGO, GIL has implemented cluster-randomized trials of club-based girls’ programs in Uganda, Tanzania, and South Sudan, based on a model developed by BRAC in Bangladesh. Life skills training was delivered through a cascading training model in which adolescent leaders from each club were taught by more experienced trainers before delivering training to club members. The emphasis of the life skills curriculum on sexual and reproductive health (SRH) topics reflects the program’s equal emphasis on reducing early marriage and teen pregnancy and improving livelihoods. Short vocational training courses in a variety of fields such as poultry rearing, usually lasting less than a month, were delivered by BRAC or outside trainers. The clubs operated for a minimum of two years, at a much lower total cost compared to the average TVET program. Results from Uganda are quite promising (Bandiera et al., 2015). The ELA program increased the likelihood that girls engage in income generating activities by 72 percent (driven by increased self-employment). On the extensive margin, adolescent girls were 62 percent more likely to have some earnings from self-employment; and on the intensive margin, self-employment earnings increased by three times their baseline level. Even more impressive were the parallel results on SRH outcomes: a 26 percent drop in teen pregnancy, 58 percent fall in early marriage/cohabitation, 7 percent drop in desired fertility, 25 percent increase in condom use, and 41 percent decline in girls reporting having had sex against their will in the past year. Program costs considering a take-up rate of 21 percent amounted to $85 per participating girl per year. As one of the only rigorously tested adolescent girls programs to demonstrate such large effects on both livelihood and SRH outcomes, this evaluation opens a number of new research questions, particularly with regards to disentangling the effect and efficiency of each intervention and investigating whether there is an optimal age threshold when empowerment programs have an effect on sexual and reproductive health (SRH). 24 The Population Council, an international NGO, has also pioneered the safe spaces approach. Several quasi-experimental evaluations (Erulkar, 2009; Erulkar and Medhin, 2014; Engebretsen, 2013) have shown impressive results. Two ongoing randomized programs in Kenya and Zambia combine safe spaces with complementary education, health, and savings interventions are expected to have results in 2017. 27 Table 2. Summary table of interventions to address constraints to young women’s employment Is evidence Constraint Intervention Included Rigorous Evidence supportive of intervention? Blattman and Ralston, 2015; Cho et al., 2013; Field et al., 2015; Hicks et al., 2015; Premand et al., 2016; Honorati, 2015; Azevedo, 2012 Technical and Business Skills Mixed training Outside Africa: Gonzalez-Velosa et al., 2012; Card et al., 2011; Ibararran, 2009; Attanasio et al., 2011; Attanasio et al., 2015; Hirshleifer et al., 2014; Maitra and Mani, 2014; Gertler et al., 2013 Microcredit Banerjee et al., 2015 (review) Mixed Insufficient Savings programs Austrian and Muthengi, 2014 Capital evidence Blattman et al., 2013; Blattman et al. Provision of Grants Positive 2014; Haushofer and Shapiro, 2013 Desai and Tarozzi, 2011; Debpuur et al., 2002; Angeles et al., 1998; Lutalo et al., 2010; Ashraf et al., 2014b Increased Reproductive Mixed Control Outside Africa: Pop-Eleches, 2010; Time and Gershoni and Low, 2015 Family Delayed marriage and Formation Baird et al., 2014; Duflo et al., 2014; childbearing through Osilii and Long, 2008; Ozier, 2011; Positive support for education Keats, 2014; Handa et al., 2015 and cash transfers Increased time Insufficient availability through Martinez et al., 2012 evidence child care provision Insufficient Information Hicks et al., 2015 Occupational evidence Choice Outside Africa: Jensen, 2012; Insufficient Recruiting services Beaman et al., 2012 evidence 28 Insufficient Training on legal rights No rigorous evidence evidence Safety and Employer-Employee Insufficient No rigorous evidence Mobility Committees evidence Insufficient Transportation subsidies Franklin, 2015 evidence Peer learning and Vasilaky & Leonard, 2014 Insufficient network building evidence Multiple Insufficient Girl-friendly TVET Adoho et al., 2014 evidence Safe space Bandiera, 2015; Erulkar, 2009; Positive empowerment programs Erulkar, 2014; Engebretsen, 2013 Note: Table includes experimental and quasi-experimental studies that we are aware of, primarily sourced from economic journals and multilateral institutions. In cases where interventions are labeled with “insufficient evidence,” no rigorous study has been conducted. 29 4. Research Agenda This section proposes an ambitious research agenda on young women’s employment focused on key evidence gaps uncovered in the literature. Within this space of unanswered questions, GIL has developed a set of priorities for its research agenda. The agenda is organized by the previously outlined constraints to young women’s employment. For each constraint, we discuss both the research needed to better understand or document the severity of the constraints, as well as the research aimed at what works to address these constraints. There is significant overlap between the “understanding constraints” and “what works” agendas. Questions regarding the nature of the constraints often require the use of a variety of research methods, including inferential research using nationally representative data sources and qualitative or mixed methods approaches. Questions regarding what works are most suitably addressed using rigorous impact evaluations. In some cases, adding survey questions to existing impact evaluations can provide insights into the causal mechanisms at play. The section concludes with priorities for cross- cutting issues such as scaling up and measurement. 4.1 Skills Given the mixed record of vocational and business skills training exhibited both internationally and in Africa, as well as the high cost of training relative to other interventions (such as cash transfers), researchers have started to question the continued investment in training programs in Africa. However, even the most ardent critics acknowledge that when training does have a positive impact, it is usually for women, particularly young adult women (Blattman and Ralston, 2015). Hence, before abandoning skills training entirely, more work is required to unpack which types of skills are most needed and by whom, as well as how skills can be more cost-effectively delivered. The review of evidence in Section 3 shows that comprehensive programs that include skills training among other interventions have shown some success for young women in Africa. However, these programs leave open the question of how important skills are relative to other constraints, such as care, capital, networks and role models, in promoting young women’s employment. The next generation of impact evaluations needs to unbundle these comprehensive interventions to examine which interventions are the most effective and cost-effective, and in so doing, shed light on the fundamental question of which constraints are the most binding, and for whom. Which types of skills? Understanding the relative importance of technical, business, and life skills in labor market success is at the forefront of the youth employment research agenda, and GIL’s work seeks to examine whether young men and women require different mixes of skills. The optimal package of skills can be investigated through impact evaluations that vary the combination and intensity of the three types of skills. An ongoing GIL evaluation of a girls’ empowerment program in Sierra Leone is testing the impact of offering life skills training alone versus in combination with vocational skills. Other GIL evaluations are testing the effectiveness of a pure life skills intervention for girls in Liberia, soft skills training as an alternative to internships in Togo, and the combination of internships and training vouchers in Togo. More such evaluations are needed to determine the most impactful and cost-effective combination of skills. Socio-emotional skills training in particular merits further investigation, both 30 because it tends to cost much less than technical or business skills, and because of variation in programs and studies as to the definition of “socio-emotional skills”. Future research should examine which types of socio-emotional skills matter most for young women, by experimentation within an impact evaluation, or by documenting skills gaps between young men and women through inferential research, for example using the World Bank’s STEP surveys.25 In terms of delivery, more experimentation is needed with lower-cost methods of delivering skills, including the use of non-professional trainers (including peer-to- peer learning or mentors) or remote learning, as well as comparing classroom-based training versus on- the-job training through internships or apprenticeships. As formal education is often not indicative of ability or skill level in certain contexts, an ongoing GIL evaluation in South Africa is running cognitive and non-cognitive assessments of young women and men, then reporting them back to individuals to examine the impact of results on job search methodology and signaling ability to employers. Which skills when? Another research area concerns the optimal sequencing of skills, in light of the path- dependence of skills acquisition. People build on the skills they already have, and skills gaps originating from younger ages become harder to close over time.26 Consequently, an important line of research would examine the impact of building foundational skills, including literacy, financial literacy, and age- appropriate life skills, for younger girls (ages 10-15). These foundational skills could be promoted either through formal school or community-based programs. In addition to their direct impacts, such interventions might better prepare girls for vocational or business skills training programs at later ages. How big is the gain (if any) from intervening sooner in the life of young women? In terms of delivery, how do school-based approaches compare to community-based ones? Can improved foundational skills for younger girls affect their labor market outcomes, even if not followed by subsequent employment- related skills training? Would the channels of impact of such interventions be direct, through the building of skills, or indirect, through potentially delayed family formation? These questions could be examined through longer term follow-up of existing programs for younger girls that promote education (e.g., the Zomba cash transfer program) or life skills (e.g., the GIL-evaluated Sisters of Success life skills program in Liberia). Does participation in such programs improve later skills acquisition when offered vocational and/or business skills training at an older age? Skills or Cash? The success of cash transfers has led a growing number of experts to question whether cash alone is the best method to improve the livelihoods of poor people in Africa.27 One idea is for every development intervention to be benchmarked against cash transfers.28 In the case of training programs, the argument is that either skills are not the binding constraint they are often assumed to be (in the absence of training, poor people are able to improve their earnings and productivity) or that the cash can allow those constrained by a lack of skills to invest in training on their own. However, even if cash outperforms skills 25 The Skills Towards Employment and Productivity (STEP) household surveys assess reading and numeracy proficiency, job- relevant skills, socio-emotional skills, and risk preferences across cohorts of working-age individuals. STEP employer surveys assess the skills that employers look for when hiring, the satisfaction of employers with the skills level of the labor force, and the provision of training and compensation. 26 http://blogs.worldbank.org/futuredevelopment/skills-gaps-and-jobs-strategies 27 Ozler, Berk. “Should we just give people cash?” Policy Research talk given on September 29, 2015. And Kenny, Charles. “Give Poor People Cash”. The Atlantic. September 25, 2015. http://www.theatlantic.com/international/archive/2015/09/welfare- reform-direct-cash-poor/407236/ 28 Blattman and Niehaus. “Show them the money”. Foreign Affairs. May/June 2014. https://www.foreignaffairs.com/articles/show-them-money 31 training for adults, or young men, we can postulate a number of circumstances in which young women might benefit from skills trainings more than from cash transfers: (i) in settings where young women have limited agency over how to use cash; (ii) in settings where wage jobs are available, and certification is more likely to improve chances of obtaining these jobs relative to cash; (iii) in settings where high-quality training, both foundational and technical, is not readily available even for those who are willing to pay, such as conflict-affected settings; or (iv) where secondary outcomes of training programs are critical, such as self-esteem or knowledge of sexual and reproductive health, but vocational training is the only way to get families to allow girls to participate. For these reasons, a key research priority is to test cash versus skills training (as well as cash plus skills) for young women in a variety of settings to determine the conditions under which one is likely to perform better than the other. GIL researchers will soon evaluate such a test among youth in Kenya and Benin. 4.2 Capital How do young women access capital to start businesses? More research is needed to first document the constraints precluding young women’s access to capital to start a business, as much of the existing research focuses on adult women, and often on those who already own enterprises. With a better understanding of how young women start businesses (through savings versus borrowing) and how their preferences and constraints regarding the use of money differ from those of young men and/or adult women, practitioners and researchers will be better positioned to design and test girl-friendly financial services and products to alleviate those constraints.29 As young women’s limited collateral can limit her access to credit, GIL is studying the use of psychometric tests as an alternative to collateral. Preliminary results of the pilot suggest high repayment rates and returns. We also need to understand whether young women are more or less affected by challenges that impact adult women such as risk of expropriation from family members or increased risk of sexual harassment from obtaining control over economic assets. A study by GIL in Côte d’Ivoire is currently examining the relationship between risk of expropriation and employment decisions and earnings, by evaluating a program offering forced saving mechanisms to agro-processing employees. 4.3 Time, family formation and care responsibilities How, when, and for which subgroups does employment affect young women’s fertility? Existing evidence is mixed with regard to the impacts of increased labor force participation on young women’s fertility preferences, choices, and behaviors. More traditional TVET-style programs, such as the EPAG program in Liberia, have not documented significant impacts on fertility outcomes, while club-based approaches, such as the ELA program in Uganda, have dramatically lowered fertility and increased contraceptive use.30 One important question should examine whether there is an optimal age threshold when job training programs have an effect on sexual behavior. For example, while the EPAG program 29 GIL currently has studies underway in Côte d’Ivoire and Nigeria examining the impact of various financial products (e.g. in- kind versus cash transfers and savings products) on adult women’s agricultural profits and the redistribution pressure faced by women. 30 See Section 3.6 for description of ELA and EPAG programs. 32 average age was 22.7, the ELA program accepted girls between 14 and 20. It is necessary to understand to what extent age played a key role in mediating the effect of such programs on sexual and reproductive health outcomes. Because these dynamics change substantially after a woman has her first child, one priority area for employment interventions is to recruit and retain young women who have not yet begun having children, which often means women under age twenty. Experience from GIL’s previous evaluations indicates that club-based programs are more effective at reaching women under age twenty than TVET models. With an eye toward uncovering the mechanisms through which employment and fertility changes occur as well as the complementarities between these mechanisms, all GIL youth employment evaluations track reproductive outcomes. Qualitative work to uncover how young women think about childrearing in relation to employment and their personal relationships will be critical. We do not have enough evidence regarding how expected returns to schooling, work, and marriage may shape family formation decisions of young women in Africa. Does child care provision improve young women’s employment outcomes? Once young women become mothers, one option to expand their labor market options is to promote access to child care. Upon reviewing the international evidence to date, the 2012 World Development Report on Gender and Development concluded that improving access to child care services does in fact improve maternal labor force participation, but cautions that the size of the impact depends on the extent to which formal child care displaces informal arrangements, the financial costs, and the perceived quality of the formal child care. As discussed in Section 2, the tradeoff between labor force participation (LFP) and fertility is not as clear in Africa as in other regions: high rates of women’s LFP and fertility co-exist in a number of African countries. However, even if child rearing does not keep young women out of the labor force entirely, it may still compel them into informal, part-time work in female-dominated, low-return industries early in their working lives. Research in African contexts is needed to determine if mitigating these care constraints early through the provision of child care could increase women’s productivity, reduce occupational segregation, or increase long-term earnings. A rigorous impact evaluation would shed light on the severity of the child care constraint in addition to documenting the potential for child care provision to address it. How can we expand access to child care? If child care provision does increase young women’s participation or earnings, then a second question is how to improve the availability, quality, and use of child care. For this, a better understanding of how child care markets function in both rural and urban contexts in Sub-Saharan Africa is needed. How does the market for formal child care evolve in response to increased demand? What are the general equilibrium effects of increased female employment on norms relating to paid child care and on the availability and cost of informal and formal child care options? Is there a threshold of the percentage of women working outside the home that tips the market toward formal, rather than informal or family-based, child care arrangements? And are younger women more or less likely to be early adopters of formal child care, and to reap larger and longer-term benefits from having access to formal child care early in their working lives? A related question, building upon ongoing GIL evaluation in Burundi, is what works to promote fathers’ participation in child care? These questions could be explored in the context of an impact evaluation that exogenously increased women’s 33 employment in a limited geographical area where one could examine the attitudes, availability, and use of informal and formal child care at the community level. 4.4 Occupational Choice Young women are a critical target group for interventions to reduce occupational segregation, as existing research points to the persistence of early career choices and the importance of sector choice in explaining productivity and earnings differentials between men and women. A cautionary approach to the issue is required, because while occupational segregation may be undesirable at a societal level, at the individual level there are various risks facing young women who venture into new fields, including safety and harassment. Which trades for whom? Looking ahead, we can think of three conceptual approaches to supporting women to enter higher-return trades. First, we can tackle the issue head-on by encouraging women to enter trades that are already male-dominated (e.g., more female electricians). The only RCT we are aware of to do this, involving an informational intervention in Kenya (Hicks et al., 2015), was not successful in increasing women’s engagement in male-dominated trades. A GIL evaluation of a vocational training program in the Republic of Congo includes a variation on the information experiment from Kenya, paired with a financial incentive for women to sign up for non-traditional trades. Another GIL evaluation Togo is examining the impact of managerial and entrepreneurship training on women’s leadership and “cross- over” to male-dominated sectors. The second approach is to encourage entry to existing higher-return trades that are not necessarily male-dominated (e.g., service sector professions), but for which other barriers exist for poor young women. One way to do this would be to encourage TVET programs to expand course offerings to new fields, as the traditional menu of options tends to reinforce gendered occupational norms rather than create new opportunities for women. An expansion of TVET course offerings could also inform the related question of whether the skills constraint binds more severely on women when entering new fields, for example fields which might require literacy more than traditional occupations. Another option is to build evidence on effective strategies for recruitment, placement, and facilitation services. The third approach encourages women to enter new industries where no clear gender assignment has yet been made, as was done in the business process outsourcing industry in India (Jensen, 2012). An ongoing GIL evaluation to promote young men’s and women’s employment in the burgeoning BPO sector in Nigeria will offer some preliminary insight on this approach in Africa. Future research could go beyond testing specific interventions that implement one or more of these conceptual approaches, to examine the relative effectiveness of the approaches. In what context is one approach more desirable than another? Is it possible to screen young women to provide more tailored guidance on sector choice, for example by identifying personality traits associated with being early adopters of a new trade? What are the benefits and costs, including social costs and harassment, of the three approaches, and how are these benefits and costs reflected in the retention of women in the trade? How to promote entry into new trades? Future research could also examine the efficacy of specific mechanisms to promote women’s entry into new trades, such as encouraging women to enter non- traditional trades in groups to reduce harassment (as was done in the Liberia EPAG program, but not 34 rigorously tested), and creating peer support and mentorship opportunities (as suggested by research in Ethiopia and Uganda (Alibhai et al. 2015; Campos et al. 2014)). Could “edutainment” approaches, such as embedding female characters in non-traditional trades on popular radio or television shows, help to create role models and shift norms around appropriate work for women? Could men be engaged to create safer work environments? Testing such mechanisms can also elucidate the relative severity of the underlying constraints, such as lack of information, lack of role models, and threat of harassment, that give rise to occupational segregation. 4.5 Deepening our understanding of Networks and Safety Networks and safety are both areas in which much more research is needed to understand the nature and extent of these constraints. The design of interventions to address these constraints would subsequently build on this knowledge. How do networks help or hinder young women’s employment? There remains a major gap in our understanding of networks, and whether and how young men and women are constrained by their networks and discrimination. Networks and social support can arise from peer support, mentorship, and the fostering of family and community environments. It is commonly believed that women and youth have smaller networks, or that their networks are less likely to provide mentorship, information about jobs, or loans. However, there is little available research in Sub-Saharan Africa to support those claims. Research is similarly scarce for interventions targeting network building. Many training programs for young women have explicit measures to build social support (e.g. through peers, role models, girls-only safe spaces), but research rarely measures their impact in isolation. Several areas of research warrant further examination: Do young women use their networks to access resources (information, loans) differently from men or older women? What is the extent of discrimination against young women in the labor market? Do young women experience discrimination directly from employers or indirectly through referral? To what extent are young women pressured by family and kin networks to share their economic assets? Does such pressure discourage the accumulation of certain kinds of assets, and ultimately labor market entry? Ongoing GIL evaluations in South Africa are investigating the roles of skills certificates, referral letters, and better information about workers in the labor market, highlighting the importance skills and information/signaling for women in contrast to men. Initial analysis suggests that referral letters from past employers can reduce the gender gap in call-backs from potential employers (Abel et al., 2017). Another program in South Africa is attempting to increase the intensity of young men and women’s job search through peer support, SMS reminders, and job search action planning. GIL researchers are examining which interventions most effectively induce changes in behavior, thus illuminating constraints to job search. Initial findings suggest that action planning is most effective in improving job search intensity and employment outcomes. Are safety concerns limiting young women’s labor market choices? Another priority with regard to understanding constraints is to gain a more systematic understanding of how harassment affects young women’s labor market experiences on the way to work and in the workplace. Lacking population-level 35 data, it is difficult to know the prevalence of different forms of sexual harassment and gender-based violence in the workplace in Sub-Saharan Africa. We also do not know how young women’s expectations and previous experiences with sexual harassment constrain their choices about whether and which types of work to pursue, a topic well-suited for qualitative research. We also need to understand how domestic violence, which remains far too common in Africa, affects labor market participation and productivity.31 An evaluation of safe transportation is currently underway in Mexico City by the Bank’s Transport practice. The program, “Hazme el Paro”, supplements sex-segregated transportation with training for bus drivers and police, and a marketing campaign and smartphone application to encourage bystanders to take action. To study both the network and safety constraints, one option would be to examine the early career experiences of young men and women, and their long-term consequences. Analysis of existing long-run panel data, such as the Young Lives survey, could shed light on how networks and exposure to unsafe environments affect participation and choice of occupation in youth. Even better would be if, in conjunction with existing panel data, we could identify plausibly exogenous shocks that might have shifted job search mechanisms or the size or composition of networks, such as the introduction of mobile phones. Another option for exploring these constraints is through qualitative work with adults to ask about the decisions and circumstances that shaped their early career choices and how those have endured (or not) over time. 4.6 Cross-cutting research priorities: Scaling Up, Cost-effectiveness, Heterogeneity, and Measurement Do girl-friendly program designs work at scale? Along with the priorities outlined above, GIL’s current research agenda explores whether the girl-friendly design elements from previous evaluations can be integrated into larger co-educational youth employment programs. Such integration would expand the reach and sustainability of efforts to include girls in the youth employment agenda. Evaluating these larger-scale youth employment programs will determine whether these girl-friendly design elements perform as well at scale as they do in smaller pilot programs. Cost-Effectiveness. Measuring costs, with the goal of comparing them to the benefits measured in an impact evaluation, is particularly important for interventions in the youth employment space, which tend to be more expensive than other common development initiatives, such as cash transfers or public health interventions. In its forthcoming research, GIL will seek to apply best practice accounting methods to measure the core costs of interventions, including fixed and variable costs. To the extent possible, this will allow for benchmarking the costs of evaluated programs against other types of interventions. Heterogeneity: Approaches and Tradeoffs to Reaching the Most Vulnerable. A common tension in the design of youth employment programs is the desire to reach the most vulnerable or hard-to-reach youth (e.g., the youngest, poorest, least educated, or otherwise marginalized) while also setting and meeting high targets for the project’s employment impact. Projects that choose to focus on vulnerable groups often 31 Several studies in GIL’s Violence thematic area evaluate the impact of behavioral change interventions on intimate-partner violence. These explicitly study women’s socioeconomic outcomes as well. 36 struggle to successfully engage or solicit enough applications from these groups, and wind up serving those who are “better off” among the eligible group. Across GIL projects, efforts will be made to document operational lessons on how to effectively engage younger or more vulnerable girls. Where sample size permits, GIL evaluations will endeavor to examine the heterogeneous impacts of employment programs on different types of young women, disaggregating results by age, education, and poverty. Gathering more evidence on these types of heterogeneous impacts can directly assist project teams in weighing the trade-offs of different eligibility criteria and offering the most appropriate intervention to each target group. Additionally, the general equilibrium effects of targeting certain groups are often unknown and not studied. Knowledge of whether youth-centric interventions displace older workers or increase consumption or available investment funds, for example, can help policy makers include helpful safeguards or scale up programs effectively. Another area for further research is how social norms and institutions interact with interventions to promote human capital or labor market outcomes. For example, Ashraf et al. (2014a) found that for ethnic groups that traditionally make bride price payments at marriage, the increased supply of schools in Zambia during the 1990s resulted in a significant increase in female education, but not for ethnic groups that do not make bride price payments. Future research will need to pay attention to such heterogeneity when exploring, for example, the relationship between employment and fertility, as well as the impact of programs. Improving measurement of non-cognitive skills and women’s empowerment. Through its wide-ranging portfolio of research on women’s economic empowerment in Africa, GIL is uniquely positioned to improve comparative policy analysis by encouraging the use of similar survey questions and indicators, and creating an evidence base of costs and impacts to allow comparison across contexts. Whenever possible, GIL impact evaluations track a common set of outcomes across projects:  Employment and earnings;  Economic empowerment, including asset ownership and control over resources;  Social empowerment, including voice and agency;  Socio-emotional skills;  Networks;  Fertility preferences, behaviors, and outcomes;  Experience of harassment and gender-based violence;  Spillovers onto partners, siblings and children. Best-practice indicators already exist in the literature on women’s empowerment and youth employment for some, but not all of these outcomes. In particular, the development and use of locally-adapted measures of young women’s social empowerment and socio-emotional skills will serve as a cross-cutting activity across GIL’s impact evaluations on youth employment. A key challenge is that commonly used international scales for measuring such outcomes have not been validated in African settings. Where possible, qualitative research to elicit girls’ perspectives on their capacity to make and act on decisions regarding their lives will inform the development of new tools to measure empowerment. To further investigate psychological dimensions of empowerment and how they mediate or moderate labor market outcomes, we will collaborate with experts in psychology and public health to develop and validate 37 locally-adapted measures of socio-emotional traits and skills, such as emotional regulation, self-efficacy and self-esteem, motivation and perseverance. These measures not only allow us to examine the interactions between psychosocial well-being and employment interventions, and the next wave of research examining job quality for women in Africa, they serve as an important public good, as reliable measures of such constructs are rarely available in many African settings. 38 Table 3. Summary table of priority research questions surrounding young women’s employment Constraint Priority Research Questions Skills  Which types of skills are most important for obtaining productive employment?  What is the optimal sequencing of skills?  Is the provision of skills or cash more effective? Capital  How do young women access capital to start businesses? Time, Care,  How, when, and for which subgroups does employment affect young women’s and Family fertility? Formation  Does child care provision improve young women’s employment outcomes?  How do child care markets function in rural and urban contexts? How can we expand the availability, quality, and use of child care?  Does time availability increase through child care provision Occupational  What trades should be promoted among young women, and in what socio-economic Choice contexts?  Which mechanisms will most effectively promote entry into new trades, and what are the strongest constraints here? Networks  How do networks help or hinder young women’s employment? and Safety  Are safety concerns and experiences limiting young women’s labor market choices and productivity?  Is mobility a constraint to employment in Sub-Saharan Africa? Cross-  Which constraints are the most binding, and for which target groups and contexts? cutting Unbundling comprehensive programs.  Do girl-friendly program designs work at scale?  Approaches and tradeoffs to reaching the most vulnerability: take-up and general equilibrium effects of various target groups.  Improving measurement of non-cognitive skills and women’s employment. 39 5. References Abel, M., R. Burger, P. Piraino. (2017). The Value of Reference Letters- Experimental Evidence from South Africa. Mimeo. Adoho, F., Chakravarty, S., Korkoyah, D. T., Lundberg, M. K., & Tasneem, A. (2014). The impact of an adolescent girls employment program: the EPAG project in Liberia. World Bank Policy Research Working Paper, (6832). AfDB, OECD., & UNDP (2012). African Economic Outlook 2012. Alibhai, S., Buehren, N., & Papineni, S. (2015). Female entrepreneurs who succeed in male dominated sectors in Ethiopia. Gender Innovation Lab Policy Brief 12. Angeles, G., Guilkey, D.K., & Mroz, T.A. (1998). Purposive program placement and the estimation of family planning program effects in Tanzania. Journal of the American Statistical Association, 93 (443):884–899. Ashraf, N., Bau, N., Nunn, N., & Voenna, A. (2014a). Bride price and the returns to education for women. Working paper, Harvard Business School. Ashraf, N., Field, E., & Leight, J. (2014b). Contraceptive access and fertility: The impact of supply-side interventions. Working paper, Harvard Business School. Aterido, R., Beck, T., & Iacovone, L. (2013). Access to finance in sub-Saharan Africa: Is there a gender gap? World Development, 47: 102–20. Attanasio, O., Guarín, A., Medina, C., & Meghir, C. (2015). Long Term Impacts of Vouchers for Vocational Training: Experimental Evidence for Colombia. National Bureau of Economic Research, No. w21390. Attanasio, O., Kugler, A., & Meghir, C. (2011). Subsidizing vocational training for disadvantaged youth in Colombia: Evidence from a randomized trial. American Economic Journal, Applied Economics 3 (3), 188-220. Attanasio, O. & Vera-Hernandez, M. (2004). Medium- and long-run effects of nutrition and child care: Evaluation of a community nursery programme in rural Colombia. Working Paper Series, Institute for Fiscal Studies. Azevedo, T. A. D., Davis, J., & Charles, M. (2013). Testing what works in youth employment: evaluating Kenya's Ninaweza program. Austrian, K., & Muthengi, E. (2014). Can economic assets increase girls' risk of sexual harassment? Evaluation results from a social, health and economic asset-building intervention for vulnerable adolescent girls in Uganda. Children and Youth Services Review, 47, 168-175. Baird S., McIntosh, C.E., Ozler, B. (2009). The short-term impacts of a schooling conditional cash transfer program on the sexual behavior of young women. World Bank, Impact Evaluation Series, No. 40. Baird, S., McIntosh, C.E., & Ozler, B. (2011). Cash or condition? Evidence from a cash transfer experiment. Quarterly Journal of Economics, 126 (4), pp. 1709-1753. Baird, S., Chirwa, E., de Hoop, J., & Ozler, B. (2014). Girl power: Cash transfer and adolescent welfare. Evidence from a cluster-randomized trial in Malawi. NBER Working Paper 19479. Bandiera, O., Buehren, N., Burgess, R., Goldstein, M., Gulesci, S., Rasul, I. & Sulaiman, M. (2015). Women’s empowerment in action: Evidence from a randomized control trial in Africa. Working paper, UCL. Banerjee, A., Duflo, E., & Glennerster, R., & Kinnan, C. (2015). Miracle of microfinance? Evidence from a randomized evaluation. American Economic Journal: Applied Economics, 7, 22-53. 40 Bardasi, E., & Wodon, Q. (2010). Working long hours and having no choice: Time poverty in Guinea. Feminist Economics, 16(3), 45-78. Barro, R. & Lee, J.W. (2010). A new data set of educational attainment in the world, 1950-2010. Journal of Development Economics, 104, pp.184-198. Beaman, L., Keleher, N., & Magruder, J. (2013). Do job networks disadvantage women? Evidence from a recruitment experiment in Malawi. Working Paper, Department of Economics, Northwestern University. Beaman, L., Duflo, E., Pande, R., & Topalova, P. (2012). Female leadership raises aspirations and educational attainment for girls: a policy experiment in India. Science. 335(6068): 582-586. Berlinski, S. & Galiani, S. (2007). The effect of a large expansion of pre-primary school facilities on preschool attendance and maternal employment. Labour Economics, 14 (3): 665–80. Berlinski, S., Galiani, S., & McEwan, P.J. (2009). Preschool and maternal labor market outcomes: Evidence from a regression discontinuity design. Economic Development and Cultural Change 59 (2): 313–44. Bernard,T., Dercon, S., Orkin, K., & Taffesse, A. S. (2014). The Future in Mind: Aspirations and Forward-Looking Behaviour in Rural Ethiopia. Blattman, C., & Annan, J. (2014). Can employment reduce lawlessness and rebellion? A field experiment with high-risk men in a fragile state. A Field Experiment with High-Risk Men in a Fragile State (July 12, 2014). Blattman, C., Fiala, N., & Martinez, S. (2013). Generating skilled self-employment in developing countries: Experimental evidence from Uganda. The Quarterly Journal of Economics, 129(2), 697-752. Blattman, C., Green, E., Jamison, J., & Annan, J. (2014). The returns to cash and microenterprise support among the ultra-poor: a field experiment. Available at SSRN 2439488. Blattman, C., Jamison, J. C., & Sheridan, M. (2015). Reducing crime and violence: Experimental evidence on adult noncognitive investments in Liberia. National Bureau of Economic Research, No. w21204. Blattman, C., & Ralston, L. (2015). Generating employment in poor and fragile states: Evidence for labor market and entrepreneurship programs. World Bank Manuscript.Bloom, D. E., Canning, D., Fink, G., & Finlay, J. E. (2009). Fertility, female labor force participation, and the demographic dividend. Journal of Economic Growth, 14(2), 79-101. Borghans, et al. (2008). The economics and psychology of personality traits. Journal of Human Resources, 43(4). Brenton, P., Bucekuderhwa, C.B., Hossein, C., Nagaki, S., & Ntagoma, J.B. (2011). Risky business: Poor women cross-border traders in the great lakes region of Africa. World Bank Africa Trade Policy Notes #11. British Council. (2016). Universities, employability and inclusive development: Repositioning higher education in Ghana, Kenya, Nigeria and South Africa. British Council. Buvinić, M., & Furst-Nichols, R. (2014). Promoting women's economic empowerment: What works? The World Bank Research Observer, lku013. Campos, F., Goldstein, M., McGorman, L., Boudet, A.M., & Pimhidzai, O. (2015). Breaking the metal ceiling: Female entrepreneurs who succeed in male-dominated sectors in Uganda. Policy Research working paper; no. WPS 7503. Washington, D.C. : World Bank Group. Canning, D., Raja, S., & Yazbeck, A. (2015). Africa’s demographic transition: dividend or disaster. World Bank and Agence Francaise de Developpement. 41 Chattopadhyay, R. & Duflo, E. (2004). Women as policy makers: Evidence from a randomized policy experiment in India. Econometrica, 72(5), 1409-1443. Chigunta, F., Chisup, N., & Elder, S. (2013). Labour market transitions of young women and men in Zambia. International Labour Organization Work4Youth Publication Series No.6. Cho, Y. & Honorati, M. (2014). Entrepreneurship programs in developing countries: A meta regression analysis. Labour Economics, 28, 110-130. Cho, Y., Kalomba, D., Mobarak, A. M., & Orozco, V. (2013). Gender differences in the effects of vocational training: Constraints on women and drop-out behavior. World Bank Policy Research Working Paper, (6545). Chong, A., Gonzalez-Navarro, M., Karlan, D., & Valdivia, M. (2013). Effectiveness and spillovers of online sex education: Evidence from a randomized evaluation in Colombian public schools. National Bureau of Economic Research. (No. w18776). Corno, L., & Voena, A. (2015). Selling Daughters: Age of Marriage, Income Shocks and Bride Price Tradition. Rockwool Foundation Research Unit, No.84. Debpuur, C., Phillips, J.F., Jackson, E.F., Nazzar, A., Ngom, P., & Binka, F.N. (2002). The Impact of the Navrongo Project on contraceptive knowledge and use, reproductive preferences, and fertility. Studies in Family Planning 33(2):141–164. Demirguc-Kunt, A. & Klapper, L. (2015). Measuring financial inclusion: The global findex database. Retrieved from http://data.worldbank.org/data-catalog/financial_inclusion. Demographic and Health Surveys (various). (2008-2015). ICF International. Retrieved from www.statcompiler.com. Desai, J. & Tarozzi, A. (2011). Microcredit, family planning programs and contraceptive behavior: evidence from a field experiment in Ethiopia. Demography 48(1):749–782. DeVries, H., Eggers, S., Jinabhai, C., Meyer-Weitz, A., Sathiparsad, R, & Taylor, M. (2014). Adolescents’ beliefs about forced sex in KwaZulu-Natal, South Africa. Archives of Sexual Behavior. Dickson, K. & Bangpan, M. (2012). Providing access to economic assets for girls and young women in low-and-lower middle-income countries: A systematic review of the evidence. University of London. Doepke, M. & Tertilt, M. (2014). Does female empowerment promote economic development? National Bureau of Economic Research. (No. w19888). Duflo, E., Dupas, P., & Kremer, M. (2014). Education, HIV, and early fertility: Experimental evidence from Kenya. MIT, working paper. Retrieved from http://economics.mit.edu/files/6951. Dupas, P. & Robinson, J. (2013). Savings constraints and microenterprise development: Evidence from a field experiment in Kenya. American Economic Journal: Applied Economics. 5(1), pp. 163-192. Elder, S. & Kone, K. (2014). Labour market transitions of young women and men in sub-Saharan Africa. International Labour Organization Work4Youth Publication Series No.9. Elder, S. & Kring, S. (2016). Young and female- a double strike? Gender analysis of school-to-work transition surveys in 32 developing countries. International Labour Organization Work4Youth Publication Series No.32. Engebretson, S. (2013). Evaluation of Filles Éveillées (Girls Awakened): A pilot program for migrant adolescent girls in domestic servitude. Population Council. Erulkar, A. & Muthengi, E. (2009). Evaluation of Berhane Hewan: A Program To Delay Child Marriage in Rural Ethiopia. International Perspectives on Sexual and Reproductive Health. 42 Erulkar A, Medhin G. (2014). Evaluation of Health and Education Impacts of a Girls’ Safe Spaces Program in Addis Ababa, Ethiopia. Population Council. Esteve-Volart, B. (2004). Gender discrimination and growth: theory and evidence from India. London School of Economics and Political Science, LSE Library. (No. 6641). Fafchamps, M., McKenzie, D., Quinn, S., & Woodruff, C. (2014). Microenterprise growth and the flypaper effect: evidence from a randomized experiment in Ghana. Journal of Development Economics, 106: 211–26. Fawole, O. I., Ajuwon, A. J., Osungbade, K. O., & Faweya, O. C. (2003). Interventions for violence prevention among young female hawkers in motor parks in south-western Nigeria: A review of effectiveness. African Journal of Reproductive Health, 71-82. Fawole, O. I., Ajuwon, A. J., & Osungbade, K. O. (2005). Evaluation of interventions to prevent gender- based violence among young female apprentices in Ibadan, Nigeria. Health Education, 105(3), 186-203. Feigenberg, B., Field, E., & Pande, R. (2013). The economic returns to social interaction: Experimental evidence from microfinance. The Review of Economic Studies, 80(4), 1459-1483. Field, E., Jayachandran, S., Pande, R., & Rigol, N. (2015). Friendship at work: Can peer effects catalyze female entrepreneurship? National Bureau of Economic Research. (No. w21093). Filmer, D. & Fox L. (2014). Youth Employment in Sub-Saharan Africa. World Bank and Agence Française de Développement. Retrieved from https://www.wdronline.worldbank.org/handle/10986/16608 License: CC BY 3.0 IGO. Fox, L., Kibbuka, K., Munoz, J.H., Sohnesen, T.P. (2011). Small is Big: Development of Household Enterprise Sector in Ghana. A quantitative look at household enterprises. World Bank, processed. Gajigo, O. & Hallward-Driemeier, M. (2010). Constraints and opportunities for new enterpreneurs in Africa. World Bank Working Paper. Gershoni, N., & Low, C. (2015). The impact of extended reproductive time horizons: Evidence from Israel’s expansion of access to IVF. IZA Working Paper. Gertler, P., Heckman, J., Pinto, R., Zanolini, A., Vermeersch, C., Walker, S., & Grantham-McGregor, S. (2014). Labor market returns to an early childhood stimulation intervention in Jamaica. Science, 344(6187), 998-1001. Gitter, S. R. & Barham, B. L. (2008). Women's power, conditional cash transfers, and schooling in Nicaragua. The World Bank Economic Review, 22(2), 271-290. Goldin, C. (1994). The U-shaped female labor force function in economic development and economic history. National Bureau of Economic Research. (No. w4707). Goldin, C. & Katz, L. (2002). The power of the pill: Oral contraceptives and women's career and marriage decisions. Journal of Political Economy, 110(4): 730-770. González-Velosa, C., Ripani, L. & Rosas-Shady, D. (2012). ¿Cómo mejorar las oportunidades de inserción laboral de los jóvenes en América Latina? Banco Interamericano de Desarrollo, Unidad de Mercados Laborales y Seguridad Social, Nota Técnica IDB-TN-345. Retrieved from http://idbdocs.iadb.org/wsdocs/getdocument.aspx?docnum=36976183. Grimm, M., Gubert, F., Koriko, O., Lay, J., & Nordman, C. J. (2013). Kinship ties and entrepreneurship in Western Africa. Journal of Small Business & Entrepreneurship, 26(2), 125-150. Guengant, J. P., & May, J. F. (2013). African demography. Global Journal of Emerging Market Economies, 5(3), 215-267. 43 Hallward-Driemeier, M. (2013). Enterprising women, expanding economic opportunities in Africa. Africa Development Forum Series, 78299. Handa, S., Peterman, A., Huang, C., Halpernc, C., Pettifor, A., & Thirumurthy, H. (2015). Impact of the Kenya cash transfer for orphans and vulnerable children on early pregnancy and marriage of adolescent girls. Social Science and Medicine, 141, p. 36-45. Haushofer, J., & Shapiro, J. (2013). Household response to income changes: Evidence from an unconditional cash transfer program in Kenya. Massachusetts Institute of Technology. Heckman J. & Kautz, T. (2012). Hard evidence on soft skills. Labour economics, 19(4), 451-464. Heckman J., Stixrud, J., & Urzua, S. (2006). The effects of cognitive and noncognitive abilities on labor market outcomes and social behavior. Journal of Labor Economics, 24(3), 411–482. Hicks, J. H., Kremer, M., Mbiti, I., & Miguel, E. (2015). Vocational education in Kenya: evidence from a randomized evaluation. International Initiative for Impact Evaluation (3IE). Hirshleifer, S., McKenzie, D., Almeida, R., & Ridao-Cano, C. (2014). The impact of vocational training for the unemployed: experimental evidence from Turkey. The Economic Journal. Holla, A., Leonard, S., & Wilson, W. (2010). Secret lives of economically-empowered women. Informal World Bank brief. Hsieh, C. T., Hurst, E., Jones, C. I., & Klenow, P. J. (2013). The allocation of talent and US economic growth National Bureau of Economic Research. (No. w18693). Honorati, M. 2015. The Impact of Private Sector Internship and Training on Urban Youth in Kenya. World Bank Policy Research Working Paper No. 7404. Ibarrarán, P., Kluve, J., Ripani, L., & Rosas, D. (2015). Experimental Evidence on the Long-Term Impacts of a Youth Training Program. IZA. Ibarrarán, P. & Shady, D.R. (2009). Evaluating the impact of job training programmes in Latin America: Evidence from IDB funded operations. Journal of Development Effectiveness 1.2 : 195-216. International Labour Organization. (2015). Global Employment Trends for Youth. International Labour Office. International Labour Organization. (2014). Labour market transitions of young women and men in sub- Saharan Africa. International Labour Office. International Labour Organization. (2013a). Key Indicators of the Labor Market, 8th Edition. International Labour Office. Retrieved from databank.worldbank.org. International Labour Organization. (2013b). Global Employment Trends for Youth 2013. International Labour Office. International Labour Organization. (2012). Global Employment Trends for Women 2012. International Labour Office. International Youth Foundation. (2014). Strengthening life skills for youth: A practical guide to quality programming. International Youth Foundation. Jensen, R. (2012). Do labor market opportunities affect young women's work and family decisions? Experimental evidence from India. Quarterly Journal of Economics. Keats, A. (2014). Women’s schooling, fertility, and child health outcomes: Evidence from Uganda’s free primary education program. Mimeo, Wesleyan University. Klugman, J., Hanmer, L., Twigg, S., Hasan, T., McCleary-Sills, J., & Santamaria, J. (2014). Voice and agency: Empowering women and girls for shared prosperity. World Bank Group. Kotikula, A., Purnik, M., & Robertson, R. (2015). Interwoven: how the better work program improves job and life quality in the apparel sector. World Bank Group 99729. 44 LISGIS. (2011). Report of the Liberia Labor Force Survey 2010. Liberian Institute of Statistics and Geo- Information Services. Lippman, L. H., Ryberg, R., Carney, R., & Moore, K. A. (2015). Workforce connections: Key 'soft skills' that foster youth workforce success: toward a consensus across fields. Child Trends, Inc. Lloyd, C. B., & Young, J. (2009). New lessons: the power of educating adolescent girls: a Girls Count report on adolescent girls. The Population Council. Lutalo, T., Kigozi, G., Kimera, E., Serwadda, D., Wawer, M.J., Zabin, L.S., & Gray, R.H. (2010). A randomized community trial of enhanced family planning outreach in Rakai, Uganda. Studies in Family Planning 41(1):55–60. Magruder, J. R. (2010). Intergenerational networks, unemployment, and persistent inequality in South Africa. American Economic Journal: Applied Economics, 2(1), 62-85. Maitra, P. & Mani, S. (2012). Learning and Earning: Evidence from a Randomized Evaluation in India. Fordham University, Department of Economics, Discussion Paper Series, 2. Martinez, S., Naudeau, S., & Pereira, V. (2012). The promise of preschool in Africa: A randomized impact evaluation of early childhood development in rural Mozambique. World Bank and Save the Children. Retrieved from http://siteresources.worldbank.org/INTAFRICA/Resources/The_Promise_of_Preschool_in_Afric a_ECD_REPORT.pdf McKenzie, D. & Woodruff, C. (2013). What are we learning from business training and entrepreneurship evaluations around the developing world? The World Bank Research Observer, lkt007. Miller, G. (2010) Contraception as development? New evidence from family planning in Colombia. The Economic Journal, 120, 709-736. Murphy, E., & Carr, D. (2007). Adolescent girls education and delayed childbearing. Population Reference Bureau. Nordman, C. J., Rakotomanana, F., & Roubaud, F. (2012). Informal versus Formal: A Panel Data Analysis of Earnings Gaps in Madagascar. Nordman, C. J., Robilliard, A. S., & Roubaud, F. (2009). Decomposing gender and ethnic earnings gaps in seven West African cities. IRD. Nordman, C. J., & Wolff, F. C. (2009). Islands through the glass ceiling? Evidence of gender wage gaps in Madagascar and Mauritius. Labor Markets and Economic Development, 25, 521-544. Nordman, C. & Vaillant, J. (2014). Inputs, gender roles or sharing norms? Assessing the gender performance gap among informal entrepreneurs in Madagascar. IZA DP No. 8046. Osili, U.O. & Long, B.T. (2008). Does female schooling reduce fertility? Evidence from Nigeria. Journal of Development Economics, 87(1): 57-75. Ozier, O. (2011). The impact of secondary schooling in Kenya: A regression discontinuity analysis. Mimeo, UC Berkeley. Pike, K. & Godfrey, S. (2015). Lesotho’s apparel industry: What is the role for better work? Unpublished. Retrieved from http://betterwork.org/global/wp-content/uploads/Session-8- Lesotho%E2%80%99s-Apparel-Industry.pdf. Pop-Eleches, C. (2010). The supply of birth control methods, education, and fertility evidence from Romania. Journal of Human Resources, 45(4), 971-997. Porter, G. (2008). Transport planning in sub-Saharan Africa II : putting gender into mobility and transport planning in Africa. Progress in Development Studies., 8 (3). pp. 281-289. 45 Porter, G., Hampshire, K., Abane, A., Tanle, A., Esia-Donkoh, K., Obilie Amoako-Sakyi, R., ... & Asiedu Owusu, S. (2011). Mobility, education and livelihood trajectories for young people in rural Ghana: a gender perspective. Children's Geographies, 9(3-4), 395-410. Premand, P., Brodmann, S., Almeida, R., Grun, R., & Barouni, M. (2016). Entrepreneurship Education and Entry into Self-Employment Among University Graduates. World Development, 77, 311-327. Ruiz Abril, M.E. (2008). Liberia: Girls vulnerability assessment. Government of Liberia, Nike Foundation, and World Bank. Retrieved from https://docs.google.com/document/d/ 1FvEgQNMsZHVfMLYxfp62DUly8nY1850lh2vGmiYHk_0/ edit?hl=en_US&pli=1. Salon, D., & Gulyani, S. (2010). Mobility, poverty, and gender: travel ‘choices’ of slum residents in Nairobi, Kenya. Transport Reviews, 30(5), 641-657. Sivasankaran, A. (2014). Essays on gender, intra-household allocation and development. Sneeringer, S.E. (2009). Fertility Transition in Sub-Saharan Africa: A Comparative Analysis of Cohort Trends in 30 Countries. DHS Comparative Reports 23. Todd, P. (2012). Effectiveness of interventions aimed at improving women's employability and quality of work: A critical review. World Bank Policy Research Working Paper, (6189). UN Habitat. (2013). State of women in cities 2012-2013: Gender and the prosperity of cities. United Nations Human Settlements Programme. UN Women. (2012). Informal cross border trade report. UN Women. Retrieved from http://www.unwomen.org/wp-content/uploads/2012/03/ICBTreport_Liberia.pdf Uteng, T. (2012). Gender and mobility in the developing world. World Development Report. Vyas, S., Mbwambo, J., & Heise, L. (2015). Women's paid work and intimate partner violence: Insights from Tanzania. Feminist Economics, 21(1), 35-58. World Bank. (2009). Unleashing the potential of Ethiopian women: Trends and options for economic empowerment. World Bank Report No. 49366. World Bank. (2012). World development report: Gender equality and development. World Bank Publications. World Bank. (2015). World Bank Databank Gender Statistics. Retrieved from http://data.worldbank.org/data-catalog/gender-statistics. World Bank. (2016). Enterprise Surveys. Retrieved from http://www.enterprisesurveys.org. 46