Interventions promoting adolescent girls' economic capabilities: what works? A rapid evidence review

Stavropoulou, Maria Research Report Interventions promoting adolescent girls' economic capabilities: what works? A rapid evidence review Provided in Cooperation with: ODI Global, London Suggested Citation: Stavropoulou, Maria (2018) : Interventions promoting adolescent girls' economic capabilities: what works? A rapid evidence review, ISBN 978-1-912942-00-8, Overseas Development Institute (ODI), London This Version is available at: https://hdl.handle.net/10419/193656 Standard-Nutzungsbedingungen: Die Dokumente auf EconStor dürfen zu eigenen wissenschaftlichen Zwecken und zum Privatgebrauch gespeichert und kopiert werden. Sie dürfen die Dokumente nicht für öffentliche oder kommerzielle Zwecke vervielfältigen, öffentlich ausstellen, öffentlich zugänglich machen, vertreiben oder anderweitig nutzen. Sofern die Verfasser die Dokumente unter Open-Content-Lizenzen (insbesondere CC-Lizenzen) zur Verfügung gestellt haben sollten, gelten abweichend von diesen Nutzungsbedingungen die in der dort genannten Lizenz gewährten Nutzungsrechte. Terms of use: Documents in EconStor may be saved and copied for your personal and scholarly purposes. You are not to copy documents for public or commercial purposes, to exhibit the documents publicly, to make them publicly available on the internet, or to distribute or otherwise use the documents in public. If the documents have been made available under an Open Content Licence (especially Creative Commons Licences), you may exercise further usage rights as specified in the indicated licence. https://creativecommons.org/licenses/by-nc-sa/4.0/ Interventions promoting adolescent girls’ economic capabilities: what works? A rapid evidence review Maria Stavropoulou October 2018 Disclaimer This document is an output of the Gender and Adolescence: Global Evidence (GAGE) programme which is funded by UK aid from the UK government. However, views expressed and information contained within do not necessarily reflect the UK government’s official policies and are not endorsed by the UK government, which accepts no responsibility for such views or information or for any reliance placed on them. Acknowledgements The author would like to thank very much Sara Stevano for designing and implementing the search strategy and providing valuable research support. She would also like to thank Sarah Baird and Rachel Marcus for their comments on the draft report, Kathryn O’Neill for editorial support, and Jojoh Faal Sy for design and typesetting support. Table of Contents Executive summary i 1 Introduction 1 1.1 Setting the scene 1 1.2 Review aim, questions, methodology and limitations 6 1.3 Overview of reviewed interventions 6 2 Financial education and/or assets interventions 9 2.1 Overview 9 2.2 Economic outcomes 10 2.3 Non-economic outcomes 13 2.4 Discussion of findings and evidence gaps 16 3 Vocational and/or business skills training interventions 19 3.1 Overview 19 3.2 Economic outcomes 20 3.3 Non-economic outcomes 27 3.4 Discussion of findings and evidence gaps 33 4 Integrated interventions with an economic component 36 4.1 Overview 36 4.2 Economic outcomes 37 4.3 Non-economic outcomes 40 4.4 Discussion of findings and evidence gaps 45 5 Addressing issues of age-specific intervention effectiveness, duration and intensity, and legacy effects 48 5.1 Timing of intervention and age-specific effectiveness 48 5.2 Intervention duration and intensity to maximise effectiveness 49 5.3 Standalone or integrated interventions 50 5.4 Intervention components most supportive of positive change 50 5.5 Legacy effects 52 6 Conclusions 54 6.1 Key findings 54 6.2 Knowledge and evidence gaps 55 References 57 Annex 1: Search strategy 61 Annex 2: Key characteristics of intervention groups 63 Annex 3: Overview of interventions 65 Annex 4: Overview of secondary studies/reviews 117 Interventions promoting adolescent girls’ economic capabilities: what works? Figures Figure 1: Involvement in household chores by gender and age 2 Figure 2: Youth labour force participation and gender gaps by region and gender, 1997–2017 2 Figure 3: Youth access to formal and informal financial services by region and gender (%) 3 Figure 4: The three key dimensions of empowerment 5 Figure 5: Regional distribution of interventions 7 Figure 6: Structure of financial education and/or assets interventions 10 Figure 7: Structure of vocational and/or business skills training interventions 20 Figure 8: Structure of integrated interventions with an economic component 37 Figure A2.1: Regional distribution of interventions 63 Figure A2.2: Regional distribution of interventions 63 Figure A2.3: Regional distribution of interventions 64 Tables Table 1: Interventions examined in this review 8 Table A1.1: Search strings 61 Table A1.2: Inclusion and exclusion criteria 62 Table A2.1: Evaluation methodology 63 Table A2.2: Target group by location 63 Table A2.3: Target group by age 63 Table A2.4: Target group by type of vulnerability 63 Table A2.5: Evaluation methodology 63 Table A2.6: Target group by location 63 Table A2.7: Target group by age 64 Table A2.8: Target group by type of vulnerability 64 Table A2.9: Evaluation methodology 64 Table A2.10: Target group by location 64 Table A2.11: Target group by age 64 Table A2.12: Target group by type of vulnerability 64 Table A3.1: Overview of interventions, evaluation methodology and key findings 65 Table A4.1: Secondary studies/reviews 115 Boxes Box 1: Gender and Adolescence: Global Evidence (GAGE) programme 5 Box 2. Key review questions 6 AGEP Adolescent Girls Empowerment Programme AGI Adolescent Girls Initiative AGI-K Adolescent Girls Initiative – Kenya AIDS Acquired Immune Deficiency Syndrome BALIKA Bangladeshi Association for Life Skills, Income, and Knowledge for Adolescents CEDPA Centre for Development and Population Activities CRS Catholic Relief Services DISHA Development Initiative Supporting Healthy Adolescents ELA Empowerment and Livelihood for Adolescents EPAG Economic Empowerment of Adolescent Girls and Young Women FGD Focus Group Discussion FGM/C Female Genital Mutilation/Cutting GAGE Gender and Adolescence: Global Evidence HIV Human Immunodeficiency Virus ICT Information and Communications Technology IDI In-depth Interview ILO International Labour Organization IMAGE Intervention with Microfinance for AIDS and Gender Equity KII Key Informant Interview KYEP Kenya Youth Empowerment Project LMIC Low- and middle-income country MCI Making Cents International MENA Middle East and North Africa NGO Non-governmental Organisation RCT Randomised Control Trial SEWA Self Employed Women Association SHAZ Shaping the Health of Adolescents in Zimbabwe SILC Savings and Internal Lending Community SoFEA Social and Financial Empowerment of Adolescents SRH Sexual and Reproductive Health SSSPVAG Safe and Smart Savings Products for Vulnerable Adolescent Girls STI Sexually Transmitted Infection TESFA Towards Improved Economic and Sexual/Reproductive Health Outcomes for Adolescent Girls TRY Tap and Reposition Youth UNESCO United Nations Educational, Scientific and Cultural Organization UNICEF United Nations Children’s Fund USAID United States Agency for International Development WWB Women’s World Banking Acronyms i Interventions promoting adolescent girls’ economic capabilities: what works? Executive summary Investing in adolescent girls and intervening to empower them at this pivotal time of their life and help them realise their full potential has become a priority in development policy and practice. Given their persistent economic disadvantage and inequality of opportunity, interventions for the economic empowerment of adolescent girls in low- and middle-income countries (LMICs) are vital to reduce youth unemployment, poverty and vulnerability, and enable adolescent girls to improve their capabilities, transition into a productive adulthood and contribute to economic growth and social development. While the importance of these interventions is well-understood, what works to strengthen girls’ economic capabilities is much less well-known. This rapid evidence review focuses on interventions that target adolescent girls in LMICs with economic strategies that seek to promote their economic capabilities and empower them. Based on 74 primary studies yielded by literature searches, the review examined 57 interventions, grouped into three categories: those that provided financial education and/or assets as their core component; those that provided vocational and/ or business skills training interventions; and integrated interventions that provided a combination of support services, including at least one economic component such as financial or skills training. All three groups of interventions demonstrated a promising and multiple potential to improve girls’ economic capabilities. In line with their stated objectives, most interventions increased girls’ access to employment (most often self-employment), enabled them to generate their own income, improved their financial literacy, helped them develop good savings habits and increase their formal and informal savings, and facilitated their access to and ownership of assets, be it poultry, small livestock or jewellery. Apart from improving girls’ economic capabilities, these interventions also led to positive changes in other interlinked and mutually reinforcing capability domains, indirectly further improving girls’ economic advancement: they thus enabled girls to invest in further skills training, improve their sexual and reproductive health, better protect themselves against violence, expand their social networks, enhance their self-esteem and voice, increase their mobility, and develop more gender-egalitarian attitudes, including towards female employment. Evidence also identified certain components that can play a critical role in enabling girls’ participation and contributing to positive effects. These include: enabling peer groups to meet in safe spaces; using well-trained and trusted female mentors; offering life skills training; mobilising families and communities to create an enabling environment; prioritising safety, including locating training venues close to girls’ homes; holding activities at flexible and convenient times; providing stipends to meet basic participation costs; and offering childcare arrangements for young mothers. However, some critical knowledge and evidence gaps remain: • More evidence is needed to clarify the relationship between girls’ economic empowerment and their vulnerability to violence and early marriage, as well as their agency and civic participation. • While early adolescence provides a better window of opportunity to intervene, most of the reviewed interventions – especially those providing vocational or livelihoods training – targeted older adolescent girls. A more nuanced analysis considering the intersection of age, location and education level could provide better evidence. • Girls from marginalised groups are difficult to reach and girls with disabilities are almost invisible in the evidence base. • Little is known about the optimal duration and intensity of interventions. • While integrated interventions have more promising potential, more evidence is needed to identify which components work best and most effectively together, shedding light on how the provision of several components can multiply and reinforce economic outcomes and lead to girls’ broader empowerment. • Evaluations tend to focus on measuring what changes or not, but not why and for how long. Methodologically rigorous longitudinal evaluations are scarce and intervention legacy effects remain largely unknown. ii • Despite its critical role, the quality of training provided – for participants and for trainers – is also underresearched. • The role of context is also neglected and there is a need for more evidence on how successful interventions can be replicated in other settings. • There is much less evidence from some regions, such as the Middle East and North Africa (MENA), and about successful strategies for strongly conservative cultural settings, wherein girls face pervasive disadvantage. 1 Interventions promoting adolescent girls’ economic capabilities: what works? 1.1 Setting the scene Adolescence is a critical period of human life during which girls and boys undergo dramatic physical, cognitive and emotional changes and build their capabilities, aspirations and goals for the rest of their lives (UNICEF, 2011). However, in many developing countries, the onset of puberty entails heightened risks and vulnerabilities for adolescent girls. In settings where poverty and gender inequality intersect, low educational attainment, economic insecurity, limited sexual and reproductive health (SRH) knowledge, social isolation, early marriage and motherhood, exposure to violence, and weak voice and agency, all have detrimental and long-lasting effects on girls’ wellbeing and hinder their transition into a healthy, safe and productive adulthood (Hallman, 2016). Investing in girls and intervening to empower them at this pivotal time and help them realise their full potential has become a priority in development policy and practice. Not only does empowering girls have intrinsic value as an issue of human dignity and social justice, but it is also important to effectively address the intergenerational transmission of poverty and contribute to economic growth, also known as the ‘girl effect dividend’ (UNICEF, 2011; Chaaban and Cunningham, 2011). Interest in adolescent girls’ economic empowerment is gaining momentum. The latest data paints an alarming picture of girls’ persistent economic disadvantage and inequality of opportunity compared to their male peers: they are involved in hazardous child labour, have to spend considerable amounts of their time on household chores, and face greater difficulty to attend training, participate in economic activities when they reach the legal working age, and access financial services. Globally, there are 64 million girls aged 5–17 years in child labour, including 17 million girls aged 12–14 and 13.5 million aged 15–17. Of that total, 27.7 million are engaged in hazardous work, including 6 million younger girls and 13.6 million older girls. While girls account for 42% of all children in child labour, the International Labour Organization (ILO) recognises that these figures understate the extent of girls’ work as girls may be more likely to be doing less visible and thus under-reported forms of child labour such as domestic work. Indeed, many girls engaged in child labour are likely to work in the services sector (ILO, 2017a). Girls are also much more likely than boys to perform household chores at home – a form of work not included in child labour estimates. Data shows that girls account for two-thirds of the 54 million children aged 5–14 years who spend at least 21 hours weekly on household chores – the threshold that is suggested to have a negative impact on a child’s ability to attend and benefit from school. Girls account for a similar share of the 29 million children (aged 5-14) who perform chores for more than 28 hours weekly, and of the 7 million who perform chores for 43 hours or more weekly (see Figure 1). Furthermore, girls are more likely than boys to have ‘double work duty’, that is, doing both household chores and other forms of employment (ibid.). Although the global youth population has grown in the past decade, their participation in the labour force has declined. Older adolescents and youth of legal working age (15-24 years) face high unemployment, lowquality jobs, longer and more insecure school-to-work transitions, and increased detachment from the labour market. Older adolescent girls and young women face greater disadvantage and exclusion than their male peers. In 2017, the labour force participation rate for young men aged 15–24 was 54% compared to 37% for young women, with larger and persistent gender gaps in South Asia and the Middle East and North Africa (MENA) (see Figure 2). Most young women and men who are employed work in the informal economy; in low-income countries, 19 out of 20 working youth are in informal employment. In low- and middle-income countries (LMICs), nearly 17% of young workers (nearly 70 million, most of them in sub-Saharan Africa) live on incomes below the extreme poverty threshold of $1.90 per day. If one includes those living in moderate poverty, then 39% of youth in these countries are working poor – young men slightly more so than women, although in some regions and countries young female workers are more disadvantaged (ILO, 2017b). Girls and young women aged 15–24 also account for the majority of youth who are neither in employment nor in education or training (NEET). They comprise 3 out of every 4 young NEETs, with the gender gap increasing in 1 Introduction 8 Table 1: Interventions examined in this review Financial education and/or assets interventions 1. Adolescent Girls Empowerment Programme (AGEP) 2. Adolescent Girls Initiative – Kenya (AGI-K) 3. Intervention with Microfinance for AIDS and Gender Equity (IMAGE) 4. Ishaka 5. Learning Games for Adolescent Girls and their Mothers 6. Mia 7. Pilot Financial Literacy Project for Adolescent Girls 8. Safe and Smart Savings Products for Vulnerable Adolescent Girls (SSSPVAG) 9. Siyakha Nentsha 10. Social and Financial Empowerment of Adolescents (SoFEA) 11. Suubi Research Programme 12. Tap and Reposition Youth (TRY) Savings and Micro-Credit for Adolescent Girls 13. Temuulel 14. YouthSave 15. YouthStart Vocational and/or business skills training interventions 1. Adolescent Girl Initiative (AGI), Haiti 2. Adolescent Girl Initiative (AGI), Rwanda 3. Adolescent Girl Initiative (AGI), South Sudan 4. Adolescent Livelihood and Reproductive Health Programme 5. Akazi Kanoze 6. Economic Empowerment of Adolescent Girls and Young Women (EPAG) 7. Employment Fund and Adolescent Girls Employment Initiative (AGEI) 8. Employment and Livelihood for Adolescents (ELA), Bangladesh 9. Empowerment and Livelihood for Adolescents (ELA), Tanzania 10. Empowerment and Livelihood for Adolescents (ELA), Uganda 11. Jóvenes en Acción 12. Juventud y Empleo 13. Kenya Youth Empowerment Project (KYEP) 14. Kenya Youth Empowerment project/Ninaweza 15. Neqdar Nesharek 16. Por un Manana 17. Programa Joven 18. ProJoven 19. SATYA-Pratham Programme 20. SEWA-livelihoods skill building intervention for adolescent girls 21. Technical Education and Vocational Education and Training Authority (TEVETA) programme 22. Technical and Vocational Vouchers Programme 23. Towards Improved Economic and Sexual/Reproductive Health Outcomes for Adolescent Girls (TESFA) 24. Youth Opportunities Programme (YOP) 25. 100 Hours to Success Integrated interventions with an economic component 1. Bangladeshi Association for Life Skills, Income, and Knowledge for Adolescents (BALIKA) 2. Berhane Hewan 3. Better Life Options 4. Biruh Tesfa 5. Development Initiative Supporting Healthy Adolescents (DISHA) 6. Empowering Adolescent Girls project 7. Filles Éveillées 8. Ishraq 9. Kishori Abhijan 10. Kishoree Konta 11. Meseret Hiwott 12. Out-of-School Adolescents Support project 13. Powering Up Biruh Tesfa 14. Rajiv Gandhi Scheme for Empowerment of Adolescent Girls-SABLA (RGSEAG-SABLA) 15. SHAZ! (Shaping the Health of Adolescents in Zimbabwe) 16. Security for Girls through Land project 17. Support for Orphans and Vulnerable Children Affected by HIV/AIDS 9 Interventions promoting adolescent girls’ economic capabilities: what works? 2.1 Overview Given the increasing importance that financial products and services have in people’s lives in LMICs, development actors see various potential benefits in programmes that target vulnerable youth, especially adolescent girls and young women, and offer them some form of financial education and/or assets (WWB, 2014b; Sebstad, 2011). Access to finance may be desirable per se to expand the financial inclusion agenda, or to achieve youth development outcomes via financial and economic empowerment (Deshpande and Zimmerman, 2010; Hirschland, 2009). For instance, several studies have investigated the role of microfinance in reducing HIV infection rates among adolescent girls (USAID, 2008). Programmes included in this group provided financial education and/or financial assets as either standalone interventions (as is the case with some youth microfinance programmes) or as integrated interventions that offer financial education and/or a financial product as their main component, combined with several other support services. These programmes thus have two types of objective: (1) to promote adolescent girls’ access to and effective use of financial services; and/or (2) to provide adolescent girls with financial means that have the potential to create benefits that are not only financial or economic, but also improve girls’ broader wellbeing. In this group, we included 15 interventions and 20 evaluation studies. All these programmes provided adolescent girls with some form of financial training and/ or a savings account or participation in a microfinance scheme. Most interventions also combined at least one financial component with other activities involving peer groups and mentors, training in life skills, SRH or livelihoods, and community awareness (see Figure 6). A summary of the key characteristics of these interventions and evaluation studies is provided in Annexes 2 and 3. Two-thirds of programmes were evaluated using a combination of quantitative and qualitative methods, while nearly one-third used an RCT and one-fifth a quasiexperimental design. Most programmes were implemented in sub-Saharan Africa (particularly in Kenya and Uganda), and targeted girls in both urban and rural areas, with most also targeting adolescent girls as a broad group (without distinguishing between younger and older adolescents). Although all programmes targeted disadvantaged and low-income girls, only a few paid attention to groups of girls with particular vulnerabilities by explicitly including out-of- school, married or orphaned girls. 2 Financial education and/or assets interventions Key points • Fifteen interventions provided financial education and/or financial assets to adolescent girls. They include both youth microfinance programmes and integrated interventions that combined a core financial component with other support services. • Almost all evaluations reported some positive changes in girls’ financial and economic capabilities: improved financial literacy, increased savings and use of formal and informal financial services, increased access to credit, involvement in income-generation activities and ownership of assets. • Interventions also led to positive changes in other capability areas: girls invested their savings or income in education, improved their SRH knowledge and practices, enhanced their self-confidence, increased their mobility, and developed more gender-egalitarian attitudes. • Interventions that combined financial literacy with a savings product demonstrated stronger financial effects, while life skills training enabled girls to better protect themselves from gender-based violence. • Participants’ age and location may have a considerable effect, yet there is mixed evidence on whether younger adolescent girls or rural girls can benefit more from this intervention type. 10 2.2 Economic outcomes We first present the direct programme outcomes in terms of girls’ financial and economic capabilities, which include enhanced financial knowledge, increased savings and use of formal financial services but also informal groups for saving purposes, increased access to credit and involvement in economic activities, income generation and ownership of assets (such as poultry and small livestock). The Tap and Reposition Youth (TRY) programme in Kenya targeted out-of-school girls and young women aged 16–22 in low-income and slum areas of Nairobi with high HIV prevalence, and provided them with group-based microfinance. As only a small number of older participants wanted to access microloans for their business, the initial microfinance approach (with its focus on credit and loan repayment) evolved to accommodate younger girls’ needs, including their ability to save money in a safe, accessible place and not at home where it could be found and misused by parents or partners. Thus TRY introduced a voluntary savings scheme for younger, more vulnerable and risk-averse girls, along with peer groups with mentors, life skills and business training. At end line, participants were more likely to have savings and to have significantly larger savings than the control group. They were also significantly more likely to save in the bank compared to the baseline and the control group. While girls aged 16–19 had greater savings and incomes compared to the control group, this difference was only marginally significant. Those older than 20 years showed stronger effects as they were significantly more likely than those in the control group to have more savings, larger incomes and more household assets, and to keep their savings in a financial institution (Erulkar and Chong, 2005). A few years later, the Safe and Smart Savings Products for Vulnerable Adolescent Girls (SSSPVAG) in Kenya and Uganda combined the safe spaces model with financial products for low-income clients. The programme targeted girls aged 10–19 years in urban slums, giving them access to peer groups, female mentors, financial education and SRH information. Girls were also assisted to open a savings account, which they could operate independently of their parents. At end line, most participants answered financial literacy questions correctly. Although participants in Kenya were less likely to report having a savings plan compared to baseline and the control group, they significantly improved their savings behaviour and the use of bank services compared to the control group. Overall, more than twothirds of participants in Kenya and Uganda saved money using not only their account but through a mix of formal and informal means (Austrian and Muthengi, 2013). However, financial effects were stronger in Uganda, with significantly more participants having a saving plan, using a bank service and saving money at end line compared to the control group. They were actually more than twice as likely to have saved any money in the previous six months and eight times more likely to have saved using a formal method. The evaluation noted that this country difference may be related to participants’ age difference, as most Kenyan girls were young adolescents, while half of Ugandan participants were older than 14 years (Austrian and Muthengi, 2013). Programme design in Uganda also enabled the comparison of girls who only opened a savings account with girls who also participated in groups and received financial education. While all girls significantly reduced the use of informal savings methods, those in groups were significantly more likely to have better financial knowledge and greater improvement in saving (Austrian and Muthengi, 2014). Using a similar model, the Adolescent Girls Empowerment Programme (AGEP) in Zambia targeted vulnerable girls aged 10–19 years in urban and rural areas with three distinct interventions: the core intervention (basic package) consisting of safe spaces groups, mentors, life skills training, SRH information and financial education; Financial education and/or assets interventions Financial education Savings product Microfinance Peer groups and mentors Life skills training Sexual and reproductive health training Livelihoods training Community awareness activities Figure 6: Structure of financial education and/or assets interventions 11 Interventions promoting adolescent girls’ economic capabilities: what works? the basic package plus a health voucher; and the basic package plus health voucher and a savings account. Findings from the midline showed positive financial effects for all participants compared to the control group. However, girls in the third intervention arm, which included all activities, had higher financial literacy scores, and those who opened an account – even if they did not use it much – were more likely to save both in that formal account and informally, and in higher amounts compared to those without an account. Interestingly, evidence indicated that younger adolescent girls in rural areas, who were also more engaged in the programme, were significantly more likely to open an account (Austrian et al., 2016). Similarly, the Adolescent Girls’ Initiative – Kenya (AGI-K) supported girls aged 11–14 years in the urban slums of Nairobi and in the north-east of the country with four interventions: school assistance; girls’ groups in safe spaces; community conversations for violence prevention; and financial education combined with assistance in opening a bank account (in urban areas) or a home bank (to use at home). Preliminary findings showed that girls who received the financial education intervention reported improved financial literacy, and a notable increase in savings and financial responsibility (Muthengi et al., 2016). The Ishaka (meaning ‘Courage for the future’) project in Burundi targeted in-school and out-of-school girls and young women aged 14–22 years in urban and rural areas with financial education and microfinance, vocational training, SRH education and life skills training; it also included community awareness-raising and advocacy activities to create a supportive environment for participants, who were mostly aged 15–19. Ishaka enabled girls to improve their money management skills, save and access credit in order to engage in income-generating activities. However, it did not link them to microfinance institutions which could provide services more tailored to their particular financial needs. In line with AGEP midline findings, Ishaka was also more successful in rural areas, where the non-governmental organisation (NGO) involved had an established presence and participants had fewer distractions affecting their recruitment and participation. The evaluation also found that while out-of-school girls were mostly interested in income-generation skills, school girls were more attracted to the life skills training (Rushdy, 2012). Within South Asia, the Social and Financial Empowerment of Adolescents (SoFEA) programme provided rural Bangladeshi girls and young women aged 11–21, both married and unmarried, with girls’ clubs, financial literacy, livelihoods and life skills training, microfinance and community sensitisation. Participants were encouraged to save a minimum amount weekly, while those aged 15 and over could access a loan on condition that they saved more for a month. Participants reported having improved their financial literacy between baseline and end line, yet they were afraid of taking loans and not being able to repay them successfully. They also improved their planning skills and increased their involvement in economic activities and their ownership of assets such as poultry and small livestock. The programme was also found to have spill-over effects as non-participating girls were influenced by those who did participate and also got involved in productive activities (Kamruzzaman et al., 2012). Nevertheless, other programmes had rather limited effects. Implemented in five Indian states, the Learning Games for Adolescent Girls and their Mothers project targeted poor women, members of self-help groups and their adolescent daughters, providing health, social and financial education to improve their wellbeing. While participation improved girls’ confidence about their ability to manage money and their appreciation of savings, it did not increase their actual savings as girls were not provided with a mechanism to save. Qualitative research revealed that about half of girls who participated noted that they had not saved because they did not have their own money. Girls were more interested in health than financial education – a finding that the evaluation attributed to the fact that participants could easily put their health knowledge into daily practice, while they had limited opportunities to do so with the financial knowledge they gained (Gray and Chanani, 2010). In the state of West Bengal, the financial education had minimal effects and participants did not demonstrate statistically significant differences in their savings-related knowledge, attitudes and behaviours, partly due to the deteriorating financial context (Spielberg et al., 2010). Apart from programmes targeting only adolescent girls (along with young women in some cases), a number of evaluated programmes have provided disadvantaged youth with a combination of support services, including financial education and savings accounts. The Suubi (meaning ‘Hope’) research programme in rural Uganda provided adolescents aged 12–16 who had been orphaned due to AIDS with counselling and educational material, financial education, peer mentors and a matched child savings account to be used for post-primary education. Both boys and girls were able to save comparable amounts 12 in their bank accounts and no statistically significant gender differences were identified (Ssewamala et al., 2010). The Siyakha Nentsha (meaning ‘Building with young people’) programme in peri-urban KwaZulu-Natal, South Africa, targeted school-going adolescents and offered life skills training and health and financial education delivered by recent school graduates. Compared to the control group, participating boys and girls were more likely to have improved their budgeting and planning skills and to have attempted to open (and succeeded in opening) a bank account. They were also significantly more likely to discuss financial decision-making, looking for work, saving, social grants and starting a business (Govender et al., n.d.). Moreover, boys who received the financial education were more likely to report undertaking an income-generating activity between baseline and end line compared to those who received only the life skills and health education (Hallman and Roca, 2011). Older adolescent girls can also participate in microfinance programmes targeting women. For instance, the Intervention with Microfinance for AIDS and Gender Equity (IMAGE) project in rural Limpopo province, South Africa, provided women aged 18 years and older with microfinance and training on issues including sexuality, HIV, gender norms and domestic violence. As a result, participants increased their household assets, expenditures and membership in informal savings groups (Kim et al., 2007). There are also a growing number of initiatives that target low-income youth or girls and primarily seek to promote their financial inclusion. The Temuulel (meaning ‘Aspire’) programme in Mongolia provided school girls aged 14–17 from low-income families with financial education and a savings product. Participants showed higher levels of knowledge of banking and savings than the control group, and by opening an account and depositing their saved money, succeeded in increasing their financial autonomy, while those in the control group were more likely to spend or lose their savings or have them stolen. However, participants did not increase their ability to plan for the future, either financially or otherwise. Interestingly, girls who received both the financial education and the account showed slightly higher levels of financial knowledge, saved more and more frequently, and opened more accounts compared to the control group and to girls who only received a savings account. They also developed teamwork skills, which were highly valued in the workplace but not acquired at school. The potential of financial education to improve the effects of a savings account is emphasised by the fact that girls who received both the training and the account exhibited equal or even slightly higher levels of financial management and savings than the savings-only girls who were one year older, and the majority had summer jobs that offered them some independent income. On the other hand, the latter were also more likely to contribute to their household income and to do so in larger amounts – mostly to meet daily needs (Tower and McGuiness, 2011). The Mia (meaning ‘Mine’) project in the Dominican Republic targeted low-income girls and young women aged 7–24 with two savings products – one for those aged 7-15 and another for older youth – and a financial education programme run through schools. According to research undertaken by the participating bank, most account-holders were girls under the age of 16, and most depended on their parents to make the transactions although they could legally do so themselves; in addition, most girls lived far from bank branches and had to spend time and money on public transport to get to them. A small-scale assessment focused on two groups of participating girls aged 10–15 living in the capital city: those who were successfully saving in their accounts and those who were not. The study found that almost all girls were saving through various mechanisms (formal and informal), including hiding money at home. Younger girls aged 10–12 reported having more longer-term savings goals, while those aged 12–15 saved more for short-term consumption; girls who saved for the future aimed to cover their university expenses linked to a better life, while those who reported saving for emergencies were concerned with family welfare (WWB, 2014a). Overall, girls who used the Mia account showed more self-control, had a clearer vision about the future and were better able to set savings goals. Successful savers acknowledged the difficulty involved in saving money, yet they also expressed confidence and optimism about their ability to do so and deposit their money in the bank where it could not be withdrawn without parental approval and presence. They also reported discussing their savings habit and trying to influence their peers. In line with the bank’s research findings, the ability to save in the bank account was indeed shaped by distance to the bank branch and girls’ dependency on parents. All those who saved in the bank depended on their parents to make a deposit and had parents who regularly visited the bank for their own needs. Girls who did go to the bank to deposit savings 13 Interventions promoting adolescent girls’ economic capabilities: what works? were actually living close to a bank branch. The study also found that most girls who saved were members of femaleheaded households, with lower income, stronger savings attitudes and expectations that children should contribute to household expenses. Parents emphasised that as girls tend to be more responsible than boys and think more about household needs, they saved more (WWB, 2014a). A similar financial initiative in Haiti targeted vulnerable girls aged 10–14 with financial education along with hygiene and protection training, provided by mentors in safe spaces. However, young girls aged 8–11 struggled to understand the material. Overall, participants started to save and manage their money more effectively, asking their parents for money less often and even contributing to household expenses occasionally; some girls also joined savings groups. Most parents reported that girls shared what they had learnt with them and in some cases their daughters helped them improve the household budget or their small business account (MCI and Save the Children, 2012). The YouthSave programme was a large savings initiative that targeted low-income youth aged 12–18 years in Colombia, Ghana, Kenya and Nepal, offering financial education and account opening services. Its evaluation concluded that when offered safe and affordable accounts, youth will open savings accounts. Once again, participants younger than 13 years saved more than older youth, partly because they withdrew less. When parents were the co-signatory, youth saved significantly more. No gender differentials in savings behaviour were identified; girls saved as much as boys and sometimes even more. The greater obstacle girls faced was not saving in itself, but access to financial institutions. The situation was even more difficult for out-of-school youth, including girls, who were harder to reach and perhaps needed more tailored products and services to ensure their inclusion in formal financial services (Johnson et al., 2015). YouthStart was another large multi-country financial inclusion programme for youth aged 12–24 years in eight sub-Saharan African countries,2 which provided them with financial education and services (mainly savings accounts and credit). Participants valued the financial training most, followed by savings, and reported that the programme improved their financial capabilities and enabled them to save more and generate more income. Most reported that programme services had a positive impact on their economic situation and life, and asked 2 Burkina Faso, Democratic Republic of Congo, Ethiopia, Malawi, Rwanda, Senegal, Togo and Uganda. for more entrepreneurship training, refresher sessions and mobile banking services. Adolescent girls and young women were a key target group and accounted for half of all clients reached. However, most found it hard to meet programme targets, and overall YouthStart lacked a clear strategy to reach out to them. In particular, involving adolescent girls was a major challenge for most financial service providers, while NGOs with a gender focus had more effective strategies for doing so. There were also problems in reaching (and promoting savings products among) rural populations (Microfinanza, 2015). 2.3 Non-economic outcomes Apart from directly improving girls’ financial and economic capabilities, several of the programmes also led to positive changes in other areas, including the remaining five GAGE domains (education, bodily integrity, physical health and nutrition, psychosocial wellbeing, and voice and agency). Education and learning The main education-related issue in the evaluations included in this review was whether programme participation had affected – positively or not – girls’ education. Indeed, participants in the SSSPVAG, especially in Kenya, commonly mentioned saving in order to cover education costs when their parents were unable to help them (Austrian and Muthengi, 2013). Ishaka participants in Burundi also used their earnings to cover their school costs (Rushdy, 2012). In Bangladesh, SoFEA girls reported that their school performance was not affected by programme participation; many girls actually pointed out that time constraints linked to studying hindered their regular attendance in SoFEA clubs, especially during school exams and harvest season (Kamruzzaman et al., 2012). In Kenya, AGI-K participants who received school assistance reported improved school enrolment and attendance (Muthengi et al., 2016). Bodily integrity and freedom from violence There is a large body of literature on the effects of microfinance programmes on women’s empowerment and wellbeing, including their vulnerability to violence. As the relation between the two is complex, evidence is inconclusive, with some studies suggesting that microfinance can reduce the risk of violence and others 14 arguing that it exacerbates it (Kim et al., 2007). Two of the evaluations3 contributed to the discussion and highlighted the importance of combining savings and credit with training for participants’ protection against violence. Girls in the SSSPVAG programme in Uganda who received the savings account along with participation in groups and training by mentors were able to protect themselves better against violence and exploitation compared to the control group. On the other hand, those who only opened savings accounts reported an increase in having been sexually harassed within the previous six months compared to the baseline and the control group (Austrian and Muthengi, 2014; 2013). The IMAGE research project in South Africa, which combined microfinance with training, explicitly aimed to address the issue of intimate partner violence. After two years, the levels of intimate partner violence experienced by participants in the past year had declined by more than half, while those levels either stayed the same or increased in the control group. Participants also developed more progressive, non-tolerant attitudes towards violence against women and girls. Qualitative data indicated that the reduction was the result of several factors, including shifts in women’s attitudes towards the acceptability of violence, expectations about better treatment from partners, improved spousal communication, reduced tensions over household finances, confidence to leave abusive relationships, solidarity and support to women survivors, speaking openly about abuse, collective action (such as organising public marches), mobilisation of community groups and activities raising community awareness of violence and HIV, with women’s overall economic and social empowerment potentially contributing to the decline in violence (Kim et al., 2007). Most importantly, the study design explored programme effects among three distinct groups: participants who received only the microfinance services, those who received the microfinance services and the training, and the control group. IMAGE participants who had received the full package showed not only improvements in their economic and social wellbeing, but also a decrease in intimate partner violence compared to the control group. On the other hand, those who had received only the microfinance support improved their 3 The evaluation of AGI-K which compares four distinct interventions (violence prevention only; violence prevention and education assistance; violence prevention, education assistance and health; and violence prevention, education, health and financial education along with a savings product) is also expected to provide valuable evidence. economic wellbeing indicators compared to the control group, but they did not reduce their risk of intimate partner violence (Kim et al., 2009). Although participants were on average middle-aged women, they also included older adolescent girls; the findings provide invaluable evidence that needs to be considered, given that young IMAGE participants also reported higher levels of partner communication related to HIV and sexuality as a result of the programme (Pronyk et al., 2008). Physical health and nutrition In line with programme stated objectives and provision of reproductive health information, evaluations of three-fifths of programmes also investigated effects on participants’ SRH knowledge and practices. Most noted improved relevant knowledge and behaviours, but as evaluations tended to take place shortly after programme completion, the sustainability of the observed or self-reported change remains unknown. In Uganda, participants of SSSPVAG girls’ groups had significantly greater knowledge about HIV and contraception at end line compared to girls who only received the savings account. Some evidence also indicated that girls who saved regularly were less likely to engage in transactional sex compared to those with irregular savings (Austrian and Muthengi, 2013). Midline findings from AGEP in Zambia also showed that participants significantly increased their contraception knowledge, were more likely to use a condom during their first sexual experience, and were less likely to engage in transactional sex compared to the control group (Austrian et al., 2016). Preliminary findings from AGI-K also suggest improved knowledge about SRH issues, including menstrual hygiene and management, along with adopting a more balanced diet (Muthengi et al., 2016). In Burundi, not only did Ishaka participants improve their SRH knowledge but there was also a sharp increase in their contraceptive use and a decrease in unwanted pregnancies. Participants were also more willing to undertake regular voluntary testing for HIV and sexually transmitted infections (STIs) (Rushdy, 2012). After taking part in the Siyakha Nentsha programme in South Africa, girls reported being more confident to obtain a condom, while those already engaged in sexual intercourse reported 15 Interventions promoting adolescent girls’ economic capabilities: what works? using condoms more consistently than those in the control group (Govender et al., n.d.). On the other hand, boys who took part in the programme reported sexual abstinence or having fewer sexual partners than the control group (Hallman and Roca, 2011). The IMAGE research project in South Africa, which combined microfinance with HIV training, contributed to a reduction in HIV-risk behaviour among female participants aged 14–35. This finding was attributed to increased levels of spousal communication about sexual issues and HIV, lower likelihood of having unprotected sex with a non-spousal partner and increased likelihood of using voluntary counselling and testing than the control group (Pronyk et al., 2008). The Learning Games for Girls intervention in West Bengal, India, which also provided health education sessions, significantly increased participants’ HIV knowledge, attitudes and prevention behaviour. Nonetheless, the evaluation noted that several questions about sexual behaviour had to be removed due to community sensitivity about that issue. In addition, when the intervention was delivered to mixed groups of mothers and daughters, participants expressed difficulty engaging in discussions, while those in girls-only groups admitted feeling comfortable and free among their peers to talk about SRH issues. The evaluation also found that the programme did not significantly improve the health, nutrition and hygiene of participants compared to the control group due to financial constraints to consume foods high in protein and iron, and to purchase soap (Gray and Chanani, 2010). In some cases, programme effects were largely insignificant. At end line, participants in the TRY programme in Kenya were less knowledgeable about SRH issues than the control group. And although they were more likely to be able to negotiate condom use and refuse sex, condom use was not statistically significantly higher compared to the control group (Erulkar and Chong, 2005). Boys in the Suubi programme in Uganda reported a significant decrease in approval of risky sexual behaviours at end line compared to the control group, yet scores remained unchanged for girls (Ssewamala et al., 2010). Midline findings from AGEP in Zambia also showed no change in rates of STIs or unwanted pregnancies and in age of sexual initiation (Austrian et al., 2016). Psychosocial wellbeing The evaluations of eight programmes noted that participation enabled adolescent girls to break their social isolation, to access a safe space in which they made friends and were supported to acquire new skills, and to improve their self-esteem. Indeed, half of participants in the SSSPVAG in Kenya and Uganda reported having made a new friend, having a place to meet friends and having a female mentor who provided guidance and counselling (Austrian and Muthengi, 2013). Likewise, Ishaka participants in Burundi increased their number of friends and had a space to talk about personal experiences and access support (Rushdy, 2012). Similar experiences were reported by participants in the Temuulel project in Mongolia (Tower and McGuiness, 2011), in AGEP in Zambia (Austrian et al., 2016), and in AGI-K in Kenya (Muthengi et al., 2016). A group with special vulnerabilities – married girls in rural areas – appreciated that the SoFEA programme in Bangladesh enabled them to socialise and interact with peers; girls also enjoyed the opportunity to participate in recreational activities. Participants also noted improved relationships with family, relatives and community members (Kamruzzaman et al., 2012). In South Africa, girls who received the full Siyakha Nentsha intervention (combining health with financial education) reported greater levels of social inclusion in their communities than girls who received the standard life skills training (Hallman and Roca, 2011). Participants in these programmes also reported increased self-esteem. Most girls in the SSSPVAG programme in Kenya and Uganda showed high levels of self-esteem at end line and were optimistic about the future, and feeling happier and safer in their neighbourhoods. An interesting finding was that girls who felt confident deciding about their money were also significantly more likely to say they had hope for the future (Austrian and Muthengi, 2013). Girls in AGEP in Zambia also reported increased confidence and reduced social isolation, with younger adolescent girls in rural areas showing stronger impact, potentially as they had higher participation rates (Austrian et al., 2016). Girls and boys in Siyakha Nentsha in South Africa were more likely to be optimistic about the future and to have increased their selfesteem compared to the control group, with boys showing a greater and statistically significant increase compared to girls (Govender et al., n.d.). Participants in IMAGE in South Africa (Kim et al., 2007) and AGI-K in Kenya (Muthengi et al., 2016) also reported increased self-confidence. Girls in Ishaka in Burundi significantly reduced their engagement in sex work or begging and thus enhanced their sense of dignity and self-respect (Rushdy, 2012). 16 Voice and agency The evaluations of 11 programmes also investigated effects on girls’ voice and agency. They reported positive changes in girls’ communication skills and confidence to interact and speak up, their mobility, their relationships and their social status in their households and communities. Participants in two programmes were also able to develop more gender-egalitarian attitudes. Participation in SSSPVAG in Kenya was associated with increased mobility and independence. Compared to baseline, at end line, more girls reported being able to attend group meetings or visit a local clinic alone, and fewer girls had to seek permission before looking for a job. Older girls aged 15–19 were significantly more likely to be independent than girls aged 10–14 years. Having a savings account and knowing how to manage money also made girls feel less vulnerable and more independent, as they could refuse sexual advances from men and meet their own and their household’s needs without having to depend on sexual relationships (Austrian and Muthengi, 2013). Always in Kenya, girls in AGI-K who participated in safe spaces reported improved communication skills, including their ability to express themselves and state their wish not to marry early but to continue their education (Muthengi et al., 2016). Boys and girls in Siyakha Nentsha in South Africa increased their ability to express their ideas to others compared to the control group, with boys showing a statistically significant increase. Boys and girls alike were significantly more likely to have discussed self-esteem, attitudes and values, sexuality, violence and sexual abuse, and gender relations than the control group (Govender et al., n.d.). Participants in the Ishaka project in Burundi improved their communication and conflict resolution skills, with subsequent improvements in their family and community relationships. As girls generated their own income and contributed to household needs, they also increased their household status (Rushdy, 2012). Girls in the SoFEA programme in Bangladesh also reported that their participation in groups improved their communication skills and enabled them to overcome their shyness and express their opinion. Parents and husbands agreed that participation in group activities led to positive changes as girls acquired new skills and knowledge, which they shared with their families and relatives, and made them value their girls more and believe in their abilities (Kamruzzaman et al., 2012). Girls in Temuulel in Mongolia reported that their participation in the savings project and the bank account made them feel they were no longer children but were growing up, as they were doing something associated with adulthood. Nearly all participants in the financial education classes talked to their parents about what they had learnt and received praise about their interest in financial issues, which further enhanced their self-esteem and raised their status within the family. Girls also discussed what they had learnt with their friends and some even tried to convince them to open their own account (Tower and McGuiness, 2011). Participants in Mia in the Dominican Republic also reported discussing their savings habit and trying to influence their peers (WWB, 2014a). Girls in the financial initiative in Haiti improved their communication skills and public speaking, and felt more confident speaking up and participating actively in school. Parents also commented that the project helped their daughters gain confidence and become more responsible (MCI and Save the Children, 2012). The IMAGE project in South Africa also enabled participants to increase their autonomy, improve their spousal communication and increase their status as they contributed to household needs. This is also the only evaluation in which participants (though only a small proportion of them were adolescent girls and young women) reported higher levels of participation in social groups and collective action, speaking in public and joining a range of social groups (formal and informal) (Kim et al., 2007). Participants in the IMAGE project also reported more gender-egalitarian attitudes (Kim et al., 2007). Similarly, TRY participants in Kenya reported more egalitarian views on three issues: that a wife should be able to refuse to have sex with her husband; that marriage is not the best option for an unschooled girl; and that having a husband is not necessary to be happy. Overall, at end line, girls who had participated in TRY scored marginally higher on the gender attitudes score compared to the control group, which had higher scores at baseline (Erulkar and Chong, 2005). Midline findings from AGEP in Zambia also showed no significant change in attitudes towards gender roles and gender-based violence (Austrian et al., 2016). 2.4 Discussion of findings and evidence gaps The reviewed evidence provides useful insights into the effectiveness of financial education and financial asset interventions targeting disadvantaged adolescent girls in LMICs. Almost all of the evaluation studies highlighted how these programmes can directly build or strengthen 17 Interventions promoting adolescent girls’ economic capabilities: what works? girls’ financial and economic capabilities by enhancing their financial knowledge, improving their money management skills, guiding them to develop good savings habits, increasing their savings and use of both formal and informal savings institutions, facilitating their access to credit and their involvement in income-generating activities, and increasing their income and ownership of assets such as small livestock. Those programmes also had positive effects in other capability areas. There is some evidence that girls from low-income households were able to use their savings or generated income to cover their school fees, while at the same time programme participation did not negatively affect their school attendance and performance. Participants were also able to improve their SRH knowledge and some related behaviours, and to better protect themselves against violence and exploitation – for instance, girls with savings were less likely to adopt risky sexual behaviours and engage in transactional sex. Programmes also had positive psychosocial effects, as girls felt confident about managing their money and setting goals, and optimistic about their future. They also reported becoming more independent and less vulnerable, overcoming their shyness, and beginning to interact and speak up, as well as increasing their mobility, enhancing their social status by contributing to household needs, and developing more gender-egalitarian attitudes. Interestingly, this review also found that combining financial education with a savings product and further training on a range of key issues affecting girls’ daily lives can clearly enhance the potential of such interventions to generate multiple positive effects. This finding confirms established understanding in current policy and research that integrated interventions are likely to be more effective than standalone ones. In particular, several of the studies reviewed here emphasised the need to combine financial training with a savings product. Girls who received group-based (in schools or in community-based girls’ groups) financial education along with a savings account demonstrated stronger savings effects compared to those who only opened an account, including greater financial knowledge and saving more regularly and in higher amounts (Austrian and Muthengi, 2014; Tower and McGuiness, 2011). On the contrary, providing financial education without a savings product appears to have more limited financial effects as it does not allow participants to put their newly acquired knowledge into practice (Austrian et al., 2016; Gray and Chanani, 2010). Given that girls’ financial vulnerability is closely linked to SRH vulnerabilities, programmes that included training on SRH were also quite successful in improving relevant knowledge and behaviours. Although evidence is still limited, compared to participants who only received a savings account or microfinance support, girls who also received training on SRH issues were less likely to resort to negative coping strategies and risky sexual behaviours, and more likely to take better care of their sexual health and even improve communication with their partners about sexual issues (Austrian and Muthengi, 2014; 2013; Kim et al., 2009; Pronyk et al., 2008). Support for the provision of training and integrated programmes is also based on the potential harmful consequences of interventions that provide girls and young women with access to savings or microfinance only, in the absence of any other support service. Indeed, some evidence indicates that participants who only opened savings accounts experienced increased levels of sexual harassment compared to girls who also received training (Austrian and Muthengi, 2014; 2013), whereas those who were only provided with microfinance continued to experience high levels of intimate partner violence compared to participants who also received training and reported that levels of violence declined significantly (Kim et al., 2009). Promoting girls’ financial capabilities while also preventing any unintended negative outcomes (such as increased violence) is critical for girls’ wellbeing and is an area where more research is urgently needed. Although most of the programmes did not consider participants’ age, location or socioeconomic status, a more nuanced analysis of outcomes could suggest that programmes may need to be tailored to the particular needs of targeted girls. Regarding age, some evidence suggested that standard microfinance interventions may not be suitable for adolescents, who could benefit more from peer groups and savings than access to credit for income-generating activities (USAID, 2008; Erulkar and Chong, 2005). In the case of savings interventions, one study found that a savings programme had stronger effects among older adolescent girls aged 15–19 than among younger girls aged 10–14 (Austrian and Muthengi, 2013). However, three other evaluations – including a study of a large-scale financial inclusion intervention in four countries – pointed out that younger (less than 13 years) adolescent girls showed better savings outcomes, as they saved more and withdrew less money (Austrian et al., 2016; Johnson et al., 2015; WWB, 2014a). In another study, younger girls who 24 to be employed than girls in control communities. Girls themselves reported that the programme enhanced their business skills. In general, girls get more involved in income-generating activities as they grow older and work more, yet the programme accelerated the transition into self-employment for participants (ibid.). While BRAC also implemented the programme in Tanzania, its evaluation did not find any significant employment effects there (Buehren et al., 2015). Similarly, the Neqdar Nesharek (meaning ‘We can participate’) programme in rural Upper Egypt targeted marginalised girls and women aged 16–29 and sought to empower them economically and socially. Nearly 40% of participants found employment; most started their own small business in their village as this activity enabled them to combine work with childcare responsibilities. Many graduates rejected employment opportunities outside the village and some accepted local jobs at lower pay. Unmarried girls with fewer household responsibilities who also wanted to save up for their marriage were more willing to get higher-paying jobs outside the locality. Despite planning, many participants were unable to start their business due to high start-up costs and inability to obtain low-interest loans locally (Ramadan et al., 2014). The Towards Improved Economic and Sexual/ Reproductive Health Outcomes for Adolescent Girls (TESFA) programme targeted married girls aged 14–19 years in rural Ethiopia, providing girls’ clubs, life skills training, access to health information and services, financial and livelihoods training, and community mobilisation. The programme included three intervention groups: girls in the first group received economic information and guidance based on an adapted Village Savings and Loan Association (VSLA) model, encouraging them to save and find incomegenerating activities; girls in the second group received SRH information and services; and girls in the third group received both components. While the proportion of girls in paid employment and income-generating activities increased across all three groups, girls who received the combined package showed the largest increase, while the control group recorded the lowest. Edmeades et al. (2014) concluded that TESFA was successful in economically empowering married girls, but also stressed that overall, there was no area in which the combined intervention 5 The Self Employed Women’s Association (SEWA) is a trade union of poor, self-employed women workers in India. 6 ILO (2003) notes that underemployment refers to worker’s time and willingness in relation to an alternative work situation in which this person is willing and available to engage in adequate employment for sufficient hours of work. group outperformed the economic empowerment group consistently; thus the combination of the two strategies did not apparently generate better outcomes. On the other hand, the combined curriculum may have improved both economic and health outcomes for participating girls. The Adolescent Livelihood and Reproductive Health programme in the urban slums of Allahabad, India, targeted girls aged 14–19 with peer groups, peer educators, reproductive health education, vocational training and savings accounts. While most girls who attended a vocational course reported that they used their skills after programme completion, only 10% were able to earn money as a result and girls did not increase the time spent on income-generating activities. The evaluation also found that the number of respondents expecting to work for pay in 10 years’ time declined from baseline to end line for both intervention and control groups, with more girls noting that youth experienced great difficulty to find a good job in the city. Qualitative research confirmed that girls had great difficulty in finding work and were often unable to assert themselves enough to get paid for what they did (Sebastian et al., 2005). A livelihoods programme initiated by the Self Employed Women’s Association (SEWA)5 in rural Gujarat, India, targeted in-school and out-of-school girls (married and unmarried) aged 13–19 years with girls’ groups (kishori mandals), life skills and vocational training, financial literacy, and opportunities to visit places such as banks, universities and places of historic interest to broaden girls’ horizons and experiences beyond their villages. The girls also received basic training in forestry, agriculture and animal husbandry and water management, and were encouraged to attend training in traditional and non-traditional trades in institutions outside the village to expand their livelihoods options. While participants were not interested in some of the programme topics, they wished to have more on livelihoods options and appreciated learning vocational skills (SEWA et al., 2006). Working hours As youth underemployment – the underutilisation of youth working capacities6 – is a large problem globally, evaluations of five programmes also examined whether vocational training not only enabled participants to 25 Interventions promoting adolescent girls’ economic capabilities: what works? access employment but also increased the number of hours they spent working. Indeed, both Jóvenes en Acción in Colombia and ProJoven in Peru increased hours worked weekly – the former by three hours and the latter especially for male participants (Attanasio et al., 2011; Nopo et al., 2007). The Youth Opportunities Programme in Uganda also increased participants’ working hours by 17% compared to the control group, with this increase almost entirely being in skilled trades (Blattman et al., 2014). The Employment Fund in Nepal increased the hours graduates spent on income-generating activities between 28% and 30% (that is, between 19 and 21 hours) monthly (Chakravarty et al., 2016a). And participants in the SATYA–Pratham project in India also worked 2.5 additional hours weekly, an effect that was sustained 18 months after programme completion (Maitra and Mani, 2014). Earnings Twelve programmes had positive effects on participants’ earnings. Evidence indicates that graduates of vocational and business training programmes are likely to improve their employment and business opportunities and increase their earnings compared to the control group. The Programa Joven in Argentina significantly increased earnings for young men and adult women, a finding which Aedo and Nuñez (2004) attributed to labour market conditions. Jóvenes en Acción in Colombia increased participants’ earnings as it enabled them to get better jobs. In particular, female participants increased their earnings by nearly 20% compared to the control group; women also increased their formal wages more than men (Attanasio et al., 2011). Similarly, the ProJoven programme in Peru increased average monthly earnings, with women benefiting more: Nopo et al. (2007) found that female participants increased their income by 93% compared to 11% for men. Diaz and Rosas (2016) noted that payroll data showed that women and young adults benefited more, while the follow-up survey indicated that men and adolescents earned more income. A previous evaluation of the programme found higher and statistically significant effects for female participants (who in almost all cases doubled their monthly income compared to the control group) and for adolescents (Diaz and Jaramillo, 2006). On the other hand, the Juventud y Empleo programme in the Dominican Republic led to a modest increase in participants’ monthly earnings compared to the control group. However, effects were larger for the bettereducated and those living in the capital city, who actually accounted for all the positive impact on earnings (Card et al., 2011). An evaluation six years after programme completion also found that it increased earnings for those in the capital city and had a statistically significant impact in the long run, particularly on earnings for women in the capital compared to those in the control group (Ibarrarán et al., 2015). Women in the SATYA–Pratham project in India earned 150% more monthly than those in the control group, with this effect being sustained 18 months after programme completion (Maitra and Mani, 2014). The Ninaweza programme in Kenya also increased participants’ weekly income, especially those who did not receive the life skills training (de Azevedo et al., 2013). The impact evaluation of the Youth Opportunities Programme in Uganda found that four years after completion, female graduates were earning incomes 73% higher than those of women in the control group. Male graduates showed a 29% increase in income compared to men in the control group (Blattman et al., 2014). Blattman et al. (2014) stressed that during this time, the gap between the cash earnings of women and men in the control group widened, as women’s earnings stagnated while men’s rose by nearly 50%. Thus, in the absence of the programme, women would have probably been unable to increase their earnings at all, while men would have continued doing so as their businesses grew. Another possible explanation for this finding is that group structure and periodic meetings provided members with support but also pressure to invest their grant and generate higher income. This is contrary to existing evidence that women are often forced to share their own resources with fathers and husbands and thus unable to invest them. KYEP in Kenya increased labour earnings for older male youth and those with higher education. On the contrary, less-educated women benefited more from the project compared to women with better education. Women in the tourism sector had higher earnings. Female trainees increased their wage earnings compared to the control group, but this finding should be interpreted cautiously due to a high attrition rate in the end line survey (Honorati, 2015). On the other hand, the Vocational Training Vouchers programme in Kenya did not increase participants’ overall earnings, though some evidence suggests that it did significantly increase hourly earnings for those in wage employment. Yet female trainees had lower earnings than male trainees (Hicks et al., 2015). 26 AGI in Rwanda doubled girls’ average earnings between baseline and end line (Botea et al., 2015). EPAG in Liberia also increased participants’ earnings by 80% compared to the control group. Older adolescent girls benefited more than young women, and those who attended the business skills training more than doubled their earnings (Adoho et al., 2014). The Employment Fund in Nepal also increased participants’ monthly earnings by 72%, including increasing the rate of those earning a monthly amount showing gainful employment. Those who attended beautician and tailoring courses had higher earnings than others in the Nepali labour market (Chakravarty et al., 2016a). Participants in the ELA programme in Uganda, and particularly those in self-employment, also increased their earnings between baseline and end line (Bandiera et al., 2015). Financial literacy and assets Most programmes also provided financial education and assets such as savings accounts or microcredit. The evaluations of 16 programmes examined their effects and several reported positive changes, helping participants to start or increase their savings (through formal and informal structures), start their own business, manage their budget successfully, and even increase their assets. The Neqdar Nesharek programme in Egypt helped participants open more than 3,800 savings accounts (Ramadan et al., 2014). Similarly, AGI in Rwanda assisted trainees to open individual savings accounts and save part of the daily stipend in their account, with savings considerably increasing over the course of the project (Botea et al., 2015). Participation in KYEP in Kenya increased the probability that participants opened a bank account in order to access their stipends, with female participants in particular starting to accumulate savings (Honorati, 2015). EPAG in Liberia also helped girls to open savings accounts at local banks. Thus trainees were significantly more likely to have savings and to save more than the control group (Adoho et al., 2014). Participation in the Akazi Kanoze programme in Rwanda was strongly associated with increased savings. Graduates were more likely to save and to save more frequently than the control group, with the primary aim to start a business (Alcid, 2014). The Employment Fund in Nepal also increased individual total savings (Chakravarty et al., 2016a). The microcredit services provided by ELA in Tanzania to older participants led to a significant increase in savings among girls as well as increased participation in informal savings groups for participants and nonparticipants in targeted communities (Buehren et al., 2015). The 100 Hours to Success intervention in Morocco aimed to enable participants to open savings accounts and access other financial services such as loans. Indeed, participants improved their financial knowledge and were more likely to open a savings account and to maintain it more than two years after the training was completed. However, there was no statistically significant increase in their savings or accessed loans compared to the control group. Thus Kluve et al. (2016) concluded that the programme improved financial literacy but not actual financial behaviour. Some evidence indicates that those aged 20 years and older benefited more from the financial literacy training. Effects were also stronger for better-off male participants, perhaps because they may already have been familiar with the use of financial services. On the other hand, female participants demonstrated stronger financial literacy results and were more likely to hold an independent savings account than men; although these findings were not statistically significant, gender differentials were smaller among participants compared to the control group. Kluve et al. stressed the role of traditional social norms, highlighting that the programme enabled women to realise the importance of having their own savings account. The evaluations of two other interventions also noted less positive outcomes. The impact evaluation of an apprenticeship programme targeting vulnerable youth in Malawi found that participation in training actually led to a decline in savings, particularly for female participants, who used them to meet training-related transport and accommodation costs (Cho et al., 2015). The livelihoods programme in the urban slums of Allahabad, India, also included opening a savings account for girls. Many girls succeeded in opening a savings account in their name at the local post office, yet they faced significant barriers to continue using those accounts. Post office staff were male and frequently expressed reluctance to work with adolescent girls; they often reprimanded girls for disturbing them and sometimes asked them to come back another day. Many girls were fearful of the staff and were reluctant to spend money on repeated trips to the post office to manage their account. Friends’ unwillingness to accompany them when they wanted to deposit money was cited as an additional barrier. Project staff acknowledged the problem and tried to improve communication between girls and post office staff and to sensitise the latter (Sebastian et al., 2005). 27 Interventions promoting adolescent girls’ economic capabilities: what works? The TESFA programme for married adolescent girls in rural Ethiopia enabled participants who had earned money to start saving part of their earnings; girls who had received livelihoods training and support were more likely to do so. Most importantly, girls in the livelihoods group and those who received the combined livelihoods and reproductive health training – i.e., the two groups which received financial training – significantly increased the use of savings for productive investment compared to girls who received only the SRH component. Girls in those two groups also considerably increased their access to loans; again, the largest increase was recorded among girls in the livelihoods group, followed by the group that received the combined package (Edmeades et al., 2014). AGI in Rwanda also assisted trainees to open cooperative accounts and encouraged them to use their deposits as start-up capital for the cooperatives (Botea et al., 2015). However, EPAG in Liberia did not improve access to credit for those girls who chose the business development track (Adoho et al., 2014). The financial literacy training provided by AGI in South Sudan enabled participating girls to improve their financial and business management and increase their financial independence. The programme also helped over 80% of participants to apply for a small business loan. However, some girls encountered problems with repayment and others had to hide money from their husbands to avoid having them misusing the money (Jurgens et al., 2013). The ELA programme in Bangladesh did not significantly improve girls’ financial literacy compared to the control group. Yet participants were more likely to take a loan. In particular, most girls who received the livelihoods training and a loan were able to maintain control over the money, use it and repay on time. On the contrary, some girls who received the microfinance support only (without skills training) gave the money to their parents or husbands. A potential negative relationship was also identified between schooling, loan intake and earning activities, as some adolescent girls – afraid that their involvement in income-generating activities would have a negative impact on their studies – did not use available loans to fund their education. Also, those from very poor households feared not being able to repay a loan, so reported no future plans to take one (Shahnaz and Karim, 2008). On the other hand, the Kenyan programme that provided vouchers for vocational training did not offer start-up capital to graduates. Thus, most of those who wanted to use their newly acquired skills and start their own business reported lack of capital as the main barrier, emphasising their inability to purchase the necessary tools for their trade (Hicks et al., 2015). In Uganda, participants in the Young Opportunities Programme used the grant to invest in tools and materials. Therefore, they increased the value of their business assets significantly compared to the control group. In particular, female participants increased their stocks more than 100% compared to women in the control group, while for men this increase was 50% compared to those in the control group (Blattman et al., 2014). Likewise, the SATYA–Pratham programme in India improved ownership of capital goods, with participants more likely to own a sewing machine 18 months after completion (Maitra and Mani, 2014). In Rwanda, many AGI participants reported using part of the daily stipend to purchase small animals, such that livestock ownership tripled. Girls also increased ownership of mobile phones. Owners of assets reported increased control over them at end line (Botea et al., 2015). Time use Two evaluations also assessed time use management. The livelihoods intervention in the slums of Allahabad, India, aimed to improve girls’ time management skills so that they would be able to engage in economic activities but also spend more time with friends. Sebastian et al. (2005) found that girls increased their leisure time between baseline and end line, but not the time spent on income-generating activities, partly due to difficulties translating their newly acquired skills into productive activities. The Neqdar Nesharek programme in Egypt also enabled participants to better manage their time as they had to combine attending the programme with their daily household tasks. Indeed, trainees realised that time is a valuable resource and they learnt how to plan and make best use of it (Ramadan et al., 2014). 3.3 Non-economic outcomes Apart from directly improving girls’ economic capabilities, most of these interventions (in line with their stated objectives and provided services) also led to positive changes in all five remaining GAGE domains: education, bodily integrity, physical health, psychosocial wellbeing, and voice and agency. Education and learning Participation in income-generating activities may have adverse effects on girls’ schooling. However, livelihoods 28 programmes targeting disadvantaged out-of-school youth often make them realise the importance of education and prompt them to return to school or invest more in skills training. Eight evaluated programmes examined educationrelated outcomes and reported mostly positive findings. Two years after the Youth Opportunities Programme ended in Uganda, Blattman et al. (2014) found that 68% of graduates were enrolled in vocational training compared to 15% of the control group, spending on average 340 hours more on training. No gender differentials were identified. However, men benefited more than women in two other programmes. The 100 Hours to Success in Morocco aimed to help participating youth invest in quality education and build their human capital. Indeed, Kluve et al. (2016) found that when confronted with a difficult labour market, men, younger participants and those from more affluent households were more likely to invest in post-secondary education and postpone their entry into the labour market. Male graduates of an apprenticeship programme in Malawi also continued investing in their human capital and spent around 3–6 hours weekly in additional skills development activities (Cho et al., 2015). Graduates of a small-scale intervention in Argentina reported that they would like to re-enrol and complete their secondary education (Hamilton et al., 2013). Participants in the Colombian Jóvenes en Acción were encouraged to complete their schooling and continue to tertiary education, up to four years after training (Kugler et al., 2015 cited in Honorati, 2015). On the other hand, KYEP graduates in Kenya also increased their participation in certified skills training or internship programmes compared to the control group, but not the probability of holding a tertiary education or college degree (Honorati, 2015). Girls-only programmes also led to similar improvements. AGI in Haiti raised out-of-school girls’ aspirations for further education and subsequently afforded them better opportunities in the labour market. Three months after programme completion, graduates were 14 percentage points more likely to expect to enrol in school than the control group (Rodella et al., 2015). Girls’ participation in ELA in Uganda did not affect their schooling, with no differences measured between girls in rural and urban areas, those from affluent and poor households, and younger and older girls. In addition, girls who had dropped out of school were more likely to want to re-enrol (Bandiera et al., 2015). Some evidence from ELA in Bangladesh indicated that participants aimed to use their earnings to finance their education, while those who only received microfinance support without training did not express any such motivation. As we have already noted, girls who were afraid that getting involved in economic activities would negatively affect their schooling, did not use available loans to fund their education. Although the evaluation did not assess the effects of long-term economic engagement, it did not find any immediate change in girls’ enrolment rates (Shahnaz and Karim, 2008). Bodily integrity and freedom from violence The evaluations of six interventions that explicitly targeted adolescent girls also examined whether participation improved their attitudes and exposure to gender-based violence within their household and the community. While acceptance of violent partner behaviour declined and spousal relationships improved, programme participation may have also unintentionally increased girls’ exposure to sexual harassment and abuse. Evidence from AGI in Haiti showed that participating girls increased their self-control, capacity for anger management in intimate relationships, and nonacceptance of intimate partner violence under some circumstances. They were less likely to accept that a husband/partner can hit his wife when she questions him about cheating, cheats on him with another man or refuses to have sex with him. However, attitudes to violence under other conditions (such as arguing with or insulting her partner) remained unchanged among both participating girls and those in the control group. Rodella et al. (2015) estimated that the programme led to an overall decline in the violence score, signalling that acceptance of violent behaviour fell. The AGI intervention in South Sudan also aimed to make participants less vulnerable to violence by encouraging girls to leave abusive relationships and report acts of violence, and teaching them ways to avoid dangerous situations such as travelling alone or taking rides from people unknown to them (Jurgens et al., 2013). On the other hand, evidence from the TESFA programme in Ethiopia indicated improved communication between married girls and their husbands and a decrease in genderbased violence (Edmeades et al., 2014). Some qualitative evidence from ELA in Bangladesh noted that many girls reported turning down proposals to marry early – yet quantitative data does not confirm this finding (Shahnaz and Karim, 2008). 29 Interventions promoting adolescent girls’ economic capabilities: what works? Programme participation may have also unintentionally exposed girls to violence. Access to loans made some participants in AGI in South Sudan vulnerable to domestic violence as their partners tried to get hold of the money and threatened girls or even used violence against them (Jurgens et al., 2013). Graduates of AGI in Haiti repeatedly faced sexual advances in their job searches and were asked for sexual favours in exchange for having their job application considered (Rodella et al., 2015). Similarly, reports of gender-based violence, and especially sexual harassment, increased at the end line of AGI in Rwanda. Although Botea et al. (2015) clarified that available data did not allow them to determine when and whether this harassment was associated with programme participation, they also noted the possibility that respondents became more comfortable or better able to recognise and report harassment at end line as a result of the training they received. Some of the few girls that started working after attending the livelihoods programme in Allahabad, India, also mentioned being sexually harassed by men and boys on their way to and from work (Sebastian et al., 2005). Physical health and nutrition Several integrated livelihoods programmes targeting adolescent girls also included a reproductive health component to improve girls’ knowledge and behaviour. The evaluations of five such programmes, along with the evaluations of four technical and vocational training programmes, investigated their health and nutrition effects. In most cases, adolescent girls demonstrated increased knowledge, improved practices, lower fertility rates, and improved nutrition. ELA in Uganda, which explicitly aimed to empower girls to overcome the economic and health challenges they faced, improved their knowledge of HIV, pregnancy and risky sexual behaviour: the fertility rate among participating adolescents fell by 26% over a two-year period and early entry into marriage/cohabitation fell by 58% compared to baseline, while marriage rates for girls in the control group rose. Girls who reported being sexually active also increased self-reported condom use compared to baseline, while the share of girls reporting having had sex against their will declined and was lower among participants than in the control group (Bandiera et al., 2015). In rural Ethiopia, all participants in TESFA improved their SRH knowledge, yet girls who received the SRH training showed the highest gains, including the largest increase in the use of modern family planning methods and taking up HIV testing. Girls receiving the combined health and livelihoods intervention showed the second largest effects (Edmeades et al., 2014). The life skills training provided by AGI in South Sudan also focused on health issues, giving girls information on family planning and empowering them to discuss issues with their partners; as a result, the programme contributed to a reduction in early pregnancy rates (Jurgens et al., 2013). Female participants in the Juventud y Empleo programme in the Dominican Republic were less likely to get pregnant compared to women in the control group. This effect was mostly driven by adolescent girls aged 16–19, for whom the pregnancy rate fell by 45% between baseline and follow-up survey compared to girls in the control group (Ibarrarán et al., 2014). Female trainees in an apprenticeship programme in Malawi were also significantly less likely to have given birth in the past year compared to women in the control group (Cho et al., 2015). While the livelihoods and reproductive health programme in the urban slums of Allahabad, India, provided reproductive health classes for targeted participants and a control group, such knowledge increased more among the former. Sebastian et al. (2005) attributed this difference to higher class attendance in intervention areas, as girls’ participation in vocational training was conditional on good attendance at those classes. They also attributed it to the time girls spent in peer groups where they could continue discussing and improving their knowledge. However, whether this knowledge was translated into practice remained unexplored (ibid.). On the other hand, ELA participants in Bangladesh did not improve their HIV and STI knowledge compared to the control group, with qualitative data showing that girls with higher education and socioeconomic status had better HIV awareness and hygiene practices (Shahnaz and Karim, 2008). Neither EPAG in Liberia nor the Employment Fund in Nepal had any impact on fertility, sexual behaviour or HIV knowledge (Adoho et al., 2014; Chakravarty et al., 2016a). Two impact evaluations also investigated programme effects on nutrition. While the Employment Fund in Nepal did not have any considerable impact on household food security or protein consumption (Chakravarty et al., 2016a), the evaluation of EPAG in Liberia found evidence of improved household food security as girls’ participation increased weekly consumption of meat and fish in their households (Adoho et al., 2014). 30 Psychosocial wellbeing Evidence from 15 programmes suggests that participation in vocational or business skills training or in livelihoods programmes can also lead to improved psychosocial wellbeing. Participants reported lower stress and anxiety levels, greater sense of control over their life, greater optimism about the future, enhanced self-esteem and improved social relationships. Only the evaluations of two other programmes did not find any such change: the evaluation of ProJoven in Peru emphasised that the programme had no significant impact on participants’ self-esteem, perhaps because such modules were not included in the training provided (Diaz and Rosas, 2016). The evaluation of the SATYA–Pratham programme in India also found no effect on happiness at home (Maitra and Mani, 2014). Participants in AGI in Haiti spoke about reduced mental stress as they felt less overwhelmed and more in control of their lives (Rodella et al., 2015). Likewise, EPAG participants in Liberia reported an increased sense of control, worrying less than non-participants about their job and their personal relationships, and being more satisfied with their job or business (Adoho et al., 2014). Participation in the Juventud y Empleo programme in the Dominican Republic improved young people’s perceptions about their current situation and expectations about the future. Graduates were more likely to perceive being in very good health (Ibarrarán et al., 2014). In Rwanda, AGI participants reported higher life satisfaction, especially in terms of employment and income, and greater optimism about the future. Anxieties about obtaining a job in the future fell, along with worries of not having enough money to meet basic family needs (Botea et al., 2015). In Kenya, girls in Ninaweza, especially those who attended life skills training, were more optimistic than those in the control group about finding quality employment (de Azevedo et al., 2013). In Uganda, ELA graduates reported lower anxiety about finding a good job in adulthood compared to baseline (Bandiera et al., 2015). In Malawi, trainees in an apprenticeship project reported increased happiness and life satisfaction and agreed that their life improved during the past year. Male trainees were twice as likely as female trainees to report improvements in wellbeing, yet gender differences were not statistically significant (Cho et al., 2015). In Uganda, the Youth Opportunities Programme increased subjective wellbeing among participants compared to the control group (Blattman et al., 2014). In Egypt, Neqdar Nesharek participants reported having experienced a transformative process and emphasised that the life skills component helped them understand themselves better. Overall, the programme increased their self-confidence, made them develop a sense of self-worth and the notion that they were able to find a purpose in life, and to set and reach goals (Ramadan et al., 2014). Although the quantitative survey of AGI in Rwanda showed that self-esteem scores for participants increased modestly, qualitative research suggested that girls became much more confident and secure as the training helped them improve their self-esteem by offering support and teaching them self-acceptance, restoring a sense of dignity and making ‘all feel valuable’ (Botea et al., 2015). In Rwanda, Akazi Kanoze participants reported greater confidence and optimism about their skills and employment opportunities (Alcid, 2014). In Nepal, female participants in the Employment Fund reported increased self-confidence in life and business activities (Chakravarty et al., 2016a). Adolescent girls in three programmes (AGI in Haiti, ELA in Bangladesh, and the SEWA-led livelihoods programme in Gujarat, India) also reported increased levels of self-esteem or self-confidence as a result of programme participation (Rodella et al., 2015; Shahnaz and Karim, 2008; SEWA et al., 2006). Participation in these programmes also provided adolescent girls and young women with the opportunity to break their isolation and expand their social networks and potential resources to draw upon, while also improving their personal relationships. AGI in Haiti and Rwanda enabled girls to expand their support network (Rodella et al., 2015; Botea et al., 2015). For instance, by the end of the programme in Rwanda, trainees reported gaining one new friend of the same age (Botea et al., 2015). In the livelihoods programme in Allahabad, India, six times as many participants belonged to a group at end line compared to the control group (Sebastian et al., 2005). Apart from making new friends, participants in ELA in Bangladesh felt that attending groups and interacting with peers released the stress of daily life and made them feel supported (Shahnaz and Karim, 2008). Girls in TESFA in Ethiopia reported increased social capital and support (Edmeades et al., 2014). Two evaluations paid particular attention to the importance of having access to mentors. In Rwanda, participants in Akazi Kanoze reported significant increases in having mentors who believed in and encouraged them (Alcid, 2014). Similarly, in Nepal, the Employment Fund 31 Interventions promoting adolescent girls’ economic capabilities: what works? provided girls with mentors who could advise and support them on work-related issues; having a mentor had a significantly larger impact on empowerment outcomes for female than male participants (Chakravarty et al., 2016a). Apart from having supportive mentors, participants also reported improved household relationships. Participation in AGI in Haiti was associated with increased parental support of wage employment and relevant opportunities for their daughters, despite initial doubts about girls working outside of the home (Rodella et al., 2015). Similarly, most participants in AGI in Rwanda felt that their parents were supportive of their new activities – yet rates were already high at baseline (Botea et al., 2015). Voice and agency Apart from providing skills and improving economic outcomes for adolescent girls, some interventions also improved gender attitudes towards female employment, increased girls’ mobility, made them develop higher aspirations, enhanced their autonomy and self-efficacy, and increased their involvement in decision-making. Overall, 13 programmes led to positive changes, while the evaluations of two others reported no relevant change. In Uganda, ELA participants were more likely to feel that both women and men should equally participate in work, education and household tasks, compared to the control group. They also reported significantly higher ideal age at marriage for women, and lower preferred number of children, while they expressed an aspiration for their daughters (and not their sons) to marry five years later, close to the average age at first marriage in developed countries (Bandiera et al., 2015). In Liberia, EPAG participation also led to a small but significant shift in gender attitudes on the division of household labour such as cooking, washing, cleaning and caring for children (Adoho et al., 2014). The SEWA-led livelihoods programme in Gujarat, India, led to participants questioning traditional gender roles and adopting more gender-egalitarian attitudes, including being significantly more likely to disapprove of wife-beating at end line compared to the control group (SEWA et al., 2006). In Haiti, AGI participants increased the belief that women and men should be equally responsible for income generation; yet they continued to believe that women were still responsible for childcare, cooking and housework (Rodella et al., 2015). The livelihoods intervention in the urban slums of Allahabad, India, did not improve attitudes around the acceptability of female wage labour, shared domestic responsibilities and recognition of women’s value. No significant differences were measured between baseline and end line or between the intervention and the control group – a finding that the authors attributed to the short duration of project activities and lack of an explicit module on gender roles (Sebastian et al., 2005). The evaluation of AGI in Rwanda noted that while graduates increased their employment, they did not reduce the hours spent on household chores, with particularly young mothers continuing to spend more time than any other group on such chores (Botea et al., 2015). On the other hand, participation in such programmes also enabled girls to increase their mobility. In Haiti, AGI participants were more likely to go to the training centres, NGOs and health centres than the control group (Rodella et al., 2015). In South Sudan, AGI participants also increased their mobility, visiting girls’ clubs and engaging in community events (Jurgens et al., 2013). Girls in the ELA programme in Bangladesh also reported increased mobility and participation in recreational activities compared to non-participants who spent more time at home (Shahnaz and Karim, 2008). Girls in the SEWA- led livelihoods programme in Gujarat, India, increased their mobility, but they did not significantly increase their freedom of movement, which remained supervised (SEWA et al., 2006). The livelihoods intervention in Allahabad, India, increased participants’ mobility as girls became members of peer groups and travelled to attend vocational courses. The vast majority of participating girls also significantly increased their knowledge of a safe space where they could meet, between baseline and end line and compared to the control group. Qualitative research indicated that parents accepted and allowed girls to make visits unaccompanied (Sebastian et al., 2005). Likewise, Neqdar Nesharek graduates in Egypt increased their mobility as those who started their business often had to travel to purchase raw materials and interact with providers and customers, while a few also found employment outside their villages (Ramadan et al., 2014). While the evaluation of the Employment Fund in Nepal did not find any increase in female out-migration (Chakravarty et al., 2016a), a follow-up of ELA in Uganda noted that four years post-intervention, graduates with the highest scores in the ‘economic empowerment’ and ‘control over the body’ indices were most likely to migrate away from their villages (Bandiera et al., 2015). 32 Programme participation also enabled participants to develop higher life aspirations. Girls in AGI in Haiti started to believe that their situation was not the result of luck or fate but of their actions, and thus they could try to change and manage their lives better. In particular, participants enhanced their education and employment aspirations compared to the control group (Rodella et al., 2015). Similarly, Adoho et al. (2014) pointed out that the ‘most striking’ empowerment finding was the high aspirations of EPAG participants who expressed their goal to use their training to acquire more skills and move into medicine, business and politics. Graduates from the Juventud y Empleo programme in the Dominican Republic were more likely to have higher expectations about the future in terms of education, business, professional aspirations, a better life and higher social status. The evaluation found that those effects were actually stronger for women and adolescent participants (Ibarrarán et al., 2014). In Bangladesh, ELA participants questioned traditional women’s roles and reported higher aspirations about the future such as improving their skills, getting a good job and marrying later (Shahnaz and Karim, 2008). Several programmes also aimed to increase girls’ agency and involvement in decision-making so that they could start transforming their lives. Evaluations used self-assessed measures to assess relevant change. The Employment Fund in Nepal significantly increased girls’ self-confidence in life and entrepreneurial activities, with positive impacts measured in six out of ten empowerment indicators for self-reported individual self-regulation ability (Chakravarty et al., 2016a). Similarly, EPAG graduates in Liberia reported increased confidence in terms of starting and managing a business compared to the control group (Adoho et al., 2014). In Rwanda, AGI improved entrepreneurial self-confidence, with girls reporting being more confident of their ability to identify business opportunities, run their business or interview for a job (Botea et al., 2015). Both female and male trainees of the apprenticeship intervention in Malawi reported increased self-confidence in their ability to change their economic activities, and move from agriculture into incomegenerating activities in other sectors (Cho et al., 2015). The livelihoods programme in the slums of Allahabad, India, also improved girls’ social skills – those related to communicating ideas and working with others in a group. In particular, girls reported improved skills in negotiating and getting better prices for their products (Sebastian et al., 2005). In Bangladesh, girls who attended ELA livelihoods training and accessed a loan were more valued by their families and communities and were able to express themselves and voice their opinions on family issues. As already mentioned, most of those who received training were able to have significant control over the loan they had taken (Shahnaz and Karim, 2008). In Liberia, EPAG participation was linked with a small but statistically significant impact on control over girls’ own resources and earnings (Adoho et al., 2014). In South Sudan, AGI participants reported increased levels of confidence to talk about family planning and sensitive issues with their partners and a greater role in decision-making on SRH issues. Girls who received a loan also reported being able to make their own decisions about the business and how to invest their money (Jurgens et al., 2013). In Rwanda, increased engagement of AGI trainees in income-generating activities was associated with increased ability to make decisions about their work, training and household tasks. At the same time, individual decision-making about childbearing fell, reflecting more joint decision-making between girls and their partners (Botea et al., 2015). Married participants in the TESFA programme in Ethiopia who generated their own income increased their participation in household economic decision-making, with the exception of loans, which most girls shared with their husbands. Qualitative research clarified that this was linked to girls feeling closer to their partners and thus should not be interpreted as a loss of autonomy. In terms of control over assets, girls reported greater control over less valuable property such as poultry, yet no change in control over larger (more valuable) livestock, where girls again collaborated with their husbands. Participants also improved their communication skills and increased their involvement in decision-making about family planning. While this improvement was significant in all three intervention groups, husbands and families continued to have the final say on family planning (Edmeades et al., 2014). AGI in Haiti improved participants’ agency, defined as the capacity to make choices and pursue desired actions. Compared to the control group, girls demonstrated increased autonomy in decision-making and enhanced social status (Rodella et al., 2015). Young participants in the Neqdar Nesharek programme in Egypt improved their communication skills and thus were able to negotiate and persuade others to support them. Being able to use their new skills, earn an income and provide for their families made them feel empowered (Ramadan et al., 2014). 33 Interventions promoting adolescent girls’ economic capabilities: what works? The Juventud y Empleo programme in the Dominican Republic and its life skills training also had positive effects on participants’ non-cognitive and emotional skills, especially for young female participants. Compared to the control group, adolescents, those with higher education and those in the capital city achieved the highest scores in leadership skills, conflict resolution, self-organisation and persistency of effort (Ibarrarán et al., 2014). Girls in the SEWA-led livelihoods programme in Gujarat, India, improved their communication skills and started voicing their wishes and participating in decision-making. Despite the measured positive change, overall, girls’ agency continued to be limited (SEWA et al., 2006). The evaluation of the 100 Hours to Success in Morocco did not find any significant changes in self-efficacy or participants’ engagement in the community (Kluve et al., 2016). Similarly, the Youth Opportunities Programme in Uganda did not have any effect on community involvement, social cohesion or protest (Blattman et al., 2014). 3.4 Discussion of findings and evidence gaps By bringing together different types of youth programming, this section has provided valuable insights into the effectiveness of vocational and/or business skills training interventions targeting adolescent girls. The reviewed evidence has clearly showed that both youth training interventions focusing on the provision of technical and/ or business skills, and integrated programmes for the economic empowerment of adolescent girls, can improve their technical skills, facilitate entry into the labour market, and help them generate their own income and acquire assets. In addition, these interventions also led to positive changes in key non-economic domains such as further skills acquisition, increased SRH knowledge, lower fertility, improved nutrition, non-acceptance of intimate partner violence and improved partner communication, a greater sense of control in life and optimism for the future, increased social capital, progressive gender attitudes towards female employment, increased autonomy and mobility, and enhanced self-esteem and decision-making power. While valuable in their own right, each one of these changes is also instrumental to further enhancing girls’ labour market access and performance, and their overall economic capabilities. The high-quality evaluations of the Latin American youth training programmes emphasised that female and in some cases adolescent participants often benefited more as the intervention enabled them to move either from unpaid family work or low labour force participation into formal employment, and increase their earnings (Diaz and Rosas, 2016; Ibarrarán et al., 2015; Attanasio et al., 2011; Nopo et al., 2007; Diaz and Jaramillo, 2006; Aedo and Nuñez, 2004). Similarly, youth training programmes elsewhere also enabled disadvantaged low-skilled youth to access technical and soft skills training, leave casual labour, find better employment and increase their income. Once again, female participants showed larger gains than males, as due to gender discrimination they had typically started from a more disadvantaged position than male participants (Chakravarty et al., 2016a; Honorati, 2015; Alcid, 2014; Blattman et al., 2014). Integrated interventions targeting adolescent girls and young women also enabled them to increase their off-farm self-employment, earnings and assets (Bandiera et al., 2015; Botea et al., 2015; Adoho et al., 2014; Edmeades et al., 2014; Ramadan et al., 2014). Although rarely acknowledged, let alone analysed, the economic context also shaped girls’ employment opportunities post-intervention. While programme participation helped girls and young women access formal employment in Latin America, participants in programmes in other regions and countries were not able to significantly increase their wage employment, increasing their entrepreneurial self-employment in the informal economy instead (Bandiera et al., 2015; Botea et al., 2015; Adoho et al., 2014). Evaluations of programmes with weak effects – where participants were more likely to reduce their labour market participation or faced great difficulty in finding employment – attributed this finding to the difficult school-to-work transition and youth integration in the local labour market, prompting participants to delay entry into the labour market and invest more in training (Kluve et al., 2016; Rodella et al., 2015). The evidence presented has also indicated that effective programmes were those that conducted labour market assessments and provided skills training rooted in local realities. In some cases, trainees were able to find employment in the trade they had studied, yet in others, only a small proportion were able to do so (Blattman et al., 2014; Botea et al., 2015). A few programmes also encouraged female participants to enrol in non-traditional trades for women, with varying levels of success. In particular, one vocational programme in Kenya provided female participants with information about the higher earnings of traditionally male-dominated trades; indeed, 40 community sensitisation were more likely than those in the control areas to have assets in their name created by their parents. Parents of participating girls reported that their daughters were more likely to inherit land compared to those in the control area. However, the study clarifies that this was reported, and whether it was realised or not remained to be seen. Parents of participants were also expected to make a bequest directly to the girl at the time of marriage (Landesa RDI, 2013). A few programmes explicitly targeted more vulnerable groups of girls such as domestic workers in urban areas, or orphaned girls, and provided them with a package of services, including financial education and participation in savings groups, with positive effects. The Filles Éveillées (meaning ‘Girls awakened’) programme in urban Burkina Faso provided adolescent girls aged 11–19 years who had migrated to cities and worked as domestic workers with safe spaces, mentors, financial literacy, health and life skills training. Participants in the first cohort (who were aged 11– 16 years) showed improved savings behaviour at end line, with more girls reporting saving in a bank or an informal savings group, and having savings goals (Engebretsen, 2012). Participants in the second cohort (who were aged 15–18 years) also reported that they improved their savings significantly between baseline and end line. Most girls continued saving in informal institutions for both short- and longer-term reasons (Engebretsen, 2013a). A follow-up was conducted one year after programme completion to assess the sustainability of girls’ knowledge, attitudes and behaviours. The proportion of girls saving between end line and follow-up increased slightly. Four out of five girls reported saving money – more than half doing so regularly – for short- and longer-term reasons such as to prepare for their marriage, pay for training or buy supplies for incomegenerating activities. While knowledge of formal savings mechanisms improved, it was still low (Engebretsen, 2013b). The Support to Orphans and Vulnerable Children Affected by HIV/AIDS programme in Rwanda included among its services microfinance groups known as Savings and Internal Lending Communities (SILCs). Members had to contribute on a weekly basis small sums that were pooled and provided as a loan, thus enabling participating adolescents to save safely and access financing when needed. Initially, girls were included in adult SILC groups where they often received support and mentorship, but gradually youth-only groups emerged. Adolescents found it easier to access and repay a loan through youth-only groups as no adults could challenge their participation; yet those groups offered smaller loans, as youth members had more limited earning capacity, and thus inhibited girls’ ability to grow their business. Overall, participants were able to improve their financial capabilities and girls used loans to invest in their small businesses and meet household needs. Despite its positive effects, participation in those groups was inadequate for sustainable livelihoods (Rowe and Miller, 2011). The Out-of-School Adolescents Support project in Zimbabwe targeted adolescents (mostly girls) and also provided them with SILC groups for adolescents where they were offered training for money management but also life skills education and psychosocial support. A primary objective of those groups was to support participants to return or remain in school by teaching them to save, access loans, and invest in small businesses to earn enough money for school fees and basic daily needs, while also enabling them to build their social skills and become selfreliant young adults. Indeed, participation in those groups enabled many girls to develop their business skills, start a small business, earn an income and pay their school fees (Miller et al., 2011). 4.3 Non-economic outcomes Apart from directly improving girls’ financial and economic capabilities and in line with their particular objectives, these interventions also led to positive changes in the five remaining GAGE domains: education, bodily integrity, physical health, psychosocial wellbeing, and voice and agency. Education and learning A common objective of programmes targeting adolescent girls is to improve their school enrolment and attendance rates. Accordingly, ten interventions led to reported positive relevant changes; a few also enabled girls to realise the importance of education and use part of their savings or earnings to cover school costs. The Kishoree Kontha (meaning ‘Adolescent girls’ voices’) programme targeted girls aged 10–19 in rural Bangladesh with four distinct interventions: a basic empowerment package of safe spaces and life skills training; this basic package augmented with a financial component; cooking oil as an incentive for older girls aged 15–17 to remain unmarried; and a combination of all three main components. The oil incentive had a positive impact on girls’ education with girls in the oil incentiveonly group being 18% more likely to be in school at end 41 Interventions promoting adolescent girls’ economic capabilities: what works? line (19% more likely for those who received both the oil and the empowerment package) (Field et al., 2016). Also in rural Bangladesh, BALIKA increased the proportion of girls in school. Girls who received the livelihoods training were 22% more likely to be in school, while those who received educational support in safe spaces particularly improved their maths skills. The programme also significantly increased girls’ exposure to information, with participants (especially those in the livelihoods training) more likely to listen to the radio, watch television and read a newspaper than the control group (Amin et al., 2016). Another programme in rural Bangladesh, Kishori Abhijan, contributed to improved enrolment rates only for younger girls aged 12–14 living in the poorest districts. Participants also increased their exposure to information. Most importantly, older girls’ involvement in paid work did not lead to them dropping out of school, but working girls reported spending fewer hours studying outside the classroom (Amin and Suran, 2005; Amin, 2011). As was the case with Kishori Abhijan, younger girls aged 10–14 in the Berhane Hewan (meaning ‘Light for Eve’) programme in rural Ethiopia were more likely to be in school and to delay their marriage compared to the control group. The programme targeted girls aged 10–19 years, both unmarried and those already married, with safe spaces and mentors, educational support, livelihoods training, an economic incentive, and community mobilisation and sensitisation in order to delay early marriage and empower girls. Participants also improved their literacy levels between baseline and end line (Erulkar and Muthengi, 2009). When asked to identify the programme components that were more important to keep girls in school, girls’ fathers, mothers and husbands emphasised the role of educational support in the form of school supplies, followed by the provision of an economic incentive and girls’ groups (Mekbib and Molla, 2010). Another programme in rural Ethiopia, Empowering Adolescent Girls, also increased their school enrolment and reduced dropout rates. Girls used the stipends provided to purchase books, rent a room to stay close to their secondary school or meet their basic needs. In addition, fuel-saving stoves greatly reduced the time girls spent on firewood collection and cooking time, and thus enabled them to spend more time studying. The project also constructed water points and latrines, enabling girls to be on time in class and to increase their enrolment and performance during menstruation (CRS, 2009). Better Life Options graduates in India were significantly more likely than the control group to be literate and to have completed secondary education (CEDPA, 2001). Participants in the Ishraq programme in Egypt also increased their literacy and numeracy compared to nonparticipants, and more than two-thirds of participants were able to enter or re-enter formal education (Brady et al., 2007; Sieverding and Elbadawy, 2016). However, continued academic and social support to those girls, especially during their first year back to school, was considered to be critical as their transition to a formal learning environment was not easy and could lead them to drop out (Selim et al., 2013). Two programmes targeting vulnerable girls in marginalised urban spaces also showed positive effects. The Powering Up Biruh Tesfa (meaning ‘Bright future’) project in the capital of Ethiopia provided out-of-school girls aged 7–18 (many of whom were migrants and domestic workers) with safe spaces and female mentors, life skills and HIV training, financial literacy and a voucher for health services. Participants significantly increased their participation in non-formal education compared to the control group, while those girls who had never been in school improved their literacy and numeracy scores (Erulkar and Medhin, 2014). Filles Éveillées in urban Burkina Faso, which also targeted domestic workers, contributed to a decrease in the proportion of girls who had never been to school by supporting those interested to enrol in evening classes. It also increased the proportion of girls who could read (Engebretsen, 2012). Girls in the second cohort also reported investing their earnings in education and training (Engebretsen, 2013a). Likewise, the Out-of-School Adolescents Support project in Zimbabwe and particularly adolescents’ involvement in savings and lending groups enabled them to access loans and use earnings from their small business to pay school fees and stay or return to school. Teachers supported the programme and ensured that participation in microfinance groups did not interfere with learning (Miller et al., 2011). Bodily integrity and freedom from violence Many interventions implemented in countries where early marriage continues to be a significant problem for adolescent girls have an explicit objective to empower girls and delay their marriage by enabling them to remain in school or engage in economic activities and build their agency to act. The evaluations of eight programmes 42 focused on bodily integrity outcomes linked to early marriage and provided mixed results. In addition, the evaluations of three programmes investigated impact on girls’ exposure to physical and sexual violence; two of them reported positive change. In rural Bangladesh, BALIKA was successful in delaying early marriage: Amin et al. (2016) found that in all three intervention groups, adolescent girls were less likely to be married early compared to those in control communities. In particular, girls who received the educational support and the gender rights training were more likely not to marry early than girls who received the livelihoods training. Age-differentiated data also shows that BALIKA was particularly successful in delaying marriage among those aged 16 and 17, and in the case of the educational intervention among those under age 16. The evaluation of Kishoree Kontha indicated that older girls aged 15–17 years who received the cooking oil incentive were less likely to be married than the control group. In particular, girls who only received the oil incentive were 21% less likely to marry early, while those who also received the empowerment training were 14% less likely (Field et al., 2016). On the other hand, Kishori Abhijan enabled only younger girls residing in the poorest districts and attending school to delay their marriage. One particular challenge that the programme encountered was that it did not consider the economic importance of dowry and thus its early marriage impact was compromised as girls who delayed their marriage had to provide higher dowry payments (Amin, 2011). In India, graduates of the Better Life Options programme were less likely than the control group to get married before age 18 (CEDPA, 2001). Its later version implemented in rural Uttar Pradesh also sought to delay girls’ marriage by making them aware of their rights and the minimum legal age at marriage, enhancing their ability to decide and negotiate when to get married, and providing them with an income-generating alternative to early marriage. Acharya et al. (2009) found that the proportion of participants who were aware of the legal minimum age at marriage increased as well as their preferred age at marriage. However, the overall percentage of those who married did not decline significantly, although age at marriage increased modestly among participants compared to the control group. The DISHA programme 7 The Meseret Hiwott programme in rural Ethiopia, which targeted married adolescent girls aged 10–24, also introduced in response to their husbands’ requests a parallel programme, Addis Birhan (meaning ‘New light’), for husbands of all ages to improve their communication and support skills, and enable them to contribute to the wellbeing of their wives and families. also increased participants’ awareness of the legal age at marriage for girls and their belief that girls should not be married until they are 18 or older. Most importantly, participating girls were more likely to marry nearly two years later between baseline and end line, with age at marriage increasing to nearly 18 years (Kanesathasan et al., 2008). Participants in the Security for Girls through Land project in rural West Bengal were also expected to marry nearly one and a half year later, and their parents to pay less cash dowry (Landesa RDI, 2013). In Ethiopia, younger participants in Berhane Hewan were more likely to delay their marriage: the proportion of married girls aged 10–14 fell in the intervention area, while it increased in the control area; on the other hand, older girls aged 15–19 in the intervention area were more likely to be married at end line and that percentage remained the same over the study period, while it declined among the control group (Erulkar and Muthengi, 2009). The fathers, mothers and husbands of participants reported that community awareness activities, followed by girls’ groups and mentors’ house-to-house visits, were the three most effective strategies to tackle early marriage (Mekbib and Molla, 2010). In Egypt, the proportion of Ishraq participants preferring to marry before age 18 dropped substantially from baseline to end line, especially among girls with longer exposure to the programme. In addition, although experience of violence did not change, participants showed a statistically significant decline in attitudes accepting violence and controlling behaviours; greater programme exposure was associated with a stronger decline (Brady et al., 2007). Two programmes provided promising evidence about reduced levels of violence. Participants in the SHAZ! project in Zimbabwe showed a greater reduction in the experience of violence compared to the control group (Dunbar et al., 2014). Married girls who participated in the Meseret Hiwott programme in Ethiopia, while their husbands joined the parallel programme Addis Birhan7 for men, reported lower levels of domestic violence compared to non-participants and girls whose partners did not participate in the men’s programme. However, girls in the wives-only group reported higher levels of sexual and physical violence than the other two groups (Erulkar and Tamrat, 2014). 43 Interventions promoting adolescent girls’ economic capabilities: what works? Physical health and nutrition The evaluations of 12 programmes presented evidence on programme effects on the SRH knowledge, attitudes and behaviours of participating girls. However, evidence about whether this improved knowledge was sustained and translated into lasting behavioural change remains inconclusive. On the other hand, evidence on improvements in other health areas or nutrition was rather limited. In rural Bangladesh, Kishoree Kontha’s oil incentive reduced the likelihood of early motherhood. Girls who only received the cooking oil were 6% less likely to have given birth before age 18, while those receiving the combined oil and empowerment intervention were 19% less likely to do so compared to the control group (Field et al., 2016). Girls in all three intervention groups of BALIKA significantly improved their reproductive health knowledge and behaviour compared to the control group. They were thus more likely to know about family planning methods, HIV transmission and menstruation; they were also significantly more likely to receive treatment for reproductive health problems (Amin et al., 2016). Similarly, participants in Kishori Abhijan improved their SRH knowledge (Amin, 2011). In India, girls of the Better Life Options programme had higher HIV awareness. Effects were stronger among married participants who were more likely than the control group to have discussed family planning with their husband, to use contraceptives, to deliver in a health institution and receive antenatal and postnatal care. Those girls also reported higher child survival practices (CEDPA, 2001). Participants in the linked intervention in Uttar Pradesh also reported increased awareness of SRH issues, including sex and pregnancy, contraception and HIV (Acharya et al., 2009). DISHA participants – girls more so than boys – significantly increased their knowledge of modern contraception and were more likely to know where to access it compared to non-participants. Married participants were also significantly more likely to report modern contraceptive use compared to non-participants (Kanesathasan et al., 2008). In rural Ethiopia, girls in Berhane Hewan were more likely than the control group to know about condoms and oral contraception. They also improved their knowledge of STIs and discussed reproductive health issues more with their friends at end line. Almost three-quarters of sexually experienced participants had used a contraception method at end line compared to less than half of girls in the control group (Erulkar and Muthengi, 2009). When asked to identify the programme components that contributed to the increased use of family planning methods, girls’ fathers, mothers and husbands reported that community awareness activities positively affected family planning behaviour, followed by group meetings and house-to-house visits by mentors (Mekbib and Molla, 2010). Participants in Meseret Hiwott were more likely to report increased use of family planning and voluntary counselling and testing. Interestingly, reported increases were higher among girls whose husbands also participated in the parallel men’s intervention, and lower among nonparticipating girls (Erulkar and Tamrat, 2014). Interventions that targeted girls with particular vulnerabilities, such as orphaned girls, out-of-school girls, those in domestic work and urban migrants, also found positive changes. Participants in SHAZ! in Zimbabwe were more likely to use a condom and less likely to engage in transactional sex compared to the baseline; they also had significantly reduced unintended pregnancy rates compared to the control group. They were also more likely to improve their food security compared to the control group (Dunbar et al., 2014). Girls in the Ishraq programme in Egypt improved their knowledge about nutrition, anaemia, smoking and SRH issues. However, they were not much more likely than non-participants to seek professional health care when sick (Brady et al., 2007; Sieverding and Elbadawy, 2016). Participants in Biruh Tesfa in urban Ethiopia were more likely than the control group to have improved their HIV knowledge, awareness of relevant services and intention to use them. Biruh Tesfa targeted marginalised out-of-school girls aged 10–19 living in slum areas – most of whom were migrants and domestic workers, but also included girls with disabilities. Girls were provided with safe spaces, mentors, financial literacy, HIV education and life skills training (Erulkar et al., 2013). Its expansion in the capital of Ethiopia, the Powering Up Biruh Tesfa, which also included a voucher subsidising the cost of basic health services, enabled participating girls to be significantly more likely to have visited a health facility compared to the control group (Erulkar and Medhin, 2014). Girls in the first cohort of Filles Éveillées in Burkina Faso also improved their overall health and SRH knowledge and attitudes towards use of local health services. Almost all girls reported that programme participation enabled them to manage their health better (Engebretsen, 2012). Likewise, participants in the second cohort also improved their SRH awareness, including knowledge of modern family planning methods, and more girls visited a health 44 centre while fewer girls used self-medication at end line. Over time, girls’ knowledge of positive health behaviours during pregnancy also improved (Engebretsen, 2013a). One year after programme completion, girls continued to improve their overall health awareness. However, changes in pregnancy knowledge were inconsistent, and awareness of when and how a woman can get pregnant decreased over time. In addition, family planning knowledge also declined and knowledge about STIs, including HIV, was also inconsistent (Engebretsen, 2013b). Psychosocial wellbeing Participation in integrated empowerment programmes with safe spaces, female mentors and life skills training enabled girls to expand their social network, acquire new friends, discuss their experiences and access support. This was particularly important for girls with increased vulnerability such as domestic workers in urban areas, out-of-school or orphaned girls. Almost all participants in the first cohort of Filles Éveillées in urban Burkina Faso reported having more friends as a result of the programme and all reported being able to talk to their mentors about the problems they faced (Engebretsen, 2012). Girls in the second cohort (who were older adolescents) were also able to increase their social capital in terms of having friends, a place to meet them and someone to turn to for advice (Engebretsen, 2013a). A follow-up with a subsample one year after programme completion found that those girls continued expanding their social network even more between end line and follow-up (Engebretsen, 2013b). Likewise, girls in Berhane Hewan and Biruh Tesfa in Ethiopia, in Better Life Options and the linked girls’ empowerment programme in India, in Ishraq in Egypt, in Kishori Abhijan in Bangladesh, and in the Out-of-School Adolescents Support project and SHAZ! in Zimbabwe, reported expanding their network of friends, participating more in local groups, and increasing their access to social support (Erulkar and Muthengi, 2009; Erulkar et al., 2013; CEDPA, 2001; Acharya et al., 2009; Brady et al., 2007; Selim et al., 2013; Amin, 2011; Miller et al., 2011; Dunbar et al., 2014). In addition, participants increased their self-confidence. In Bangladesh, Kishori Abhijan girls reported improved sense of self-worth (Amin, 2011), and in Egypt, regular participants in Ishraq reported higher levels of selfconfidence than non-participants (Brady et al., 2007). In Burkina Faso, fewer girls in Filles Éveillées continued feeling weak and vulnerable at end line (Engebretsen, 2012; 2013a), while vulnerable and orphaned girls in the Out-of- School Adolescents Support project in Zimbabwe also reported improved self-image and enhanced self-esteem (Miller et al., 2011). Voice and agency In line with their stated objective to empower girls, nine empowerment programmes increased participants’ knowledge about their rights, enhanced their confidence and communication skills to express their opinion about issues that affected their lives, improved their relationships with their parents and husbands, increased their mobility and decision-making power, and helped them develop more progressive gender attitudes. In rural Bangladesh, Kishori Abhijan enabled participants to improve their mobility, with girls reporting being more likely to visit their friends or go to the movies compared to the control group (Amin, 2011). Similarly, girls in BALIKA increased their mobility and held more genderequitable attitudes than the control group (Amin et al., 2016). Kishoree Kontha participants who received the oil incentive were more likely than any other group to enjoy increased mobility and attend community events outside their village alone. They were also more likely to say that they would negotiate with their parents if they were to receive a marriage proposal (Field et al., 2016). Better Life Options participants in India were more likely than the control group to increase their mobility, use public transport, go alone to the market, travel outside their village and visit a health centre. They were also more likely to be involved in decision-making about going to the market, spending their own earnings, continuing their education and getting married, with girls feeling more confident to speak in a group and in front of elders. Participants also developed more gender-egalitarian attitudes, while almost one in four were community volunteers and involved in organising community events (CEDPA, 2001). Participation in the linked girls’ empowerment project in Uttar Pradesh, India, also enabled girls to develop more gender-egalitarian attitudes, particularly towards work, such as the view that it is important for both men and women to work. The evaluation attributed this finding to the inclusion of the livelihoods training component, which emphasised the importance of girls learning a skill and becoming economically independent. Once again, girls also improved their decision-making ability and mobility: they reported being less afraid to go out alone, increasing 45 Interventions promoting adolescent girls’ economic capabilities: what works? their confidence and becoming more able to speak without fear, and being more likely to speak out if they disagreed with another opinion. They also reported becoming better informed about their rights (Acharya et al., 2009). The Security for Girls through Land project in West Bengal, India, also significantly improved participants’ knowledge about their legal rights and increased their confidence to exercise those rights (Landesa RDI, 2013). The DISHA programme in India enabled unmarried girls to talk to their parents about getting married, and to feel that their marriage wishes were being considered by their parents. Irrespective of their marital status, female participants also increased their mobility at end line. In addition, spousal communication improved, with significantly more married girls and women reporting being able to talk to their husband about contraception at end line (Kanesathasan et al., 2008). Married girls in the Meseret Hiwott in Ethiopia were also more likely to report increased spousal help with domestic work, accompaniment to the clinic, family planning, and voluntary counselling and testing compared to non-participants. Erulkar and Tamrat (2014) noted that the programme enabled girls to build their confidence, communication skills and awareness of gender-equitable relationships. When both spouses participated in the two related interventions, increases were even higher. On the other hand, girls who did not participate in the programme were more likely to need permission to leave the house than girls who did (ibid.). The first cohort of Filles Éveillées participants in Burkina Faso reported feeling more confident to express themselves, make decisions and resolve conflicts with others. Participants also improved their gender-egalitarian attitudes (Engebretsen, 2012). Likewise, girls in the second cohort improved their confidence and ability to decide for themselves. More girls reported deciding alone on how to spend their money – yet the difference was not statistically significant. A statistically significant increase was found in gender-egalitarian attitudes (Engebretsen, 2013a). Twothirds of participants continued to decide alone how to spend their earnings between end line and the followup (Engebretsen, 2013b). Employers and community members also reported that participants were more respectful and stable in their behaviours, learnt new skills and were willing to share their new knowledge with others (Jarvis and Kabore, 2012). Girls in the Ishraq programme in Egypt also adopted more gender-egalitarian attitudes as more agreed that girls should be allowed to play sports like boys, that the appropriate age for marriage was 18 or older, and that they should be involved in the decision over whom to marry (Brady et al., 2007; Selim et al., 2013; Sieverding and Elbadawy, 2016). Girls also expressed the desire to have fewer children and their intention not to perform female genital mutilation/cutting (FGM/C) on their daughters (Brady et al., 2007; Sieverding and Elbadawy, 2016). Overall, the programme did not have an effect on girls’ empowerment as measured by self-esteem, attitudes towards violence against girls, and mobility, and had only a marginally significant effect on their decision-making. Sieverding and Elbadawy (2016) attribute this finding to the fact that behavioural change among young women can only take place when interventions effectively address the underlying gender unequal relations and sufficiently engage with family and community members that restrict girls’ control over their life. For instance, Ishraq did not succeed in changing the attitudes of mothers or brothers. 4.4 Discussion of findings and evidence gaps Despite being a rather diverse set, the interventions in this group offer interesting insights into what works when economic and non-economic components are bundled together to improve adolescent girls’ capabilities – most often by delaying their marriage, improving their SRH knowledge and practices, strengthening their skills and enhancing their economic outcomes. It is thus clear that these programmes have the potential to increase girls’ access to employment and particularly involvement in income-generating activities, their earnings, savings and economic assets. They also appear to be effective at inducing positive change in a number of non-economic domains, such as girls’ schooling and learning outcomes, psychosocial wellbeing, voice, mobility, relationships with significant others, gender attitudes (including towards women’s work) and SRH knowledge and practices, which further strengthen and reinforce girls’ economic capabilities. With regard to programme effectiveness at delaying marriage, the evidence is less clear-cut, with younger girls apparently benefiting more (Amin, 2011; Erulkar and Muthengi, 2009) or economic incentives holding a promising potential (Field et al., 2016). However, the particular effects of livelihoods training and involvement in income-generating activities as an alternative to early marriage are not well analysed. 46 If the development of strong economic capabilities is critical for girls’ empowerment and wellbeing, as the literature claims, then it would be analytically important that all studies of integrated multi-component interventions measure economic outcomes alongside the other outcomes they currently tend to focus on and which are more in line with their broader objectives. Moreover, indicators measuring change should be refined to help capture actual change, as improved knowledge or selfreported intention does not automatically translate into better, let alone sustainable, practice. The evaluation design of these interventions should also become more rigorous in order to identify the relative effectiveness of different components and the key pathways through which change occurs. Despite its critical importance, this is an area where information is still very limited but (encouragingly) growing, as indicated by the inclusion of different intervention groups and the comparison of their outcomes. Apart from not being always measured and analysed, the integration of economic components into a broader empowerment programme targeting disadvantaged girls in settings with scarce resources may experience a number of challenges, largely due to weak programme design and implementation. This is particularly true with the provision of vocational training, despite evidence that it can be used as an entry point motivating girls to join and attend an intervention (Acharya et al., 2009; CEDPA, 2001). Limited implementation experience, shortage of professional providers and trainers, lack of market assessments to align technical skills with market demand, lack of access to credit upon graduation to start a small business, along with limited attention to issues such as distance to the training facility, safety concerns, girls’ household responsibilities or lost earnings, and restrictive gender norms – all of these factors compromised the effectiveness of the vocational training component of an integrated empowerment intervention in several cases. This highlights the need to consider not only which services are provided but also their quality (ASCI, 2013; Rowe and Miller, 2011; Miller et al., 2011; Acharya et al., 2009; Kanesathasan et al., 2008). On the other hand, community engagement and intensive mobilisation activities that include parents, partners, brothers and other stakeholders appear promising in terms of enabling girls to practise their newly acquired knowledge and skills and become agents of change in a supportive environment. Emerging evidence suggests that when significant others are also targeted and participate in programme activities, effects are stronger, as indicated by decreased rates of early marriage, girls’ greater control over their own income and probability to have financial assets created by their parents in their name, or increased use of reproductive health services, spousal assistance with household chores, and lower levels of domestic violence for married girls (Erulkar and Tamrat, 2014; Landesa RDI, 2013; Mekbib and Molla, 2010). However, when programming focused primarily on girls and paid inadequate attention to the close interdependence of the individual with the group, not investing in regular community activities to transform traditional gender norms and attitudes in conservative settings, there was a lack of effects signalling some promising change in the underlying gender unequal relations such as more egalitarian attitudes and non-acceptance of gender-based violence (Sieverding and Elbadawy, 2016). Integrated interventions aiming to empower girls tend to work across a wide adolescent age range and may target girls with different components. A few programmes divided girls in different groups depending on their age or marital status and adjusted the services provided accordingly. Thus only older girls received livelihoods training or married girls met less often in safe spaces in recognition of their greater time constraints (Amin, 2011; Erulkar and Muthengi, 2009). However, most evaluation studies did not specify which components or activities targeted which age groups or with which age groups outcomes were measured. Nonetheless, the few that did so offer some useful insights. Young adolescent girls had difficulty understanding sexuality and reproduction issues, and thus needed age-appropriate SRH curricula (Erulkar and Medhin, 2014; Jarvis and Kabore, 2012). A number of evaluations also found that younger girls benefited more from early marriage and education interventions (Amin 2011; Erulkar and Muthengi, 2009; Amin and Suran, 2005), while a recent study suggested that when targeting out- of-school girls, younger is better, as older girls with more household responsibilities are less likely to enrol and more likely to drop out of the programme (Sieverding and Elbadawy, 2016). Finally, there are significant evidence gaps on programme length, intensity and legacy effects. While many evaluations provided some basic information about programme length and intensity, only three studies provided interesting information on intervention length and intensity. Assessing a girls’ empowerment programme in Egypt, the first study found that those exposed to the programme for a longer period of time showed the greatest 47 Interventions promoting adolescent girls’ economic capabilities: what works? change in attitudes about early marriage and decisionmaking, about childbearing, about gender roles, health knowledge, FGM/C, and in their level of self-confidence (Brady et al., 2007). The second study of another girls’ empowerment programme in India noted that those who attended regularly and were more exposed to the programme increased their awareness of rights and of reproductive health issues, their agency (including savings behaviour), and progressive gender role attitudes much more so than irregular programme participants (Acharya et al., 2009). Similarly, the third study of a youth SRH programme in India reported that participants with intensive exposure showed greater attitude and knowledge change (Kanesathasan et al., 2008). In addition, five studies offered some evidence on programme legacy effects. A tracer study of a longterm intervention suggested that graduates continued to experience the benefits of the programme even years after its completion, mostly due to the skills and knowledge they acquired (CEDPA, 2001). The second study reported that in order to support girls who had completed the programme, graduates’ groups were created in each village where girls could continue to meet every week and access additional livelihoods training (Acharya et al., 2009). The need to enable girls to maintain support was also highlighted by another evaluation with regards to girls who succeeded in re-entering formal education, yet also continued to be in need of social and academic support during the first year so that they did not drop out (Selim et al., 2013). Depending on their educational needs and wishes, girls in another intervention were allowed to remain in programme groups and to continue receiving training after completion of the basic curriculum (Erulkar et al., 2013). The final study, a small-scale follow-up conducted one year after programme completion, showed that while girls continued expanding their social network and sustaining their self-confidence and general health awareness, their SRH knowledge and financial behaviour deteriorated over time, indicating the need for refresher training (Engebretsen, 2013b). 48 The analysis of these 57 interventions seeking to improve adolescent girls’ economic capabilities provides interesting findings on three key issues: age-specific effectiveness and optimal age to intervene; the significance of intervention duration and intensity (or of specific components) to maximise effectiveness; and legacy effects for participants. While several studies have started investigating these issues and offer the evidence that follows, significant gaps remain to be addressed. 5.1 Timing of intervention and age-specific effectiveness Out of 57 interventions, only four focused exclusively on younger adolescent girls, while most targeted older adolescent girls. Most programmes targeting older girls provided vocational and/or business skills training to disadvantaged female and male youth aged 16–24, with mixed results: a few measured more positive effects for older adolescents than young adults, and for girls more than boys, in terms of increasing their likelihood of finding paid employment or improving their earnings; yet other such interventions have benefited young adults more. Evidence also suggests that effects may also be different in the short run as opposed to the medium or long run. Two studies of the same Latin American youth training intervention – the Peruvian ProJoven – which used data from different followup surveys conducted 6, 12 or 18 months after programme completion, identified considerable changes in effects, with those aged 16–20 years initially benefiting more but young adults benefiting later on (Diaz and Jaramillo, 2006), and female participants benefiting more than their male peers in the long run in terms of employment and total monthly earnings (Nopo et al., 2007). Multi-component programmes explicitly targeting adolescent girls and young women with vocational and/or business skills training and other support services also targeted older adolescents; we did not find any such programme that included girls aged 10–13. 5 Addressing issues of agespecific intervention effectiveness, duration and intensity, and legacy effects Key points • Few evaluations distinguished outcomes by age group, so clear evidence on age-specific effectiveness is limited. Younger adolescent girls may experience difficulty participating in financial interventions or understanding the themes of sexuality and reproduction. Older girls – especially married and working girls – often have increased responsibilities and are thus less likely to attend and complete a programme. • Most evaluations do not assess the effectiveness of an intervention’s duration and intensity. Some evidence suggests that increased programme exposure generates stronger effects. • Integrated interventions appear more promising and able to achieve a broader set of interlinked and mutually strengthening outcomes to empower girls compared to standalone interventions. • A number of intervention components, including mentors and life skills training, can play a critical role in enabling girls’ participation and contributing to positive programme effects. • Evidence on the legacy effects and the sustainability of reported change is scarce. Yet it is increasingly recognised that girls may need to continue accessing some support after programme completion. 49 Interventions promoting adolescent girls’ economic capabilities: what works? Integrated interventions seeking to empower adolescent girls targeted the whole adolescent age range – that is, girls aged 10–19. Clear evidence on their age-specific effectiveness is again limited as very few distinguish key outcomes – largely non-economic – by age group. For instance, younger girls appear to benefit more from interventions aiming to keep them in school and delay their marriage (Amin, 2011; Erulkar and Muthengi, 2009; Amin and Suran, 2005). On the other hand, younger girls, especially those aged 10–12, may experience difficulty understanding the themes of sexuality and reproduction (Muthengi et al., 2016; Erulkar and Medhin, 2014; Jarvis and Kabore, 2012). This may also be related to mentors’ difficulty teaching these themes and their tendency to avoid them, alongside the need for a more age-appropriate and less technical curriculum (ibid.). One recent evaluation of Ishraq in Egypt that targeted rural out-of-school girls recommended intervening early, as younger girls are more likely to participate and complete the programme with fewer distractions and household responsibilities (Sieverding and Elbadawy, 2016). Similarly, the ongoing evaluation of AGEP in Zambia emphasised that it was more challenging to ensure older girls’ participation and attendance, as they have more responsibilities and competing demands than their younger peers (Austrian et al., 2016). On the other hand, a more nuanced analysis of available participation data from two interventions suggested that even when programmes target both younger and older adolescent girls, younger adolescents and particularly those aged 10–12 may benefit less. For instance, the BALIKA evaluation in Bangladesh found that participation was highest among girls aged 14–15 (Amin et al., 2016). Likewise, data from Biruh Tesfa in Ethiopia also showed that girls aged 15–19 accounted for the majority, while those aged 11–14 comprised only 18% of participants (Erulkar et al., 2013). Similar conflicting evidence emerged from a number of programmes providing financial literacy and assets. Younger participants across diverse settings showed higher and more regular programme participation and savings and thus stronger financial effects (Austrian et al., 2016; Johnson et al., 2015; WWB, 2014a). However, other studies indicated that such interventions, especially those with microfinance support, work better with older adolescents and young adults, whereas safe spaces and voluntary savings schemes may indeed be more suitable for younger girls (Austrian and Muthengi, 2013; USAID, 2008; Erulkar and Chong, 2005). 5.2 Intervention duration and intensity to maximise effectiveness Most studies provided information on the length of the assessed interventions and of specific activities within them. For instance, integrated interventions with several components tend to last one year and include meetings where a female mentor provides training once or twice a week. Similarly, vocational training programmes typically last for almost a year and provide technical training for three to six months followed by a three-month internship. Several studies also mentioned the particular number of hours of training: the Employment Fund in Nepal, AGI in Rwanda and Ninaweza in Kenya, all provided girls and young women with 40 hours of life skills training (Chakravarty et al., 2016a; Botea et al., 2015; de Azevedo et al., 2013), while the Juventud y Empleo programme in the Dominican Republic provided 75 hours of such training (Ibarrarán et al., 2015). Both the aforementioned programmes in Kenya and the Dominican Republic also provided 150 hours of technical training (de Azevedo et al., 2013; Ibarrarán et al., 2015). However, they did not generate the same employment outcomes for participating youth. Most importantly, their impact evaluations did not assess the effectiveness of different intervention length and intensity. Thus even when such information is mentioned, it is rarely linked to programme effectiveness. Overall, intervention intensity needs to be compatible with the daily lives and time demands of participants. Failure to consider these aspects may lead to the exclusion of particular groups of adolescent girls or poorer outcomes for those who are unable to attend regularly. We actually found only six studies with some relevant information. An evaluation of Ishraq that compared different lengths of programme participation found that full-term participants (those who participated for 30 months) showed the greatest increase in academic skills and improvement in attitudes towards early marriage, lower fertility, FGM/C and gender roles (Brady et al., 2007). The evaluation of an intervention for unmarried adolescent girls in India measured programme effects for regular and nonregular participants and pointed out that the intervention was more effective for participants who took part in the activities regularly (defined as attending half or more of the sessions); these were the ones to demonstrate the greatest positive changes in agency, attitudes and awareness (Acharya et al., 2009). Likewise, participants who attended regularly (defined as three to four times a week) in a 56 different outcomes. While girls in Uganda focused on the programme’s economic components, girls in Tanzania had different priorities and were more interested in receiving educational support. Resource constraints also hindered implementation, mainly the inability to provide permanent club spaces, poor quality of club materials and repetitive activities unable to keep members’ interest, inadequate provisions for training of new mentors, and non-frequent monitoring and supervision visits (Buehren et al., 2015). This is linked to another considerable evidence gap. Most of the reviewed studies assessed interventions implemented in sub-Saharan Africa and South Asia, particularly in India, Ethiopia, Kenya, Uganda and Bangladesh. We found only three studies of interventions in the Middle East and North Africa – actually none in the Middle East – despite the persistent disadvantage that adolescent girls and young women face in the labour and financial markets in that region. Moreover, two of the three interventions had limited effects, either economic (due to difficult labour market conditions) or non-economic (particularly regarding girls’ agency due to restrictive gender norms). More evidence is thus necessary from some regions such as the Middle East and North Africa and about how to work effectively in conservative cultural settings, wherein girls face pervasive and persistent economic disadvantage and limited agency. To conclude, this review has provided evidence that interventions with economic components have promising potential to improve, directly and indirectly, the economic capabilities of adolescent girls in diverse settings. However, there are significant evidence gaps and thus we cannot claim that we know beyond doubt what works effectively. These gaps are closely linked to weak quality of evidence and lack of methodologically rigorous longitudinal evaluations that investigate short- and long-term effects, differentiate effects by age and location, assess the relative effectiveness of different intervention strategies, and identify the pathways of change, while considering the importance of context. GAGE thus provides a critical opportunity to fill these evidence and knowledge gaps and move the research and policy agenda forward. 57 Interventions promoting adolescent girls’ economic capabilities: what works? Primary studies assessing included interventions Acharya, R., Kalyanwala, S., Jejeebhoy, S.J. and Nathani, V. (2009) Broadening girls’ horizons: effects of a life skills education programme in rural Uttar Pradesh. New Delhi: Population Council and Prerana Adoho, F., Chakravarty, S., Korkoyah, D.T., Lundberg, M. and Tasneen, A. (2014) The impact of an Adolescent Girls Employment Program: the EPAG Project in Liberia. Policy Research Working Paper 6832. Washington DC: World Bank Aedo, C. and Nuñez, S. (2004) The impact of training policies in Latin America and the Caribbean: the case of Programa Joven. Washington DC: Inter-American Development Bank Alcid, A. (2014) A randomized controlled trial of Akazi Kanoze youth in rural Rwanda. Final evaluation report. Washington DC: United States Agency for International Development (USAID) and Education Development Center (EDC) Amin, S. (2011) Empowering adolescent girls in rural Bangladesh: Kishori Abhijan. New York: Population Council Amin, S., Ahmed, J., Saha, J., Hossain, M.I. and Haque, E.F. (2016) Delaying child marriage through community-based skills-de- velopment programs for girls: results from a randomized controlled study in rural Bangladesh. New York and Dhaka: Population Council Amin, S. and Suran, L. (2005) Program efforts to delay marriage through improved opportunities: some evidence from rural Bangladesh. New York: Population Council ASCI (2013) Evaluation of SABLA scheme: a report submitted to Ministry of Women and Child Development, Government of India. Hyderabad: Administrative Staff College of India (ASCI) Attanasio, O., Kugler, A. and 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 Austrian, K. and Muthengi, E. (2013) Safe and Smart Savings Products for Vulnerable Adolescent Girls in Kenya and Uganda: evaluation report. Nairobi: Population Council Austrian, K. and 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(2): 168–175 Austrian, K., Hewett, P.C., Soler-Hampejsek, E., Bozzani, F., Behrman, J.R. and Digitale, J. (2016) Adolescent Girls Empowerment Programme: research and evaluation mid-term technical report. Zambia: Population Council Bandiera, O., Buehren, N., Burgess, B., Goldstein, M., Gulesci, S., Rasul, I. and Sulaiman, M. (2015) Women’s empowerment in action: evidence from a randomized control trial in Africa (www.ucl.ac.uk/~uctpimr/research/ELA.pdf) Blattman, C., Fiala, N. and Martinez, S. (2014) ‘Generating skilled self-employment in developing countries: experimental evidence from Uganda,’ Quarterly Journal of Economics 129(2): 697–752 Botea, I., Chakravarty, S. and Haddock, S. (2015) The Adolescent Girls Initiative in Rwanda. Final evaluation report. Washington DC: World Bank Brady, M., Assaad, R., Ibrahim, B., Salem, A., Salem, R. and Zibani, N. (2007) Providing new opportunities to adolescent girls in socially conservative settings: the Ishraq Program in rural Upper Egypt. New York: Population Council Buehren, N., Goldstein, M., Gulesci, S., Sulaiman, M. and Yam, V. (2015) Evaluation of Layering Microfinance on an Adolescent Development Program for Girls in Tanzania. Unpublished manuscript. Card, D., Ibarrarán, P., Regalia, F., Rosas-Shady, D. and Soares, Y. (2011) ‘The labor market impacts of youth training in the Dominican Republic: evidence from a randomized evaluation’, Journal of Labor Economics 29(2): 267–300 CEDPA (2001) Adolescent girls in India choose a better future: an impact assessment. India: Centre for Development and Population Activities (CEDPA) Chakravarty, S., Lundberg, M., Nikolov, P. and Zenker, J. (2016a) The role of training programs for youth employment in Nepal: impact evaluation report on the Employment Fund. Policy Research Working Paper 7656. Washington DC: World Bank Cho, Y., Kalomba, D., Mobarak, A.M. and Orozco, V. (2015) Differences in the effects of vocational training on men and wom- en: constraints on women and drop-out behaviour. Working Paper No. 189. Geneva: International Labour Office Catholic Relief Services (CRS) (2009) Applying a girls' lens to the palette of integrated empowerment interventions. Addis Ababa: Catholic Relief Services and Nike Foundation de Azevedo, T.A., Davis, J. and Charles, M. (2013) Testing what works in youth employment: evaluating Kenya’s Ninaweza program. Washington DC: World Bank and International Youth Foundation Diaz, J.J. and Jaramillo, M. (2006) An evaluation of the Peruvian ‘Youth Labor Training Program’ -ProJoven. Working Paper: OVE/WP-10/06. Washington DC: Inter-American Development Bank Diaz, J.J. and Rosas, D. (2016) Impact evaluation of the job youth training program Projoven. IDB Working Paper 693. Washington DC: Inter-American Development Bank References 58 Dunbar, M.S., Dufour, M.S.K., Lambdin, B., Mudekunye-Mahaka, I., Nhamo, D. and Padian, N.S. (2014) ‘The SHAZ! project: results from a pilot randomized trial of a structural intervention to prevent HIV among adolescent women in Zimbabwe’, PLoS One 9(11): e113621 Edmeades, J., Hayes, R. and Gaynair, G. (2014) Improving the lives of married adolescent girls in Amhara, Ethiopia. A summary of the evidence. Washington DC: International Center for Research on Women (ICRW) Engebretsen, S. (2012) Baseline and endline findings of Filles Éveillées ('Girls Awakened'): a pilot program for migrant adolescent girls in domestic service. Cohort 1 (2011–2012), Bobo-Dioulasso. New York: Population Council Engebretsen, S. (2013a) Evaluation of Filles Éveillées ('Girls Awakened'): a pilot program for migrant adolescent girls in domestic service. Evaluation report. New York: Population Council Engebretsen, S. (2013b) Follow-up study of migrant adolescent girls in domestic service who participated in the first cohort of the Filles Éveillées (Girls Awakened) Program. New York: Population Council Erulkar, A.S. and Chong, E. (2005) Evaluation of a savings and micro-credit program for vulnerable young women in Nairobi. Nairobi: Population Council Erulkar, A. and Medhin, G. (2014) Evaluation of health and education impacts of a girls' safe spaces program in Addis Ababa, Ethiopia. Addis Ababa: Population Council Erulkar, A. and Muthengi, E. (2009) ‘Evaluation of Berhane Hewan: a program to delay child marriage in rural Ethiopia’, International Perspectives on Sexual and Reproductive Health 35(3): 6–14 Erulkar, A. and Tamrat, T. (2014) ‘Evaluation of a reproductive health program to support married adolescent girls in rural Ethiopia’, African Journal of Reproductive Health 18(2): 68–76 Erulkar, A., Ferede, A., Girma, W. and Ambelu, W. (2013) ‘Evaluation of ‘Biruh Tesfa’ (Bright Future) program for vulnerable girls in Ethiopia’, Vulnerable Children and Youth Studies 8(2): 182–192 Field, E., Glennerster, R., Nazneen, S., Pimkina, S., Sen, I. and Buchmann, N. (2016) Age at marriage, women’s education and mother and child outcomes in Bangladesh. 3ie Grantee Final Report. New Delhi: International Initiative for Impact Evaluation (3ie) Govender, K., Rogan, M. and Hallman, K. (n.d.) Siyakha Nentsha - ‘Building with young people’: enhancing the health, economic, and leadership capabilities of highly vulnerable youth (http:// doc.ukdataservice.ac.uk/doc/7006/mrdoc/pdf/7006_si- yakhanentshacele5nearfinalupdated.pdf) Gray, B. and Chanani, S. (2010) Advancing women's and adolescent girls' access to resources and influence in rural India. Freedom from Hunger Research Report No. 11. India: Freedom from Hunger and Reach in India Hallman, K. and Roca, E. (2011) Siyakha Nentsha: building economic, health, and social capabilities among highly vulnerable adolescents in KwaZulu-Natal, South Africa. New York: Population Council Hamilton, M.A., Hamilton, S.F., Bianchi, L. and Bran, J. (2013) ‘Opening pathways for vulnerable young people in Patagonia’, Journal of Research on Adolescence 23(1): 162–170 Hicks, J., Kremer, M., Mbiti, I. and Miguel, E. (2015) Vocational education in Kenya – a randomized evaluation. 3ie grantee final report. New Delhi: International Initiative for Impact Evaluation (3ie) Honorati, M. (2015) The impact of private sector internship and training on urban youth in Kenya. Policy Research Working Paper 7404. Washington DC: World Bank Ibarrarán, P., Ripani, L., Taboada, B., Villa, J.M. and Garcia, B. (2014) ‘Life skills, employability and training for disadvantaged youth: evidence from a randomized evaluation design’, IZA, Journal of Labor and Development 3: 10 Ibarrarán, P., Kluve, J., Ripani, L. and Shady, D.R. (2015) Experimental evidence on the long term impacts of a youth training program. RUHR Economic Paper #562. Bochum: Ruhr-Uni- versität Bochum (RUB) Jarvis, L. and Kaboré, G. (2012) Process evaluation: the Filles Éveillées (Girls Awakened) program for migrant adolescent girls in domestic service in urban Burkina Faso. Ouagadougou: Population Council Johnson, L., Lee, Y., Ansong, D., Sherraden, M., Chowa, G., Ssewamala, F. and Zou, L. (2015) Youth savings patterns and performance in Colombia, Ghana, Kenya, and Nepal. Youth- Save Research Report. Washington DC: Center for Social Development Jurgens, E., Isolio, M., Jadi, C. and Elia, L. (2013) Report on qualitative study of the Adolescent Girls Initiative in South Sudan. London: Integrity Research and Consultancy Kamruzzaman, M., Hossain, N.Z., Bhattacharjee, A., Akter, A., Raza, W., Malek, M.A. and Das, N.C. (2012) Social network and financial literacy among rural adolescent girls: qualitative assessment of BRAC’s SoFEA programme. Dhaka: BRAC Research and Evaluation Division Kanesathasan, A., Cardinal, L.J., Pearson, E., Das Gupta, S., Mukherjee, S. and Malhotra, A. (2008) Catalyzing change: improving youth sexual and reproductive health through DISHA, an integrated programme in India. Washington DC: ICRW Kim, J.C., Watts, C.H., Hargreaves, J.R., Ndhlovu, L.X., Phetla, G., Morison, L.A., Busza, J., Porter, J.D.H. and Pronyk, P. (2007) ‘Understanding the impact of a microfinance-based intervention on women's empowerment and the reduction of intimate partner violence in South Africa’, American Journal of Public Health 97: 1794–1802 59 Interventions promoting adolescent girls’ economic capabilities: what works? Kim, J., Ferrari, G., Abramsky, T., Watts, C., Hargreaves, J., Morison, L., Phetla, G., Porter J. and Pronyk, P. (2009) ‘Assessing the incremental effects of combining economic and health interventions: the IMAGE study in South Africa’, Bulletin of the World Health Organization 87: 824–832 Kluve, J., Bausch, J., Gardiner, D. and Mizrokhi, E. (2016) The impact of skills training on financial behaviour, employability, and educational choice of youth in Morocco. 3ie grantee final report. New Delhi: International Initiative for Impact Evaluation (3ie) Landesa RDI (2013) Security for Girls through Land: pilot evaluation (2012-2013). India: Landesa Rural Development Institute (RDI) Maitra, P. and Mani, S. (2014) Learning and earning: evidence from a randomized evaluation in India. IZA Discussion Paper 8552. Bonn: IZA MCI and Save the Children (2012) Lessons learned of pilot financial literacy project for adolescent girls in Haiti. Washington DC: Making Cents International (MCI) and Save the Children Haiti Mekbib, T.A. and Molla, M. (2010) ‘Community based reproductive health (RH) intervention resulted in increasing age at marriage: the case of Berhane Hewan project, in East Gojam zone, Amhara region, Ethiopia’, Ethiopian Journal of Reproductive Health 4(1): 16–25 Microfinanza (2015) Final evaluation of the YouthStart program. Vicenza, Italy: Microfinanza Miller, C., Swayer, M. and Rowe, W.-A. (2011) My skills, my money, my brighter future in Zimbabwe. An assessment of economic strengthening interventions for adolescent girls. Baltimore: Catholic Relief Services Muthengi, E., Austrian, K., Landrian, A., Abuya, B.A., Mumah, J. and Kabiru, C.W. (2016) Adolescent Girls Initiative-Kenya: qualitative report. Nairobi, Kenya: Population Council Nopo, H., Robles, M. and Saavedra, J. (2007) Occupational training to reduce gender segregation: the impacts of ProJoven. Washington DC: Inter-American Development Bank Pronyk, P.M., Kim, J.C., Abramsky, T., Phetla, G., Hargreaves, J.R., Morison, L.A., Watts, C., Busza, J. Porter, J.D. (2008) ‘A combined microfinance and training intervention can reduce HIV risk behaviour in young female participants’, AIDS 22(13): 1659–1665 Ramadan, N., Abdel-Tawad, N., El Sayed, K. and Roushdy, R. (2014) Enhancing livelihood opportunities for young women in rural Upper Egypt: the Neqdar Nesharek program. Cairo: Population Council Rodella, A.-S., Cuevas, F., Atuesta, B., Ferrer-Rincon, C. and Scot, T. (2015) Haiti Adolescent Girl Initiative (AGI): project report. Washington DC: World Bank Rowe, W.-A. and Miller, C. (2011) My skills, my money, my brighter future in Rwanda. An assessment of economic strengthening interventions for adolescent girls. Baltimore: Catholic Relief Services Rushdy, S. (2012) Analysis of the ‘Ishaka’ experience (www. care.org/sites/default/files/documents/MF-2012-Analy- sis-of-the-Ishaka-Experience.pdf) Sebastian, M.P., Grant, M. and Mensch, B. (2005) Integrating adolescent livelihood activities within a reproductive health programme for urban slum dwellers in India. New Delhi: Population Council Selim, M., Abdel-Tawab, N., Elsayed, K., El Badawy, A. and El Kalaawy, H. (2013) The Ishraq program for out-of-school girls: from pilot to scale-up. Cairo: Population Council SEWA, SEWA Academy, Kalyanwala, S. et al. (2006) Influencing girls’ lives: acceptability and effectiveness of a livelihoods skill building intervention in Gujarat. New Delhi: Population Council Shahnaz, R. and Karim, R. (2008) Providing microfinance and social space to empower adolescent girls: an evaluation of BRAC’s ELA centres. Dhaka: BRAC Research and Evaluation Division Sieverding, M. and Elbadawy, A. (2016) ‘Empowering adolescent girls in socially conservative settings: impacts and lessons learned from the Ishraq program in rural Upper Egypt’, Studies in Family Planning 47(2): 129–144 Spielberg, F., Crookston, B., Chanani, S., Kim, J., Kline, S. and Gray, B. (2010) Leveraging microfinance networks to scale up HIV and financial education among adolescents and their mothers in West Bengal: a cluster randomized trial and mixed-method evaluation. Freedom from Hunger Research Paper No. 15. CA, USA: Freedom from Hunger Ssewamala, F.M., Ismayilova, L., McKay, M., Sperber, E., Bannon, W., Jr. and Alicea, S. (2010) ‘Gender and the effects of an economic empowerment program on attitudes toward sexual risk-taking among AIDS-orphaned adolescent youth in Uganda’, Journal of Adolescent Health 46(4): 372–378 Tower, C. and McGuiness, E. (2011) Savings and Financial Education for Girls in Mongolia. Impact Assessment Study. Washington DC: Microfinance Opportunities. Women’s World Banking (WWB) (2014a) Final report: Banco ADOPEM/ WWB case study-based research project. New York: Women’s World Banking (http://womensworldbanking. org/youthsavingsguidelinks-EN/Ch4_1_WWBDRCaseStud- iesSummaryReport_Final.pdf) Secondary studies and reviews Baird, S. and Ozler, B. (2016) Sustained effects on economic empowerment of interventions for adolescent girls: existing evidence and knowledge gaps. CGD Background Paper. Washington DC: Center for Global Development Blanc, A.K., Melnikas, A., Chau, M. and Stoner, M. (2013) A review of the evidence on multi-sectoral interventions to reduce violence against adolescent girls. London: Girl Hub, Department for International Development (DFID) 60 Chakravarty, S., Haddock S. and Botea, I. (2016b) Providing out- of-school girls with skills: a review of the global evidence. Policy Brief. Washington DC: World Bank Deshpande, R. and Zimmerman, J.M. (eds.) (2010) Youth savings in developing countries: trends in practice, gaps in knowledge. A report of the YouthSave Consortium. Ontario: The MasterCard Foundation, Save the Children, Centre for Social Development, New America Foundation, and Consultative Group to Assist the Poor Dickson, K. and Bangpan, M. (2012) Providing access to economic assets for girls and young women in low- and lower-mid- dle-income countries. A systematic review of the evidence. London: EPPI-Centre, Social Science Research Unit, Institute of Education, University of London Fulu, E., Kerr-Wilson, A. and Lang, J. (2014) What works to prevent violence against women and girls? Evidence review of interventions to prevent violence against women and girls. London: DFID Hallman, K.K. (2016) Population Council lessons on increasing and measuring girls’ economic empowerment. New York: Population Council Hirschland, M. (2009) Youth savings accounts: a financial service perspective. A literature and program review. MicroReport#163. Washington DC: USAID Katz, E. (2008) Programs promoting young women’s employment: what works? Washington DC: World Bank Quisumbing, A.R. and Kovarik, C. (2013) Investments in adolescent girls’ physical and financial assets: issues and review of evidence. London: DFID and Girl Hub Sebstad, J. (2011) Girls and their money: strategies for promoting savings, financial education and social support for adolescent girls in low-income countries. Synthesis report. Washington DC: Microfinance Opportunities USAID (2008) Reducing adolescent girls' vulnerability to HIV infection: examining microfinance and sustainable livelihood approaches. A literature and program review. Washington DC: USAID World Bank (2015) Adolescent Girls Initiative: pilot implementation and results. Washington DC: World Bank WWB (2014b) Banking on youth. A guide to developing innovative youth savings programs. New York: Women’s World Banking (WWB) and the Nike Foundation Other sources Chaaban, J. and Cunningham, W. (2011) Measuring the economic gain of investing in girls. The girl effect dividend. Policy research Working Paper 5753. Washington DC: World Bank Elder, S. and Kring, S. (2016) Young and female – a double strike? Gender analysis of school-to-work transition surveys in 32 developing countries. Geneva: International Labour Office and the MasterCard Foundation ILO (2003) International training compendium on labour statistics. Statistics of employment, unemployment, underemployment: economically active population. Geneva: International Labour Office ILO (2017a) Global estimates of child labour. Results and trends, 2012-2016. Geneva: International Labour Office ILO (2017b) Global employment trends for youth 2017. Paths to a better working future. Geneva: International Labour Office ILO (2017c) World employment social outlook. Trends for women 2017. Geneva: International Labour Office Kabeer, N. (1999) ‘Resources, agency, achievements: reflections on the measurement of women’s empowerment’, Development and Change 30: 435-464 Sykes, J., Elder, S., Gurbuzer, Y. and Principi, M. (2016) Exploring the linkages between youth financial inclusion and job creation. Evidence from ILO school-to-work transition surveys. Geneva: International Labour Office and the MasterCard Foundation UNESCO (2012) EFA global monitoring report 2012. Youth and skills: putting education to work. Paris: UNESCO UNICEF (2011) The state of the world’s children 2011. Adolescence: an age of opportunity. New York: UNICEF 61 Interventions promoting adolescent girls’ economic capabilities: what works? Annex 1: Search strategy The search strategy included: • academic databases and academic journals • grey literature • expert advice and reference snowballing. Search strings were composed of keywords referring to four areas: a) adolescence and gender; b) economic capabilities; c) intervention; and d) focus on low- and middle-income countries (LMICs). Search terms (adolescen* OR girl* OR teenage* OR youth*) AND (land OR livestock OR asset* OR homestead plot* OR transfer* OR loan* OR credit OR saving* OR skill* OR training OR employment) AND (approach* OR evaluat* OR initiative* OR intervention* OR model* OR package* OR pilot* OR program* OR project* OR provision* OR trial* OR scheme* OR strategy*) AND (developing countr* OR middle income OR low* income) Academic databases and academic journals • EBSCOhost • Web of Science • Journal of Development Effectiveness Grey literature • BRAC Evaluation Database • CARE International • Catholic Relief Services (CRS) • International Labour Organization (ILO) • International Center for Research on Women (ICRW) • Overseas Development Institute (ODI) • Population Council • UN Women • United States Agency for International Development (USAID) • World Bank Open Knowledge Repository • 3ie Evaluation Database Expert advice and reference snowballing We asked experts with knowledge of the programming landscape in this area to indicate relevant studies that previous searches had not captured. Expert advice as well as reference snowballing have led us to include additional studies, mostly grey literature, from various sources, including BRAC Bangladesh, the UN Capital Development Fund and WWB (Women’s World Banking). Table A1.1: Search strings All databases All databases All databases All databases 1. Adolescent Terms AND 2. Economic Capabilities Terms AND 3. Intervention/Study Type Terms AND 4. Developing Countries Terms adolescen* land approach* developing countr* girl* livestock evaluat* middle income teenage* asset* initiative* low* income youth* homestead plot* intervention* transfer* model* loan* package* credit pilot* saving* program* skill* project* training provision* employment trial* scheme* strateg* 62 Table A1.2: Inclusion and exclusion criteria Parameters Inclusion criteria Excluson criteria Location LMICs in Africa, Asia, Latin American and the Caribbean High-income countries Any country in Europe, North America and Oceania Language English Any language other than English Timeframe Published between 2000 and 2016 Published before 2000 and after 2016 Population Adolescent girls aged 10–19 Every person who is not an adolescent girl aged 10–19 Type of intervention Any intervention that aims to strengthen the economic capabilities of adolescent girls where economic capabilities refer to forms of capabilities which improve their livelihoods opportunities during relevant periods of their life-course. They include land and other economic assets, life skills and other forms of market-related training and education, cash transfers, loans and savings and livelihoods opportunities. Any intervention not aimed at strengthening adolescent girls’ economic capabilities AND • cash transfers for education • interventions explicitly aimed at changing social norms Type of study Primary research: experimental, quasi-experimental, observational; Secondary research: systematic reviews, other reviews Theoretical or conceptual 63 Interventions promoting adolescent girls’ economic capabilities: what works? Annex 2: Key characteristics of intervention groups Table A2.1: Evaluation methodology Evaluation methodology Number of interventions (15) Mixed methods 11 Quantitative 2 Qualitative 2 RCT 6 Quasi-experimental 3 Table A2.2: Target group by location Location Number of interventions (15) Urban 2 Urban and rural 6 Urban slums 2 Rural 4 Not specified 1 Table A2.3: Target group by age Age group Number of interventions (15) Younger adolescent girls (aged 10–14) 3 Older adolescents girls (aged 15–19) 5 Adolescent girls (aged 10–19) 7 Table A2.4: Target group by type of vulnerability Vulnerable group Number of interventions (15) Married adolescent girls 2 Orphaned girls 1 Out-of-school girls 2 Figure A2.1: Regional distribution of interventions Sub-Saharan Africa Multiple countries South Asia Latin America & Caribbean East Asia, South-East Asia & Pacific 0 1 2 3 4 5 6 7 7 3 2 2 1 Figure A2.2: Regional distribution of interventions Sub-Saharan Africa South Asia Latin America & Caribbean Middle East & North Africa 0 2 4 6 8 10 12 12 6 5 2 A2.1 Financial education and/or assets interventions A2.2 Vocational and/or business skills training interventions Table A2.5: Evaluation methodology Evaluation methodology Number of interventions (25) Mixed methods 14 Quantitative 9 Qualitative 2 RCT 16 Quasi-experimental 6 Table A2.6: Target group by location Location Number of interventions (25) Urban 6 Urban and rural 7 Urban slums 3 Rural 5 Not specified 4 64 Table A2.7: Target group by age Age group Number of interventions (25) Younger adolescent girls (aged 10–14) 0 Older adolescents girls (aged 15–19) 22 Adolescent girls (aged 10–19) 2 Not specified 1 Table A2.8: Target group by type of vulnerability Vulnerable group Number of interventions (25) Married adolescent girls 5 Orphaned girls 1 Out-of-school girls 8 Figure A2.3: Regional distribution of interventions Sub-Saharan Africa South Asia Middle East & North Africa 0 2 4 6 8 10 1 9 7 A2.3 Integrated interventions with an economic component Table A2.9: Evaluation methodology Evaluation methodology Number of interventions (17) Mixed methods 7 Quantitative 8 Qualitative 2 RCT 3 Quasi-experimental 8 Table A2.10: Target group by location Location Number of interventions (17) Urban 3 Urban and rural 4 Urban slums 2 Rural 9 Please note that 1 intervention was implemented in both rural and urban slum areas, thus it is counted twice. Table A2.11: Target group by age Age group Number of interventions (17) Younger adolescent girls (aged 10–14) 1 Older adolescents girls (aged 15–19) 4 Adolescent girls (aged 10–19) 12 Table A2.12: Target group by type of vulnerability Vulnerable group Number of interventions (17) Married adolescent girls 3 Orphaned girls 3 Out-of-school girls 8 65 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Financial education and/or assets interventions Adolescent Girls Empowerment Programme (AGEP) Zambia Vulnerable girls aged 10–19 in urban and rural areas. The 2-year programme had 3 distinct components: the core intervention consisting of safe spaces groups, trained mentors, life skills training, sexual and reproductive health (SRH) information and financial education; this basic package and a health voucher; and the basic package with the health voucher and a savings bank account. Austrian et al. (2016). RCT with 4 arms: the 3 intervention arms and the control arm. Qualitative research was also used. AGEP also conducted a mentor quality assessment. This study presents findings from the midline. Economic outcomes: Positive financial effects for all participants compared to the control group, but girls in the third arm had higher financial literacy scores, and those who opened an account – even if they did not use it much – were more likely to save both in their account and informally, and in higher amounts compared to those without an account. Younger girls in rural areas, who were also more engaged in the programme, were significantly more likely to open an account, followed by younger urban girls. Older urban girls were significantly more likely to have worked for cash or in-kind. Bodily integrity: No change in norms about gender-based violence. Health: Participants increased their SRH knowledge – statistically significant increase in contraceptive knowledge. They were also more likely to use a condom during their first sexual experience, and were less likely to engage in transactional sex compared to the control group (a statistically significant finding). Yet there was no change in rates of sexually transmitted infections (STIs) or unwanted pregnancies and in age of sexual initiation. Only 1 in 5 girls used the voucher and onethird of girls who used it did so for SRH services. Psychosocial wellbeing: Participants improved their self-confidence, and reduced their social isolation. Younger, rural adolescent girls showed stronger impact, perhaps as they had higher participation rates. Table A3.1: Overview of interventions, evaluation methodology and key findings Annex 3: Overview of interventions 72 Title and country Target group Intervention activities Study Evaluation methodology Key findings Hallman and Roca (2011). Two rounds of panel survey before programme start and 18 months thereafter; also qualitative interviews with participants. Two versions of the programme were tested – one focused on the basic package addressing social and health capabilities; the other was an enhanced version with the added financial training, both compared against a control group, which received standard life skills training. Economic outcomes: All programme participants (regardless of gender or programme version) reported a large increase in knowledge of social grants requirements and criteria, and improved budgeting and planning skills. They were also more likely to have attempted to open a bank account compared to the control group. Boys who received the enhanced package were more likely to report undertaking an income-generating activity compared to those who received the standard package. Health: All programme participants were more likely than the control group to know a place to get condoms. Boys were more likely to have remained sexually abstinent and reported having fewer sexual partners. Girls reported feeling higher self-esteem and greater confidence in their ability to find a condom. Psychosocial wellbeing: Girls who received the full package (health + finance) felt greater levels of social inclusion in their communities than girls who received the standard package. Social and Financial Empowerment of Adolescents (SoFEA) Bangladesh Adolescent girls aged 11–21, including married girls, in rural areas. The programme provided: a secure place for adolescents to socialise; life skills training, including financial education; savings and credit facilities; and community sensitisation. Kamruzzaman et al. (2012). Qualitative IDIs and FGDs with participants. Economic outcomes: Girls reported improved financial literacy at and line, and increased involvement in economic activities and ownership of small livestock. Education: Programme participation did not affect girls’ school performance. Psychosocial wellbeing: Participants reported improved relationships with family and relatives; value given to having a space where they could interact with their peers and make new friends. Voice and agency: Participants became more confident to express their opinion. 73 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Suubi (‘Hope’) Research Programme Uganda Adolescents who had lost one or both of their parents due to HIV/AIDS, average age 13.7, rural areas. The overall aim was to develop and test the feasibility of an intervention for improving the health outcomes and life chances of AIDS-orphaned adolescents through the use of microfinance—in the form of Child Savings Accounts (CSAs). CSA programmes, including Suubi, involved youth in basic financial education, introduced them to formal financial institutions, and incentivised them to save small amounts by offering a subsidised match rate. Ssewamala et al. (2010). RCT. Participants assigned to the control condition received the usual care for orphaned children – counselling and education-related supplies. By virtue of being in school, all participants received health education. Participants assigned to the experimental condition (Suubi) received the usual care plus an economic empowerment intervention consisting of 3 components: 1) twelve 1–2-hour workshops over a 10-month period, on asset building and financial planning. The workshops introduced the participants to asset-building strategies, including saving, education, and small business development; 2) a monthly mentorship programme with peer mentors; and 3) a matched CSA, dedicated to paying for post-primary schooling. Economic outcomes: Girls and boys saved comparable amounts. This may indicate that caregivers perceived both saving money and pursuit of secondary education to be as desirable for girls as for boys. Health: Girls did not experience equally protective improvements in attitudes toward sexual risk–taking behaviours. Boys benefited more from the intervention than girls. 74 Title and country Target group Intervention activities Study Evaluation methodology Key findings Tap and Reposition Youth (TRY) Savings and Micro-Cre- dit for Adolescent Girls Kenya Out-of- school girls and young women aged 16–22 in Nairobi. The programme included savings, microcredit, training in business and life skills, reproductive health education, and mentoring by adults from the community. Erulk- ar and Chong (2005). Quasi-experiment. Structured interviews at baseline and end line. Economic outcomes: Participants increased household assets at end line, increased their proportion of working for pay, increased their weekly earnings, had significantly more savings and were more likely to keep savings in a safer place than the control group. Health: No significant effect on reproductive health knowledge, which increased among both TRY girls and control groups, likely due to age. Yet some evidence indicating that participants had greater ability to refuse sex and insist on condom use than the control group. Voice and agency: Participants showed positive improvements in gender attitudes. Problems: High levels of drop-out with younger girls more likely to do so than older ones. Programme has worked better for young women aged 20 and over. Temuulel (‘Aspire’) programme Mongolia Adolescent girls aged 13–17 years. Temuulel included financial education classes and a savings product. Tower et al. (2011). Quasi-experiment, comparison of 3 groups: savings-only girls, who had an account; financial education + savings girls, who completed the financial education training programme and opened an account; and a control group of girls who did not have an account or financial education training. Also, qualitative interviews. Economic outcomes: Positive effects on knowledge of financial services, increased control of savings, more open attitudes towards money and financial products. Increased savings, with varied patterns of contribution to the families. Psychosocial wellbeing: Temuulel provided girls with the opportunity to make new friends. YouthSave Colombia, Ghana, Kenya and Nepal Low-income youth aged 12–18. A large savings initiative that provided financial education and account opening services. Johnson et al. (2015). Mixed methods using quantitative and qualitative techniques. Economic outcomes: Participants younger than 13 years saved more than older youth, partly because they withdrew less. When parents were the co-signatory, youth saved significantly more. No gender differentials in savings behaviour were identified; girls saved as much as boys and sometimes even more. The greater obstacle girls faced was not saving itself, but access to financial institutions. The situation was even more difficult for out- of-school youth, including girls, who were harder to reach and perhaps needed more tailored products and services to ensure their formal financial inclusion. 75 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings YouthStart Burkina Faso, Democratic Republic of Congo, Ethiopia, Malawi, Rwanda, Senegal, Togo, and Uganda Young people aged 12–24. YouthStart provided capacity-building and information-sharing to promote financial services for youth – current savings accounts and individual/group credit. Microfinanza (2015). Quantitative and qualitative assessment (interviews with finance providers and clients). Economic outcomes: Programme aimed to reach girls and young women, yet no strategy was developed to do so – no best practice to reach female clients can be identified and thus replicated in other contexts and/or countries. Market studies, staff interviews and FGDs with clients all showed that female youth had the same needs as male youth around financial products and financial literacy training – i.e. the differentiation of product/service characteristics according to gender does not seem to have been perceived as necessary. Very limited programme influence. No evidence of YouthStart having affected the youth financial inclusion environment in any of the countries. Possible positive impacts on girls’/women’s financial capabilities and savings. 76 Title and country Target group Intervention activities Study Evaluation methodology Key findings Vocational and/or business skills training interventions Adolescent Girls Initiative (AGI) Haiti Vulnerable women aged 17–21 years, out of school for over a year, with basic literacy and numeracy, living in Portau-Prince. AGI provided soft skills and technical skills training in sectors not traditionally associated with women. Also provided a stipend to cover transportation costs and other attendance-re- lated expenses. Rodella et al. (2015). The programme was evaluated through an RCT and qualitative studies. The RCT included eligible women randomly assigned to participate in 1 of 2 cohorts: the first included trainees, while the second was the control group, which received training in the second round. A baseline was conducted before cohort assignment, while 3 months after completion of the training of the first round of the treatment group, a follow-up survey took place. This study used data from these two surveys. The impact evaluation was completed with 2 more rounds of data collection. Economic outcomes: 3 months after programme completion, the share of graduates engaged in income-generating activities declined and their earnings did not change. Yet graduates were changing the type of work they did, transitioning towards higher-skilled jobs related to their training. Graduates faced various challenges in their job search, including limited relevant job openings, which had already been filled or were publicised mainly through word-to-mouth and social networks. Education: Participants developed higher expectations towards school enrolment than the control group. Bodily integrity: Graduates faced sexual advances during job search. Psychosocial wellbeing: Mental stress levels declined and graduates managed their lives better. They also benefited from expanding their support network and increased parental support of wage employment. Graduates also increased their self-assertiveness and self-esteem. Voice and agency: Graduates experienced increased personal mobility compared to the control group. They also reported that the training helped make them more in control of employment choices. They were also less likely to accept IPV. Trainees showed an increase of 7 percentage points in the belief that income generation is a shared responsibility between women and men. Yet no change was recorded on beliefs of co-responsibility for domestic tasks and childcare. Their aspirations for more education and better future employment increased compared to the control group. They also increased their decision-ma- king and social status. Success factors: AGI paid attention to gender safety and security issues, including the need to limit trainees’ travel time to reduce risk of assault. 77 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Adolescent Girls Initiative (AGI) Rwanda Vulnerable girls and young women, aged 16–24, out of school for at least 1 year, with some primary education, urban and rural areas. AGI Rwanda had 2 main components: provision of scholarships for formal education; and vocational training and entrepreneurship support. The latter included 2-week induction (including 40 hours of life skills training), 6-month technical skills training, and 5.5 month placement and mentoring support. Girls were also assisted to open individual accounts, and were provided with a stipend and childcare facilities. Botea et al. (2015). The skills and entrepreneurship support component was evaluated using a tracer methodology following 160 trainees before, during and after the programme and combining quantitative and qualitative research tools, such as surveys, FGDs and semi-struc- tured interviews. Economic outcomes: Graduates were 1.5 times more likely to have non-farm employment – a 25 percentage point increase driven mainly by higher self-em- ployment, including small businesses. Those who attended arts and crafts training were less likely to use their training. Participants also increased their earnings significantly. They were also more likely to be members of savings groups, to save up, and save large amounts. Many used part of their savings to buy livestock (livestock ownership tripled). Ownership of mobile phones also increased. Bodily integrity: Participants reported increased sexual harassment. Psychosocial wellbeing: Participants reported increased social capital and moderate improvements in their relationships with family, friends and community members. They reported higher satisfaction with their lives and greater optimism for the future. They also spoke about increased self-esteem and self-acceptance. Voice and agency: No change in time spent on household chores. Improved entrepreneurial self-confidence. Increased ability to make decisions about their work, training and household tasks. Increased joint decision-making between girls and their partners. 78 Title and country Target group Intervention activities Study Evaluation methodology Key findings Adolescent Girls Initiative (AGI) South Sudan Girls and young women aged 15–24. AGI South Sudan provided safe spaces, livelihoods and financial literacy training, savings and credit services, life skills and community sensitisation. Jurgens et al. (2013). Qualitative methodology with FGDs and KIIs. Economic outcomes: Skills training was based on local market demands, but trainers and employers noted that it was often insufficient as trainees lacked the capital they needed to access the equipment to start their business. Girls improved their financial literacy, and financial and business management, and increased their financial independence. The programme helped over 80% of participants to apply for a small business loan. Yet some girls encountered problems with repayment and others had to hide money from their husbands to avoid having them misuse it. Bodily integrity: Access to loans made some participants vulnerable to domestic violence as their partners tried to get hold of the money and threatened girls or even used violence. Health: Reduction in early pregnancy rates. Voice and agency: Girls increased their mobility, visiting girls’ clubs and engaging in community events. They also reported increased levels of confidence to talk about family planning and sensitive issues with their partners and enhanced decision-making ability for SRH issues. Girls who received a loan also reported being able to decide themselves for their business and how to invest their money. 79 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Adolescent Livelihood and Reproductive Health Programme India Adolescent girls aged 14–19 in urban slums of Allahabad, Uttar Pradesh. The 19–month intervention provided groups of girls with reproductive health education from a trained peer educator at her home for 7–10 weeks, followed by livelihoods counselling, vocational training, assistance with savings formation and follow-up counselling and support. Sebastian et al. (2005). Quasi-experimen- tal design with girls in treatment areas receiving all components, while girls in control areas received only the reproductive health education. Both girls and boys in study areas participated in baseline and end line surveys, but the analysis focused on 3 groups of girls: those in control areas; those in treatment areas who did not attend the livelihoods component; and those who did so. A survey of parents also took place. IDIs with adolescents and 1 adult in the household (often the parent) were also conducted to supplement the end line survey. Economic outcomes: More than 80% of participants in the treatment group continued to use their vocational skills after programme completion. Yet only 10% were able to earn an income using those skills. More than half opened savings accounts in their name. Bodily integrity: The few girls who started working noted being harassed by boys and men. Health: Improved reproductive health knowledge compared to the control group. Psychosocial wellbeing: There was no significant change in girls’ self-esteem. Yet at end line, 6 times as many participants belonged to a group compared to the control group. Voice and agency: Compared to the control group, girls in treatment areas who attended the intervention were significantly more likely to know about safe locations where unmarried girls could meet, which was used as a proxy for girls’ mobility and its acceptability. They were thus more likely to increase their mobility, be members of a group at end line, and attain higher social skill scores (including communication skills). They improved their skills in negotiating and getting better prices for their products. Yet attitudes around the acceptability of female wage labour, shared domestic responsibilities and recognition of women’s value did not improve. 80 Title and country Target group Intervention activities Study Evaluation methodology Key findings Akazi Kanoze (‘Work well done’) Rwanda Youth aged 14–35, with low education levels, in urban and rural areas. The programme provided 5-month work readiness training and technical skills training along with 3-mon- th internships. Alcid (2014). RCT and qualitative research (case studies and FGDs). The RCT included eligible rural youth who were randomly assigned to a treatment and control group. Follow-up survey took place 4 months after programme completion. Economic outcomes: Although employment rates overall declined, programme graduates in rural areas showed a smaller fall and were 12% more likely to be employed than the control group. The programme also improved their competency in major work readiness skills: graduates were better at finding new employment faster; knew how to apply for a job or improve their current position; understood business plan development; and felt comfortable using marketing techniques. Although female participants started further behind men, they almost completely caught up over the course of a year. Women reported significantly greater increases in knowledge about applying for and finding work. No other significant gender differences were identified. Graduates were more likely to save and significantly increased their frequency of saving compared to the control group. Psychosocial wellbeing: Participants reported greater confidence and optimism about their job and business opportunities. They also reported statistically significant increases in having mentors who encouraged them and believed in their skills. 81 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Economic Empowerment of Adolescent Girls and Young Women (EPAG) Liberia Adolescent girls and young women aged 16–27 with basic literacy and numeracy skills, out of school, in target communities in and around Monrovia. EPAG provided 6-month technical and life skills training and 6-month placement and support for self or wage employment. The programme had 2 tracks: job skills (which provided training in 6 areas); and business development services (focused on business creation and management). Trainees were also provided with a small stipend, a completion bonus, free childcare at each training site, assistance to open a savings account, and a mentor. Adoho et al. (2014). The impact evaluation of the first round of training used an RCT: girls who were assigned by lottery to the first cohort/round of training were the treatment group, and those assigned to the second cohort/round served as the control group. Data was collected using 3 quantitative surveys and 2 sets of qualitative FGDs. This study was based on panel data from the baseline and the midline surveys and monitoring data from programme implementation. Economic outcomes: 47% increase in employment and 80% growth in earnings of trainees compared to the control group. The programme was more successful in promoting self-employment than wage employment, with those in the business skills track nearly twice as likely to find employment as those in the job skills track due to fewer wage opportunities available. Moreover, business development trainees more than doubled their earnings. Programme economic impact was greater for older adolescent girls (aged 16–19) than for younger girls. Trainees were also nearly 50 percentage points more likely to have savings and save more than the control group. Health: No net impact on fertility or sexual behaviour, but improved food security and dietary composition in trainees’ households. Psychosocial wellbeing: Trainees reported increased sense of control and confidence in personal business ability along with lower anxiety about jobs, income and relationships, and improved satisfaction with one’s job or business. Voice and agency: A small but significant impact on control over resources and earnings. Trainees also reported high aspirations, and increased entrepreneurial self-efficacy. There was also a small but significant change in gender-egalitarian attitudes (on the division of household chores) among the heads of participating households. 88 Title and country Target group Intervention activities Study Evaluation methodology Key findings Kenya Youth Empowerment Programme, also known as Ninaweza Kenya Young women aged 18–35 who had completed high school, were out-of- school for a year, were unemployed and living in any of the targeted informal settlements around Nairobi. The programme provided 8-week life skills training (40 hours) and technical training in ICT (150 hours) along with 8-week work experience through internships, and 6-month job placement support. De Azevedo et al. (2013). The evaluation used mixed methods, including an RCT with 3 arms. Participants were randomly assigned to one of 3 groups: those who received life skills and ICT training, internships and job placement support; those who received the same apart from life skills training; and a control group (no activities). The baseline took place prior to the intervention and follow-up was after completion of the internship. Attrition was high, with 50.2% of those in the baseline not participating at end line. Economic outcomes: Participants were more likely to obtain a job. The first group were 14% more likely to do so than those in the control group (a statistically significant difference). Participants were more likely to hold a full-time position while the control group were more likely to be casual labourers. The proportion of those who worked in the ICT sector was 5 times higher for those in the first group and 3 times higher for those in the second compared with the control group. In terms of quality of employment, more participants in the first group indicated receiving social benefits, including paid leave, compared to the control group. Participants in the first and second groups were also able to increase their weekly income. The 8-week training increased life skills and ICT skills knowledge. Largest and statistically significant gains were seen among the first group compared to the other two. Those in the first group showed the largest gains on life skills related to workplace behaviour and job search. Participants in the first and second groups were more likely to apply for a job and use different search strategies compared to the control group. Psychosocial wellbeing: Participants who were not confident about their skills at baseline improved their confidence at end line compared to the control group. Participants were also more optimistic about finding employment (with those in the first group more optimistic than those in the second). 89 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Neqdar Nesharek (‘We can participate’) Egypt Marginalised adolescent girls and young women aged 16–29 rural areas. The programme brought participants together regularly (around 3 times per week) in safe spaces and provided support by promoters/ local mentors and training on business and vocational skills, life skills, financial literacy, civic engagement awareness, and reproductive health. It also included community mobilisation with families and communities to make women’s work and engagement in the public sphere more acceptable. Ramadan et al. (2014). A quasi-experi- mental impact evaluation comparing girls in 30 intervention and 15 control villages. Programme impact was assessed through a set of strategies comparing end-of-pro- gramme responses to interim-survey responses across beneficiaries (by level of participation) in intervention villages, and non-beneficiaries in both intervention and control villages. IDIs and FGDs were also conducted throughout the programme with promoters, participants, dropouts and guardians/spouses. Economic outcomes: More than 1,290 out of the 4,500 targeted girls and young women were able to start a business in their own village, and more than 640 participants found employment. Most participants chose to start a business due to flexible hours and location, while unmarried girls with fewer family responsibilities were more willing to find higher-paying jobs in the city. Participants also opened more than 3,830 savings accounts for their earnings – for those younger than 18 years a custodian opened the account for them. Yet many were unable to start a business due to high start-up costs, lack of opportunities to get a low-inte- rest loan or the stigma of taking a loan. Participants also had difficulty to sell their products and services outside their villages. Psychosocial wellbeing: Participants reported having undergone a transformational experience. They spoke about increased self-confidence, better understanding of oneself, ability to set goals and strategise on how to reach them, and finding meaning in life. Voice and agency: Participants increased their mobility and improved their social skills. They also learnt to value and manage their time, effectively combining household and work responsibilities. New skills, income-generation and providing for their families empowered participants. The programme also helped them get a national identification card, necessary to access social services, open a bank account and vote. 90 Title and country Target group Intervention activities Study Evaluation methodology Key findings Por un Mañana (‘For a future’) Argentina Youth aged 17–24 from socially and economically vulnerable families in the poorest urban areas, and unemployed, Patagonia. 87% of participants were female, some of indigenous or mixed origin, including young mothers. The programme provided 10 months of Life Project (group activities to foster self-es- teem, life skills and self-reflection on life plan) and Social Labour workshops (labour laws, occupational safety and health), job training, and internship by partner institutions. Tutors accompanied participants in all activities. Hamilton et al. (2013). 9-month action research with 3 cohorts of graduates, aiming to reassess the programme and improve employment prospects of graduates. Observation and interviews with recent graduates, including young mothers, and staff. Problems: Graduates reported that the intervention helped them break their social isolation, build their social capital, clarify their life goals and plan for the future. Yet many were unable to act on their life plans, did not search to find employment after programme completion, and asked for continuing assistance with job searches. Staff questioned graduates’ readiness to undertake such searches, ability to change their routine and take action, and noted their difficulty to implement what they had learnt. The study also found that graduates also lacked the network to link them to employers, who discriminated on the basis of applicants’ appearance and residence. The programme enabled young mothers to leave the domestic sphere and enter new spaces where they could imagine alternative futures as working women and mothers. Yet they also found difficulty to translate their life plans into action, with their childcare responsibilities being a key barrier. The study asked staff to redesign the life skills training, create a social support network for mothers, assist graduates with job searches and placements and address employers’ attitudes. Programa Joven (‘Youth programme’) Argentina Young women and men aged at least 16 years, from poor households, with low education levels, no working experience, who were unemployed, underemployed or inactive. The programme provided vocational skills training and internships. Duration varied between 14 and 20 weeks and included 6 to 12 weeks of technical training, followed by 8-week internships. It also covered transportation expenses, provided a stipend for women with young children, and offered medical checkups, materials and work clothing. Aedo and Nunez (2004). The evaluation used propensity score matching with data from those who registered and qualified for the programme, data from a sample of beneficiaries and a control group, and the same sample of beneficiaries with information collected in a survey 1 year after programme completion. Economic outcomes: Participants were able to increase their employment. Yet programme impact was statistically significant for older women (aged 21–35). Participants also increased their earnings, yet outcomes on earnings were statistically significant for young men and adult women; they were insignificant for younger women (below 21 years). The study attributes these differences not to programme components, but to the realities of the local labour market. Having children was positively and significantly correlated to programme participation for women. The variables related to unemployed with working experience and school attendance were significant and negatively related to programme participation for young women. 91 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings ProJoven (Youth Labour Training Programme) Peru Youth (girls and boys) aged 16–24 from poor families, with low levels of formal education, no or minimum labour market experience, underemployed, unemployed or inactive, in urban areas. ProJoven provided 3-month training and paid internships at private firms for a period not shorter than 3 months. During the internship, trainees received a small stipend lower than the minimum wage, covering transportation, meals and medical insurance. Women with children under 5 of age received a double stipend. Participants also received health insurance. Diaz and Jaramillo (2006). Quasi-experi- mental evaluation with programme participants being the treatment group and eligible non-participants being the control group. A baseline was conducted and 3 follow-up surveys took place 6, 12 and 18 months post-in- tervention. Economic outcomes: Positive and statistically significant effects in terms of increased paid employment, likelihood of formal employment, and monthly earnings. Formal employment increased after programme participation by 7 to 18 percentage points. Effects were higher 6 month after programme participation than after 18 months. Girls and adolescents benefited more as they were more likely to have employment, formal employment and higher monthly earnings than boys and young adults. Programme impacts on real monthly earnings initially declined but then rebounded and grew in the more recent follow-up, presenting a U-shape. Effects on hourly earnings were also positive but less clear. Nopo et al. (2007). The study used data from the baseline at the beginning of the internships, and 3 follow-ups at 6, 12 and 18 months after internship completion. Participants were the treatment group. Propensity score matching was used to identify those who became the control group. Economic outcomes: Positive programme impact was found only after 18 months of programme completion. Outcomes were better for female participants. The study estimated that women improved their employment rates by nearly 6% and 15%, 12 and 18 months respectively after programme completion. Overall, ProJoven did not substantially increase the employment rate among participants. But it increased their weekly hours worked compared to the baseline. In terms of gender differentials, men worked more hours than women. ProJoven also increased weekly earnings initially more for women. It also increased monthly earnings, but more for women, who increased their earnings by nearly 93% (men only by 11%) after 18 months. Gender occupational segregation among participants was also reduced by 30% as women were trained and worked on occupations with a high prevalence of male workers. 92 Title and country Target group Intervention activities Study Evaluation methodology Key findings Diaz and Rosas (2016). The study presents the same experimental impact evaluation, yet using another follow-up survey, and measuring impacts almost 3 years after training, combining data from a follow-up survey and administrative data from Peru’s electronic payroll to measure formal employment. Economic outcomes: The study found no statistically significant impact on employment but a positive impact on the quality of employment: formal employment reached 20% according to data from the follow-up and 17% according to the payroll. The follow-up survey indicated that this impact was almost twice as much for male youth and for those aged 14–18. On the other hand, the payroll data showed statistically significant change in formal employment, which reached 26% for women and 35% for those over 18. The latter was explained by the fact that formal firms preferred not to hire or register under-aged youth. Psychosocial wellbeing: No significant change in socio-emotional skills, such as motivation and self-esteem. SATYA-Pra- tham programme India Young women aged 18–39, with at least 5 grades of schooling, in low-income households in two slums, New Delhi. The programme provided 6-month subsidised vocational training in stitching and tailoring. Participants had to attend 2 hours daily for 5 days. Maitra and Mani (2014). Randomised field experiment using a treatment and control group and two rounds of post-intervention data collected 6 months and 18 months later. In the sample, 56% of women assigned to the treatment group completed the programme. Economic outcomes: 6 months post-in- tervention, participants were 6 percentage points more likely to be employed, 4 points more likely to be self-employed, to work 2.5 more hours weekly, and earn 150% more monthly than the control group. These outcomes were sustained according to the second follow-up 18 months post-intervention. Participants were also 13 percentage points more likely to own a sewing machine in the second follow-up. Psychosocial wellbeing, voice and agency: No effect on happiness at home or agency. Problems: Barriers to programme uptake and completion included credit constraints, distance to training facility and lack of childcare support. A 10-mi- nute increase in walking time to the training centre was associated with a 14 percentage point reduction in the likelihood of programme completion. Married women with a mother-in-law in the household were 29 percentage points more likely to complete training – with the mother-in-law providing childcare support. Those who had already completed secondary schooling were 25% more likely to complete the training. 93 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings SEWA- li- velihoods skill building intervention for adolescent girls India Girls aged 13–19, married and unmarried, rural Gujarat. This programme provided girls’ groups (kishori mandals), life skills and vocational training, financial literacy and opportunities to visit places such as banks, universities and places of historic interest to broaden girls’ horizons and experiences beyond their villages. Girls were also offered basic training in forestry, agriculture and animal husbandry, and water management, and were encouraged to attend training in traditional and non-traditional trades in institutions outside the village in order to expand their livelihoods options. SEWA et al. (2006). Mixed methods with quasi-experimental design. Economic outcomes: Girls appreciated learning vocational skills and wished to have more on livelihoods options. Psychosocial wellbeing: Girls reported increased self-esteem. Larger increases were reported among regular participants compared to girls in the control group and those with irregular attendance. Voice and agency: Girls increased their mobility, but did not significantly increase their freedom of movement, which remained supervised. They also reported increased participation in decision-ma- king, and adopting more gender-ega- litarian attitudes. Larger increases in decision-making participation, and the adoption of more gender-egalitarian attitudes were found among regular participants compared to girls in both the control group and those with irregular attendance. Despite the measured positive change, overall girls’ agency continued to be limited. 94 Title and country Target group Intervention activities Study Evaluation methodology Key findings Technical Education and Vocational Education and Training Authority (TEVETA) programme Malawi Vulnerable youth (mostly defined as orphans and school dropouts), girls and boys, aged 15–24. The programme provided 3-mon- th on-the-job vocational and entrepreneurial skill training through placements as apprentices to master craftspeople in the area of interest of the participant. A small stipend to cover meals and accommodation was also offered. Cho et al. (2015). RCT with 1,900 youth randomly assigned to 2 cohorts to receive training; the treatment group started immediately, and the control group around 4 months later, around the time of the follow-up survey. About a third of respondents in the baseline survey could not be located at follow-up, especially girls. Dropout rates were high due to administrative errors and because some trainees chose not to attend or complete training. The study tracked down a large number of programme dropouts and examined the determinants and consequences of their dropping out. The study also included a qualitative survey with implementation officers. Economic outcomes: No improvements in labour market outcomes (hours worked, total earnings, total monthly expenditure) were identified in the short run. Yet boys were more likely to attend regularly and 3 times more likely to receive a paid job offer after the training period. Negative outcomes such as reduced personal savings used to participate in training and decreased earnings-related activities were found among girls. Girls were also less likely to have started a business. Education: Male graduates spent around 3–6 hours weekly in additional skills development activities. Health: Participating girls were significantly less likely to have given birth in the past year compared to the control group. Psychosocial wellbeing: Increased happiness and life satisfaction. Improvements in life satisfaction were greater for boys. Voice and agency: Increased self-confi- dence in ability to change economic activities, and move from agriculture into other income-generating activities. While gains in self-reported skills were similar between girls and boys, improvements in confidence to get non-farm employment measures were more marked for boys. Problems: The study attributes gender differences in programme participation and outcomes to the conditions under which women participated in training and gender differences in the training experience. Girls were more likely to participate when alternative opportunities disappeared, and to drop out more often due to adverse shocks or distance to the training facility. Girls were also less educated at baseline, had more dependents and spent more time on domestic chores. They reported family obligations and marriage as the key reasons for dropping out. Participation was also more expensive for them as the stipend was insufficient to cover costs, and girls had to draw down their savings to do that, while boys were more likely to be financially supported by trainers. Thus location, accessibility and convenience of training sessions, and household responsibilities appear to be important determinants of female attendance. 95 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Technical and Vocational Vouchers Programme Kenya Youth aged 17–28, out of school, in urban and rural areas. These youth were part of a longitudinal study known as the Kenyan Life Panel Survey (KLPS). The programme provided vouchers for vocational training. Vouchers covered tuition, materials, uniforms and trade test fees, but not accommodation, meals and transportation costs. Hicks et al. (2015). The programme randomly selected half of 2,163 eligible youth (63% were women) to get the voucher, while the other half served as the control group. Of those who received the voucher, a random half were awarded a voucher for a public institution and the other half for either a public or private institution. Finally, a randomly selected half of all treatment and control individuals were exposed to information about the actual returns to vocational education along with earnings differences by trades, especially traditionally male-dominated and female-domi- nated trades. Economic outcomes: Those who received vouchers were substantially more likely to enrol in vocational training and were able to acquire an additional 0.55 years of education compared to the control group. The programme significantly increased hourly wage earnings among those earning a wage, yet it did not have an overall statistically significant effect on labour earnings. Nor did it increase work hours. Some evidence indicates that it increased the likelihood of a shift out of agriculture and into wage employment among those who had been out of school longer. Trainees also stressed the lack of start-up capital to use their newly acquired skills and engage in self-employment. Women were much less likely to attend training every term compared to men (by 9.5 percentage points). Evidence suggests that those who completed secondary school were more likely to complete their training course. Overall, female and male trainees conformed to traditional gendered trades, with men choosing mechanics or driving, and women tailoring or hairdressing. Yet the information intervention encouraged female youth to choose and enrol in traditionally male-dominated trades; those exposed to the intervention were more than 5 percentage points more likely to prefer and 5 percentage points more likely to actually enrol in a male-dominated course. The differences between private and public vocational institutions are not yet fully studied, yet those who got the voucher for both types had higher take-up rates. Perhaps this is because private institutions provide more flexible and relevant training to participants’ needs. Non-economic outcomes: No significant improvements in consumption or wellbeing, or age at marriage or fertility. 96 Title and country Target group Intervention activities Study Evaluation methodology Key findings Towards Improved Economic and Sexual/ Reproductive Health Outcomes for Adolescent Girls (TESFA) Ethiopia Married girls, aged 14–19, rural areas. TESFA provided girls’ clubs, life skills training, access to health information and services, financial and livelihoods training and community mobilisation. Edmeades et al. (2014). Quasi-experi- ment comparing 4 groups: those who received economic empowerment education; those who received SRH education; those who received both economic empowerment and SRH education; and a control group. Also qualitative research with girls and their families. Economic outcomes: All groups, including the control group, experienced increases in the proportion of girls participating in paid employment and income-generating activities (increases were more pronounced in the combined intervention group). Girls who saved more had received the economic empowerment training, with some girls forming savings groups; loans increased significantly for the economic empowerment and combined groups, but not in the other two. At end line, more working girls reported saving their earnings for future health expenditures or investment in income-generating activities: 28% from the first group, 20% from the second, 23% from the combined group and just 3% from the control group. Also, girls who had participated in the economic empowerment and combined group (and had thus received financial training) increased the use of savings for productive investments such as small businesses and agricultural supplies. Moreover, girls increased their borrowing – by 45 percentage points in the first group, 10 points in the second, 35 points in the combined group, and 2 points in the control group. Girls also increased their access to and sole control over less valuable property such as chickens, but their control over larger livestock did not change or even reduced – a finding again attributed to greater collaboration with their husbands. Bodily integrity: Some participants mentioned improved communication with their husbands and decreased levels of gender-based violence. Health: Improvements in knowledge of contraceptive methods and STIs, and increase in family planning use, especially in the SRH and combined group. Psychosocial wellbeing: Participants reported improved mental health along with increased social capital and support. Voice and agency: Girls improved communication with their husbands and increased their involvement in household economic decisions. Most girls shared their loans with their husbands, a sign of increased spousal collaboration. 97 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Youth Opportunities Programme (YOP) Uganda Youth, aged 16–35, poor and un/ under-em- ployed, who on average reached 8th grade, in rural areas. YOP invited targeted youth to form groups and submit grant proposals to receive vocational training and business start-up to access tools and materials. A third of applicants were women. Blattman et al. (2014). RCT with the funding randomly assigned to groups. Treatment and control groups were surveyed 2 and 4 years after grant disbursement. Attrition was high after 4 years, higher in the control group. Economic outcomes: After 4 years, beneficiaries were more than twice as likely to practise a skilled trade. They also increased their business assets by 57%, work hours by 17% and earnings by 38% compared to the control group. They were also more likely to keep records, register their business, pay taxes, and hire a part-time employee. While outcomes were similar for male and female participants, women benefited more as they were poorer at baseline. After 4 years, women in the treatment group had 73% higher incomes compared to women in the control group and compared to a 29% gain for men. Education: 2 years after programme completion, 68% of graduates were enrolled in vocational training compared to 15% of the control group, and spent on average 340 more hours of vocational training than the control group. Psychosocial wellbeing: Participants reported increased sense of wellbeing. Voice and agency: No or little programme effect on social cohesion, local and national collective action or violent protest. 104 Title and country Target group Intervention activities Study Evaluation methodology Key findings Empowering Adolescent Girls project Ethiopia In-school and out- of-school adolescent girls aged 10–19. This integrated programme included provision of economic opportunities (facilitation of girls’ access to land, girls’ irrigation groups, savings and credit groups); leadership, voice and rights; education; social opportunity; health and security. CRS (2009). Baseline and midline survey (no control group). Economic outcomes: Girls started getting increasingly involved in agro-en- terprises, selling agricultural products and earning their own income. Nearly 1 in 3 generated their own income and used it to cover schooling costs and improve their nutrition. Girls who accessed a plot of land cultivated it jointly with family members and thus shared their profits with their family. Vulnerable girls who received ruminants or stipends invested in small livestock, got involved in animal breeding and doubled or tripled their profits. The proportion of girls with savings also increased; girls used savings to access small loans for off-farm income-generating activities. Those who did not invest their savings used them at critical times such as during the school registration period, dry season or at times of household hardship. Education: Increased girls’ school enrolment and reduced dropout rates. Girls used the stipends provided to purchase books, rent a room to stay close to their secondary school or meet their basic needs. In addition, fuel-saving stoves greatly reduced the time girls spent on firewood collection and cooking time, and thus enabled them to spend more time studying. The project also constructed water points and latrines, enabling girls to be on time in class and to increase their enrolment and performance during menstruation. 105 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Filles Éveillées (‘Girls awakened’) Burkina Faso Adolescent girls aged 11– 19, migrants working as domestic workers in urban areas. Girls in the first cohort were younger, aged 11–16, and lived in Bobo Dioulasso. The programme offered safe spaces, female mentors, financial literacy, health and life skills training. Engebretsen (2012). Baseline and end line surveys. Economic outcomes: Improved savings behaviour at end line and more girls reporting saving in a bank or in an informal savings group, and having savings goals. Education: Reduced proportion of girls who had never been to school as those interested were supported to enrol in evening classes; increased proportion of girls who could read. Health: Girls improved their overall health and reproductive health knowledge and attitudes towards use of local health services. Almost all girls reported that participation enabled them to manage their health better. Psychosocial wellbeing: Girls reported having more friends as a result of the programme and all reported being able to talk to their mentors for the problems they faced. Fewer girls at end line felt weak and vulnerable. Voice and agency: More girls could decide for themselves and could express ideas in front of others. Participants also had more gender-egalitarian attitudes. Adolescent girls aged 11– 19, migrants working as domestic workers in urban areas. Girls in the second cohort were older, aged 15–19, and lived in Ouagadougou. Engebretsen (2013a). Baseline and end line surveys. Economic outcomes: Girls reported that savings improved significantly between baseline and end line. Most girls continued saving in informal institutions for both short- and longer-term reasons. Education: Girls reported investing their earnings in education and training. Health: Improved SRH awareness, including knowledge of modern family planning methods, and more girls visiting a health centre while fewer girls used self-medication. Over time, girls’ knowledge of positive health behaviours during pregnancy also improved. Psychosocial wellbeing: Girls reported increased social capital, in terms of having friends, a place to meet them and someone to turn to for advice. Girls improved their self-confidence and reduced their sense of vulnerability. Voice and agency: Girls improved their ability to decide for themselves. A statistically significant increase was found in gender-egalitarian attitudes. 106 Title and country Target group Intervention activities Study Evaluation methodology Key findings Adolescent girls aged 11– 19, migrants working as domestic workers in urban areas. Girls in the first cohort were younger, aged 11–16 and lived in Bobo Dioulasso. Engebretsen (2013b). Follow-up survey with a subsample of the girls in the first cohort. Economic outcomes: The proportion of girls saving between end line and follow-up increased slightly. 4 out of 5 girls reported saving money – more than half doing so regularly – for short- and longer-term reasons such as to prepare for their marriage, pay for training or buy supplies for income-generating activities. While knowledge of formal savings mechanisms improved, it was still low. Health: 1 year after programme completion, girls continued to improve their overall health awareness. Yet changes in pregnancy knowledge were inconsistent and awareness of when and how a woman can get pregnant decreased over time. Family planning knowledge also declined and knowledge about STIs and HIV was inconsistent. Psychosocial wellbeing: Girls continued expanding their social network even more between end line and follow-up. Voice and agency: Two-thirds of participants continued to decide alone how to spend their earnings between end line and follow-up. Adolescent girls aged 11– 19, migrants working as domestic workers in urban areas. Jarvis and Kabore (2012). Primary qualitative data and analysis of primary quantitative and qualitative data collected over programme course. Voice and agency: Employers and community members also reported that participants were more respectful and stable, learnt new skills and were willing to share their new knowledge with others. 107 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Ishraq (‘Enlightenment/ Sunrise’) Egypt Girls initially aged 13–15, then 11–15, out-of- school, rural areas. Ishraq provided safe spaces, trained female mentors, literacy classes, home and livelihoods skills, life skills training, sports and community sensitisation along with the introduction of a life skills programme for boys aged 13–17, particularly brothers of female participants, to improve their awareness on gender equality issues. Brady et al. (2007). Baseline and end line survey after the pilot phase ended. The study compared girls from 4 villages divided into 4 categories: non-participants; participants for 1 year or less; those who participated for 13-–29 months; and participants for the full 30 months; plus the control group drawn from 2 control villages. Economic outcomes: Most participants reported learning new practical skills such as making cheeses and butter or sewing and needlework. Those who attended vocational training acquired new skills. Education: 92% of participants passed the government literacy exam and 68.5% who completed the programme entered or re-entered school. Bodily integrity: The proportion of girls preferring to marry early substantially dropped between baseline and end line, especially among girls with longer programme exposure. Experience of violence did not change, yet participants showed a statistically significant decline in attitudes accepting violence and controlling behaviours, with the decline increasing with greater programme exposure. Health: Participants improved their knowledge about nutrition, anaemia, smoking and SRH issues. Psychosocial wellbeing: Participants built friendships. Those with regular participation reported higher levels of self-confidence than non-participants. And 49% of participants were members of a local group or association. Voice and agency: Participants expressed the desire to marry later and have a say in choosing their husband. Those with longer exposure to the programme were more likely to support having a say. Participants also reported desiring smaller family sizes, especially those with greatest programme exposure. Nearly all (99%) objected to FGM/C for their future daughters. 108 Title and country Target group Intervention activities Study Evaluation methodology Key findings Selim et al. (2013). Quasi-experiment. Interviews with participants and non-participants. Economic outcomes: Participants were more confident than non-participants about starting their own business. A few – a higher proportion than non-participants – planned to or had already started their business at end line. Although participants were not more likely than non-participants to save and budget, they were more likely to save money for an emergency and to know someone from whom to borrow money in case of emergency. Education: Participants scored better than non-participants on a number of writing, reading, comprehension and multiplication tests. Psychosocial wellbeing: The youth centre became a safe space for girls and gave them the opportunity to expand their social networks. Voice and agency: Participants improved their gender-equitable attitudes as more agreed that girls should be allowed to play sports like boys, that the appropriate age for marriage was 18 or older, and that they should be involved in decisions about whom to marry. They also expressed the desire to have fewer children and their intention not to perform FGM/C on their daughters. Sieverding and Elbadawy (2016). Quasi-experiment with participants and girls who were eligible but did not join the programme. Economic outcomes: Girls were assisted to open savings accounts at the local post office. Education: Participants increased their literacy and numeracy compared to non-participants, and many girls were able to re-enter formal education. Health: Girls improved their reproductive health knowledge, but were not much more likely than non-participants to seek professional health care when sick. Voice and agency: Some gender attitudes improved, such as those towards girls playing sports, marrying at age 18 or older and getting involved in decisions over whom to marry, having fewer children and intending not to perform FGM/C on their daughters. Yet no effect was found on girls’ attitudes towards beating a girl. Overall, the programme did not have an effect on girls’ empowerment as measured by self-esteem, attitudes towards violence against girls and mobility, and had only a marginally significant effect on their decision-making. This was linked to the fact that the programme did not sufficiently work with family and community members – the ‘gatekeepers’ that shape girls’ lives. 109 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Kishori Abhijan (‘Adolescent girls’ adventure’) Bangladesh Girls aged 13–22, rural areas. The programme provided safe spaces, life skills training (self-es- teem, leadership skills, health, gender roles, legal rights), and in addition, to older girls aged 15–18, livelihoods training (poultry care, handicrafts, sewing, teacher training). Amin and Suran (2005). Longitudinal qua- si-experimental approach combined with qualitative methodology. Economic outcomes: Participants increased their economic participation, but school enrolment and living in better-off families significantly reduced the likelihood of working for pay. Health: Percentage of respondents with knowledge of STIs increased from 23% in 2001 to 45% in 2003, with participants exhibiting a greater increase. Amin (2011). Quasi-experiment and IDIs. Economic outcomes: Increases in the number of girls working for cash (self-em- ployment and part-time work) and in earnings (outcomes more pronounced when programme included microcredit). Education: Participation in paid work did not lead to school dropout but working girls reported spending fewer hours on school work outside the classroom. Bodily integrity: Younger (aged 12–14) and poorer girls living in the poorest district delayed their marriage. The fact that the intervention did not have a dowry component limited its effectiveness, because the longer the respondents waited to marry, the higher the dowry their families had to pay. Health: Improved knowledge about HIV transmission and about aspects of female reproductive biology. Psychosocial wellbeing: Positive effects, with girls improving their sense of self-worth and social connectedness. Voice and agency: Girls improved their mobility. 110 Title and country Target group Intervention activities Study Evaluation methodology Key findings Kishoree Konta (‘Adolescent girls’ voices’) Bangladesh Adolescent girls, aged 10–19, in rural south central region. The programme provided girls with 4 main interventions: basic empowerment package (community mobilisation, safe spaces, mentors, basic education and life skills training); basic package plus financial training; oil incentive for girls aged 15–17 to delay marriage (4 litres of cooking oil collected by girls); and the full package (all 3 components combined). Later on some villages were randomly selected from the ‘basic package plus financial literacy’ and ‘full package’ groups to establish girls’ savings groups. Field et al. (2016). RCT with 460 communities randomly allocated to each of the 4 intervention groups and a control group. Qualitative methods were used to complement some findings. Education: The oil incentive had a positive impact on education. Girls in the incentive-only group were 23% more likely to be in school at midline and 18% more likely to do so at end line. Girls in the incentive plus empowerment group were 25% and 19% more likely respectively. Bodily integrity: The oil incentive reduced the likelihood of early marriage. In particular, the likelihood to get married before age 18 fell by 21% in the incentive-only group and by 14% in the incentive plus empowerment group. The likelihood of marrying under age 16 fell by 16% and by 23% in each of these two groups respectively. Health: Girls in the incentive-only group were 5% less likely to have given birth at end line and 6% less likely to have done so before age 18. Girls in the incentive plus empowerment group were 4% and 19% less likely respectively. Voice and agency: At midline, 73% of girls who received the full package said that they would negotiate with their parents if they were to receive a marriage proposal compared to 69% of girls who received the oil incentive, 67% of girls who only received basic empowerment training and 67% of the control group. Girls who received the incentive were also 12.5 percentage points more likely to attend community events outside their village alone, indicating increased mobility. 111 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Meseret Hiwott (‘Base of life’) Ethiopia Married adolescent girls aged 10–24, rural areas. The programme provided girls with safe spaces and female mentors, life skills training with a focus on SRH and gender and power dynamics, and financial literacy. Participants were also provided with a youth savings scheme to promote their financial inclusion. An additional programme, Addis Birhan (‘New light’) was designed to contribute to the objectives of Meseret Hiwott by providing rural husbands of all ages with communication and support skills to improve the wellbeing of their wives and families. Erulkar and Tamrat (2014). Quasi-experimental design, and analysis of cross-sectional data from population-based surveys among married girls. Bodily integrity: Participating girls whose husbands also joined the parallel programme for men reported lower levels of both sexual and physical violence compared to non-participants and to girls whose partners did not participate in the men’s programme. Health: Girls in the programme were more likely to report increased use of family planning and voluntary counselling and testing for HIV. Reported increases were higher among girls whose husbands also participated in the men’s intervention, and lower among non-participating girls. Voice and agency: Participants were more likely to report increased spousal help with domestic work and accompaniment to the clinic compared to non-parti- cipants. When both spouses participated in the two related interventions, increases were even higher. On the other hand, girls who did not participate in the programme were also more likely to need permission to leave the house than girls who did. 112 Title and country Target group Intervention activities Study Evaluation methodology Key findings Out-of- School Adolescents Support Project Zimbabwe Adolescents aged 10–19, out of school (or at risk of being out of school); 60% of participants were girls. The project provided vocational training (to girls and boys 16–19), including 6 months of classroom training (along with life skills training and recreational activities for psychosocial development), a 6-month apprenticeship, and savings and internal lending communities (SILCs). Communities were provided with gender-awar- eness training. Miller et al. (2011). Qualitative assessment with girls and boys participating in different sections of the programme. Economic outcomes: Apprenticeships were critical to enhance employment prospects as they often led directly to employment after graduation. Training helped girls earn an income and girls were less likely to engage in transactional sex. SILCs provided safe spaces for adolescents to save money and access loans. Education: SILCs enabled girls to access loans and use the earnings of their small business to pay school fees and stay in (or return to) school. Teachers supported the programme and ensured that participation in SILCs did not interfere with learning. Psychosocial wellbeing: Girls improved their self-image and self-esteem and increased their social capital. Voice and agency: Girls felt empowered to make informed decisions about their life. The programme created safe spaces and provided positive role models (mentors). Community gender-awareness training created a supportive environment for girls, including reducing their household responsibilities during training. Problems: Skills offered in the vocational training programme were based more on the supply of trainers rather than market demands. Despite job counselling, participants often chose traditionally female trades. Lack of start-up capital to begin self-employment and purchase the necessary equipment did not allow most graduates to engage in self-employment, which they preferred. The programme also lacked business skills training to help graduates to start and run their own business. Powering Up Biruh Tesfa Ethiopia Girls aged 7–18 years, out of school, living in the poorest and most marginalised urban areas of Addis Ababa. Many girls were migrants and domestic workers. The programme provided girls with safe spaces and female mentors, life skills and HIV training, financial literacy and a voucher for health services. Erulkar and Medhin (2014). The quasi-experi- mental evaluation compared data from a baseline conducted prior to the establishment of girls’ groups with data from the end line, 6 months after the programme started. Education: Participants significantly increased their participation in non-for- mal education compared to the control group, while those girls who had never been in school improved their literacy and numeracy scores. Health: Girls were significantly more likely to have visited a health facility compared to the control group. 113 Interventions promoting adolescent girls’ economic capabilities: what works? Title and country Target group Intervention activities Study Evaluation methodology Key findings Rajiv Gandhi Scheme for Empowerment of Adolescent Girls-Sabla (RGSEAG-Sa- bla) India Adolescent girls aged 11– 18, in school and out of school. RGSEAG-Sabla is a government programme to empower adolescent girls across the country with takehome rations or hot cooked meals, peer groups, health checkups, adolescent reproductive health counselling, life skills training, exposure visits and vocational training. ASCI (2013). Mixed methods: quantitative survey, semi-structured and unstructured qualitative interviews, and FGDs. Problems: The vocational training component faced great challenges. Less than half of eligible girls aged 15–18 received information about vocational training, and only half of them attended training. Distance and gender norms were identified as part of the problem, with parents not allowing their daughters to attend training. Another problem was low and irregular attendance, as poor girls already working were not provided with a cash incentive to compensate for their lost earnings when attending training. Many girls also lacked the credit to purchase the required material to start their own business. Those who did receive training reported being unable to find relevant employment opportunities, and others emphasised their inability to assert themselves enough to get paid for the work they did. SHAZ! (Shaping the Health of Adolescents in Zimbabwe) Zimbabwe Adolescent girls (orphans) aged 16–19, urban areas. The SHAZ! pilot aimed to reduce HIV transmission and SRH risks through 4 activities: provision of adequate and consistent HIV and SRH services; life skills-based HIV education; vocational training and a micro-grant; and integrated social support. Dunbar et al. (2014). RCT: full intervention (treatment group) compared to life skills and reproductive health service only (control group). Economic outcomes: Participants were more likely to generate their own income compared to the control group. Bodily integrity: Greater reduction in the experience of violence over time than the control group. Health: Greater decrease in food insecurity. Reduced risk of transactional sex; higher likelihood to use condoms. No difference in HIV and HSV-2 (herpes simplex virus type 2) incidence. Psychosocial wellbeing: Participants expanded their network of friends and support. Study Target group Study focus Review type Study findings World Bank (2015). Adolescent girls and young women in Afghanistan, Haiti, Jordan, Lao PDR, Liberia, Nepal, Rwanda and South Sudan (age range varies depending on country, some programmes also targeted young women and men). Adolescent Girl Initiative, a multi-country pilot programme providing technical, business and life skill training. Programmes had country-specific characteristics. Review of 5 programmes evaluated through RCTs or quasi-ex- periments. Positive economic and non-economic outcomes for participants, including increased access to employment, income and assets, increased self-confidence, mobility and autonomy, less stress and increased aspirations. WWB (2014b). Low-income youth, including adolescent girls, in developing countries. Women’s World Banking savings programmes, including youth savings accounts and financial education. Programme review. Provision of convenient, affordable and safe savings products along with relevant education is crucial to build girls’ financial capability and equip them with the resources and the knowledge to enter adulthood. Such interventions enable girls to develop skills in asset accumulation, risk management and goal setting. Interventions promoting adolescent girls’ economic capabilities: what works? About GAGE Gender and Adolescence: Global Evidence (GAGE) is a nine-year longitudinal research programme generating evidence on what works to transform the lives of adolescent girls in the Global South. Visit www.gage. odi.org.uk for more information. GAGE Programme Office Overseas Development Institute 203 Blackfriars Road London SE1 8NJ United Kingdom Email: [email protected]g.uk Web: www.gage.odi.org Disclaimer This document is an output of the Gender and Adolescence: Global Evidence (GAGE) programme which is funded by UK aid from the UK government. However, views expressed and information contained within do not necessarily reflect the UK government’s official policies and are not endorsed by the UK government, which accepts no responsibility for such views or information or for any reliance placed on them. Copyright Readers are encouraged to quote and reproduce material from this report for their own non-commercial publications (any commercial use must be cleared with the GAGE Programme Office first by contacting gage@odi. org.uk). As copyright holder, GAGE requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the GAGE website, www.gage.odi.org © GAGE 2018. This work is licensed under a Creative Commons Attribution – NonCommercial-ShareAlike 4.0 International Licence (CC BY-NC-SA 4.0). Front cover: © Graham Crouch/World BankISBN: 978-1-912942-00-8