For instance, early marriage may be considered a “better” option when the alternative of remaining unmarried or delaying marriage fails to protect adolescents from risks to well-being such as early pregnancy, increased likelihood of raising children outside of marriage, and amplified exposure to sexually transmitted infections (Schaffnit et al., 2021).

In the context of constrained choices, women’s agency may contribute to outcomes that contradict what outsiders would define as beneficial in the realm of marital stability (Baraka et al., 2022; Reniers, 2003). For example, in northern Burkina Faso, Guirkinger et al. (2021) found that as women were increasingly involved in the selection of their spouses, a higher number of divorces were initiated by women, many women had a positive perception of polygamy, and remarriage was on the rise and showed no negative associations with women’s well-being. Therefore, they concluded that women showed self-efficacy and internal locus of control in accepting an early marriage, in that they anticipated the ability to divorce and later remarry if they deemed their first union unsatisfactory or unsuccessful. Lastly, focusing on marital satisfaction (rather than stability), and attempting to move scholarship in a more causal direction, Li (2023) used an instrumental-variable approach based on genealogical data and found that higher intrahousehold bargaining power was associated with lower marital satisfaction for women, particularly among older women constrained by external norms.

A related line of inquiry explores the indirect role of agency in marital outcomes by studying outcomes across varying couple configurations, such as couples in which both partners hold the same versus different levels of education or couples in which women’s relative earnings are higher than men’s. One well-established finding, at least in low- and middle-income societies, is that couples that did not share “status consistency” (heterogamous couples) tended to exhibit higher rates of conflict and marital instability, including more common instances of intimate partner violence (IPV; Behrman, 2019; Pesando, 2022; Weitzman, 2014; Yount, 2005; Yount & Carrera, 2006). In contexts in which new and atypical roles may threaten male dominance (Cools & Kotsadam, 2017; Hornung et al., 1981), when women had higher education, better occupations, or more resources than men, increased conflict between partners could result, including less gender-equal decision making, less female autonomy, and more violence (Ackerson et al., 2008; Flake, 2005; Rocca et al., 2009). Research across 28 countries on women’s asset ownership and experiences of IPV suggested that the relationship of women’s sole (status inconsistent) and joint ownership of a house and/or land with IPV was highly context-specific (Peterman et al., 2017). Much of this literature, however, adopted proxies for agency that may be seen as developmental outcomes per se, namely female educational

attainment, earnings, and occupation. Research using actual measures of agency, including interpersonal agency at the couple level, is lacking.

In contrast to observations on forms of status inconsistency related to resources such as education or occupation, age inconsistency was shown to favor men—age gaps between spouses (e.g., older husbands) were associated with earlier onset of sexual intercourse, less agreement on sexual practices, and less overall use of contraception (Barbieri & Hertrich, 2005; Kaestle et al., 2002). So, increases in income/education for women were shown to create higher vulnerability, while increases in women’s age (e.g., reducing the age gap favoring men) were associated with lower coercion. Research also documented that in Indian communities in which social norms dictate greater separation of men and women and in which women have low decision-making power, girls tended to marry at an earlier age (Desai & Andrist, 2010).

Moving to fertility, several analyses of Demographic and Health Surveys (DHS) data from low- and middle-income countries (LMICs) examined the association between women’s empowerment and fertility preferences. Haque et al. (2021) analyzed large national surveys for women older than 35 years from 53 LMICs, focusing on the association between women’s empowerment (measured by participation in household decision making and attitude toward wife-beating) and fertility preferences (measured by women’s perceived ideal number of children and their ability to achieve that preference). The study found that empowered women had a relatively low ideal number of children, irrespective of the measures used to assess women’s empowerment (Haque et al., 2021). Atake and Ali (2019) also used DHS data from four countries in West Africa and found that women who were more empowered desired significantly fewer children compared with their less-empowered counterparts, although results varied by country.

On the relationship between women’s empowerment and the number of children born, in a systematic literature review, Upadhyay et al. (2014) documented that most studies found an inverse relationship between number of children and at least one empowerment measure. The review reported that most studies including at least some aspects of women’s empowerment were linked to increased spousal communication and the capacity to decide whether to have children (Upadhyay et al., 2014). Moreover, the analysis included two studies using DHS data from the Philippines and Bangladesh, which documented a significant association between empowerment measures and unwanted pregnancy. Women’s final say in household decisions and sexual matters with husbands was found to be associated with lower odds of unwanted pregnancy (Abada & Tenkorang, 2012; Rahman, 2012). Albeit associational, these findings suggest higher sexual and reproductive autonomy among women with higher levels of agency. Schierl et al.’s (2023) pooled analysis of DHS data from 31 Sub-Saharan African

countries corroborated this conclusion: women’s empowerment was found to be strongly associated with the ability to make decisions about sexual behavior, including highly empowered women refusing sex or asking their partners to use condoms, which increased women’s control over the risk of acquiring HIV (Schierl et al., 2023).

A recent study from Mozambique on the influences of women’s empowerment on fertility-related outcomes also showed that women’s ability to make independent decisions increased the odds of not wanting more children. Furthermore, beliefs about violence against women and control over resources were positively associated with having fewer children (Lopes et al., 2022). While almost all the evidence reviewed came from cross-sectional studies, one longitudinal study from Egypt showed that women with greater agency were more likely to have had a birth and to have had a greater total number of births than women with less agency. Due to social norms in Egypt favoring a higher number of births, women with higher agency could be fulfilling these social expectations and choosing to have more children (Samari, 2017).

Part of the challenge in evaluating the role of women’s interpersonal agency in shaping fertility arises from the fact that, in many instances, spouses have similar fertility preferences (Mason & Taj, 1987). However, when spousal preferences do not coincide, the relative weight of partners’ preferences defies generalization, even within the same setting (Blanc, 2001). In contrast to studies focusing on individual-level measures of discordance between spouses, Mason and Smith (2000), looking across LMICs with differing levels of gender inequality in intrahousehold power and across diverse gender contexts within the same country, compared outcomes when spouses disagree about desire for more children. They found that in countries where women have greater autonomy, the decision to use contraception was more likely to be affected by the wife’s preferences than it was in countries where gender inequality was greater. This finding was observed for India, Pakistan, Thailand, and Malaysia but not for the Philippines.

A similar importance of social context was observed when interventions were implemented. Many interventions aim at influencing family formation and fertility outcomes by boosting various aspects of women’s empowerment. Substantial differences exist, however, across contexts in terms of these issues (J-PAL Policy Bulletin, 2018, Figure 1). The timing of first union formation or marriage is heavily influenced by social norms. In some regions, the practice of dowry payments by the bride’s family to the groom’s family fuels early marriages, as dowry tends to increase with the bride’s age (Chiplunkar & Weaver, 2023; Corno & Voena, 2023). This practice can lead brides’ parents to favor early marriages, to reduce dowry costs. Even in the absence of dowry practices, women from disadvantaged households may marry early due to reduced opportunities, such as limited

access to education or employment (Batyra & Pesando, 2023; Corno et al., 2020; Kohno et al., 2020). Household income not only affects marriage decisions but also influences decisions related to sexual activity. In contexts in which sex outside of marriage is not against the norm, adolescents in low-income households may engage in relationships with older partners who can support them financially (Luke, 2003). Additionally, income plays a role in adolescents’ ability to develop human capital, including education and skills, which can empower them to make informed decisions about sexual activity, marriage, and childbearing and to pursue economic opportunities.

Because family formation and fertility outcomes often co-move (e.g., early marriage is closely tied to decreased sexual and reproductive health and earlier fertility), reviewing the causal evidence by outcome is challenging and risks redundancy. Instead, the text below is organized by type of program investment, distinguishing between economic programs (e.g., financial incentives and/or schooling interventions), social norms interventions (often called “empowerment” programs), life skills training programs, political initiatives, or combinations of these investments. Chapter 5 includes a review focused on sexual and reproductive health, and thus that topic is not a focus here.

Impacts of Women’s Empowerment Interventions on Union Formation and Fertility Outcomes

Cash and Asset Transfers

Much of the literature on cash transfers and their potential to influence union and fertility outcomes has targeted adolescent girls, to support their retention in school and delayed marriage and childbearing. Baird et al. (2011) assessed the effectiveness of randomized conditional cash transfer programs (CCTs) and unconditional cash transfer programs (UCTs) on adolescent girls’ life trajectories in Malawi. When focusing on adolescent marriage and pregnancy rates, results were particularly strong in the UCT arm. The likelihood of ever being pregnant or married were 27% and 44% lower, respectively, in the UCT arm than in the control group at the end of the two-year intervention. These substantial delays in marriage and fertility in the UCT arm were seen entirely among adolescent girls who dropped out of school after the start of the two-year intervention; rates of marriage and fertility among girls still enrolled in school at follow-up were negligible, regardless of the treatment arm. Additional positive effects of the cash transfers included lower risk for stunting and HIV, as well as delayed marriage and age at first birth (Baird et al., 2011, 2016, 2019). Importantly, effects were lost after the cash transfers ended, and there is some indication that delayed marriage may have affected the quality of available partners

for marriage in adulthood. Hence, while the findings from this work, as well as reviews of cash transfer studies (Baird et al., 2013; Kalamar et al., 2016), document the value of conditional and unconditional cash transfers, unanticipated negative consequences of nonadherence to norms related to age at marriage can result from such interventions. These findings highlight the importance of providing broader community- and society-level interventions focused on norms when engaging in economic empowerment interventions targeting women and girls.

In contexts without dowries and in which women had agency over their own marriages, money transfers conditional on staying in school were associated with delayed marriage or childbearing (Alam et al., 2021; Austrian et al., 2022). The primarily nonexperimental evidence from contexts with dowry and arranged marriage is mixed. Alam et al. (2021), in Pakistan, and Hahn et al. (2018), in Bangladesh, found that transfers conditional on education and marriage were associated with reduced child marriage. However, Heath and Mobarak (2015) documented no impact of the same Bangladeshi program using a different comparison group. The only policy effort to condition financial incentives on marriage age alone was the Apni Beti Apna Dhan (ABAD) program in Haryana, India. Despite the ABAD CCT’s intent to enhance the value of girls in Haryana, the evaluation showed that ABAD did not significantly delay marriages or encourage secondary and higher levels of education, suggesting that financial incentives alone may not change deeply entrenched and gendered social norms (Nanda et al., 2016).

Targeting child marriage, specifically by lessening families’ economic pressure to marry girls early through a combination of economic and social norm interventions, Buchmann et al. (2021) evaluated a randomized program in Bangladesh providing financial incentive to families to delay marriage, alongside a girls’ empowerment program. They found that girls eligible for two years of incentive were 19% less likely to marry below the age of 18 (the age of consent), while the empowerment program failed to reduce adolescent marriage. As such, the findings suggest that empowerment programs may be ineffective at reducing underage marriage when adolescent brides have limited influence on marriage timing, underscoring the key role of young girls’ agency and the influence of an enabling normative environment. This same finding was also reflected in several additional studies (Ashraf et al., 2020; Bandiera et al., 2020; Buehren et al., 2017; Edmonds et al., 2023).

In Bihar, India, the state-run Mukhhyamantri Balika Cycle Yojana (Chief Minister’s Scheme on Cycle for Girls), designed to close the education gap among girls in the state, provided girls with funds to buy bicycles. This is an example of a “successful” asset-transfer program. A randomized impact evaluation found that the bicycle program improved school enrollment by 32% and reduced the corresponding gender gap by 40%

(Muralidharan & Prakash, 2017). Furthermore, the authors found a 22% decline in sexual harassment on the way to school and documented that the program led girls to express an increased desire to delay marriage and pregnancy. Now piloted in Zambia, a similar randomized evaluation of a bicycle program found that the program had significant effects on enrollment, graduation, absenteeism, punctuality, and empowerment (Fiala et al., 2022).

Education Subsidies

Similarly, several types of educational subsidies have been found to support school retention, delayed marriage, and delayed first births for adolescent girls, though relatively less research on this type of intervention exists compared to research on cash transfers. Duflo et al. (2015) randomized the provision of free school uniforms to students in their last three years of elementary school in Kenya and observed a reduction in adolescent pregnancy among women beneficiaries. In Zimbabwe, using a randomized controlled trial (RCT) design, Hallfors et al. (2011) found that a comprehensive support program for orphan girls, including uniforms and a school-based helper/tutor to solve problems, combined with food programs, increased the likelihood of participants reporting over time that the consequences of sex were an important factor in their desire not to have sexual intercourse. However, the authors did not find self-reported sexual intercourse to be a reliable measure among girls who reported ever having had sexual intercourse at any one of the survey time points. Duflo et al. (2024) randomized secondary school scholarships in Ghana and found large impacts on childbearing onset, age at first cohabitation, partner characteristics, and incidence of unwanted pregnancy. Using a quasi-experimental approach in five Sub-Saharan African countries, Bhuwania et al. (2023) found that tuition-free secondary education reduced the probability of marriage and childbearing before 15 and 18 years of age, respectively, by providing girls with more autonomy and independence in decision making.

Quasi-experimental research in this area includes a study conducted in the context of changes that eliminated primary school fees in Uganda (Keats, 2018). The study found that women with more education tended to delay fertility and had lower fertility overall. A higher level of maternal education was also correlated with improved child health indicators. A similar study examined the role of education on fertility, in the context of the introduction of universal primary education in Nigeria (Osili & Long, 2008). The study found that each additional year of education reduced fertility before the age of 25 by 0.26 births. Ozier (2018) linked eighth grade standardized test score data to responses from a survey of young adults in Kenya and found a reduction in teen pregnancy among women with higher test scores, believed to be associated with admission to government

secondary school. A recent study that leveraged the expansion of access to girls’ or mixed schools in Jordan, to investigate the effects of access to school on women’s education and fertility, found that an additional year of schooling reduced total fertility by 0.3–0.4 births (Zhang & Assaad, 2024).

Knowledge and Skills Training

When social norms allow girls to have some control over their sexual relationships, access to health information may change their beliefs about sexual behaviors and marriage, and building skills to support options beyond marriage and childbearing may, in turn, influence marriage and childbearing outcomes (Chang et al., 2020). Among women regardless of age, contraceptive knowledge and access was shown to contribute to improved birth spacing, which can reduce maternal and infant morbidity and mortality; birth spacing is thus viewed as a proxy indicator of reproductive agency for women (Molitoris et al., 2019; Wolfers & Scrimshaw, 1975). A review by Upadhyay et al. (2014) identified six studies that examined associations between women’s empowerment indicators, including decision-making control, reproductive coercion, and birth spacing, and five of these six studies found significant associations, despite differing measures of birth spacing. One recent study from Kenya that used attitudes toward IPV as an indicator of empowerment found that shorter birth intervals were associated with more permissive attitudes toward IPV (Kudeva et al., 2020).

Intervention evaluation studies also show the potential utility of knowledge and skill building on early marriage and fertility outcomes. Using an RCT design, Dupas (2011) showed that equipping young women with information about the health risks associated with sexual relationships with older partners (e.g., “sugar daddies”) reduced (typically unwanted) adolescent pregnancies, especially pregnancies with older partners. Blending knowledge and skills training, Amin et al. (2016) used an RCT design to evaluate the effectiveness of a program named BALIKA, aimed at delaying child marriage through community-based skills development for girls in Bangladesh. All girls participating in the BALIKA program met weekly with mentors and peers in safe, girl-only locations called BALIKA centers, which helped girls develop friendships, receive training on new technologies, borrow books, and acquire the skills needed to navigate the transition from girlhood to adulthood. Girls used these skills within their communities, which helped build confidence, demonstrate their achievements, and elevate their profiles within their communities. Girls living in BALIKA communities were one-third less likely to be married as children than were girls living in communities not reached by the BALIKA program. Girls who were single at the beginning of the study were one-fourth less likely to be

married by the end of the study. Relevant for the purposes of this committee, in communities where girls received life skills training on gender rights and negotiation, critical thinking, and decision making, girls were 31% less likely to be married as children at endline than were girls in the control communities. Similar programs leading to declines in child marriage are documented by Stark et al. (2018) in Ethiopia. In contrast, recent studies in Nepal and Bangladesh evaluated the impact of sexual and reproductive health education and training girls for leadership and advocacy, in the context of efforts to change community norms among parents and leaders, but found no impact on delayed first marriage (Naved et al., 2024; Yount et al., 2023). Reduced child marriage showed an association with level of engagement in the program in Bangladesh (Naved et al., 2024). A positive effect of vocational skills training is also evident from some studies. Bandiera et al. (2020) evaluated a policy intervention aimed at jump-starting adolescent girls’ empowerment in Uganda by providing adolescents with vocational training and information on sex, reproduction, and marriage. After four years, the program led to sharp declines in teen pregnancy, delayed entry into marriage or cohabitation, and declines in the share of girls reporting sex against their will. Weaker impacts were observed on outcomes such as views and aspirations tied to the ideal age at marriage. One concern is that scale up of the program was limited. Furthermore, similar programs in Tanzania, Bangladesh, and Zambia found small or no impacts on education, marriage, or childbearing (Austrian et al., 2020; Buchmann et al., 2021; Buehren et al., 2017; Waidler et al., 2022).

Adoho et al. (2014) evaluated the impact of the Economic Empowerment of Adolescent Girls and Young Women (EPAG) project in Liberia, a randomized intervention aimed at providing livelihood and life skills training for 2,500 young Liberian women. The EPAG program increased women’s employment by 47% and earnings by 80% through positive effects on a variety of empowerment measures, including increased access to money, improved self-confidence, and decreased anxiety about circumstances and the future. Despite evidence of shifts in gender norms and more gender-equitable allocation of household tasks, the evaluation found no net impact on fertility, sexual behavior, or opinion about the suitable age at marriage for women. Also, in the context of Liberia, Özler et al. (2020) evaluated the effectiveness of Girl Empower, a mentoring program combined with a cluster-randomized cash-transfer intervention, which sought to equip adolescent girls with skills to make healthy, strategic life decisions that would keep them safe from sexual abuse. The program led to noticeable positive changes in gender attitudes, life skills, and sexual and reproductive health but did not affect girls’ reports of sexual violence, level of schooling attained, or psychosocial well-being. Nonetheless, adding a monetary

incentive to the empowerment program reduced the likelihood of marriage and the number of sexual partners in the previous 12 months by over 50%.

Makino et al. (2021) documented positive impacts of two RCTs in Bangladesh and Zambia. The former, providing academic skills training, had a significant impact on discouraging child marriage in villages where girls’ participation in paid work was particularly high. The second study, providing empowerment intervention and safe spaces, had a significant impact on delaying pregnancy, especially for illiterate girls in communities where premarital sex was relatively common.

Employment and Other Economic Opportunities

Studying the effects of the explosive growth in the Bangladeshi garment industry through a natural experiment exploiting the timing of access to garment sector jobs, Heath and Mobarak (2015) found that access to factory jobs lowered the risk of early marriage and childbirth due to two factors: older girls postponing marriage because of a higher likelihood of being employed, and younger girls remaining in school.

Jensen (2012) performed one of the few studies seeking to empirically establish the causal effect of increased opportunity for paid work on marriage and fertility decisions, in contexts in which those decisions are taken by parents. The author worked with business process outsourcing centers to organize recruiting sessions in randomly selected villages in Northern India, providing an exogenous increase in women’s labor force opportunities for women in these villages. The study found that women in treated villages were less likely to marry and have children and more likely to pursue additional schooling and training. The study also reported that women in treated areas wanted fewer children and desired to work steadily throughout their lifetimes.

Heath and Jayachandran (2016) reviewed evidence on the effects of increases in female education and labor supply on socioeconomic development. They concluded that

in response to the expansion of a pension program in the country (Rossi & Godard, 2022). Donald et al. (2024) analyzed data from six RCTs of business-training and land-titling programs in Sub-Saharan Africa that increased women’s earnings and household wealth, and found increases in women’s fertility—with results driven by those women most in need of sons for support in old age or widowhood. Future research investigating the relevance of these findings for other contexts is important, as is research to causally assess the impact of women’s access to social protection and insurance programs on their fertility.

Fertility Decline and Gender Equality

Evidence on the impact of the fertility decline on changes in gender equality is limited. One study of fertility decline and women’s empowerment in China (Wu et al., 2012) showed that women with lower fertility do less housework and tend to be more satisfied with their family status than women with higher fertility. Other studies indicated that fertility decline does not predict other aspects of women’s welfare, including household gender equality and gender-based violence (Stoebenau et al., 2013). In many contexts around the world, women’s employment and earnings decline precipitously after the birth of the first child, whereas men’s do not (Kleven et al., 2023). Gendered norms and roles have not shifted alongside or following fertility decline. This resistance to change is mostly attributed to deeply entrenched inequitable gender norms and systems (Yount et al., 2016).

Fertility declines can have detrimental impacts on gender equality through two primary mechanisms. First, when the son preference is strong, a decline in fertility can increase the skewness of the sex ratio. A vast body of demographic research has documented that declining fertility, coupled with continued preference for sons, is associated with declining proportion of girls to boys (Das Gupta & Bhat, 1997; Larsen et al., 1998). Jayachandran (2017) provided evidence that, in India, this association is likely causal: using an innovative survey in Haryana, she showed that respondents reported preferring a more skewed sex composition when asked about a hypothetically low versus hypothetically high total fertility level. The study estimated that the fertility decline could explain up to 50% of India’s recent increase in sex ratio skewness in the past 50 years.

Second, increased sex ratio skew can exacerbate the practice of bride importing. To the extent that bride importing is associated with low agency of exported/imported women, this may be detrimental for gender equality (Kawaguchi & Lee, 2017; Mukherjee, 2013). Proportions of marriages with international brides grew rapidly in the late 1990s in Taiwan and early 2000s in Korea (Lee, 2010). This change occurred as the effects of

an increasingly adverse sex ratio became noticeable. Bride importing in Taiwan, Korea, and China increased as adverse sex ratios began to create a marriage squeeze (Cho, 2018). In South Korea, for example, male-to-female sex ratios for individuals ages 25–34 began to diverge from 2003 onward. In the same period, the number of foreign brides increased from about 25,000 to about 140,000. Interestingly, despite many economic similarities, Japan saw neither a substantial rise in adverse sex ratio over the period of fertility decline nor a vast increase in foreign marriages, providing a useful counterfactual (Lee, 2010). While international bride import receives considerable attention, substantial import of brides within countries with demographic diversity also exists, generally from distant regions to areas with marriage squeeze. Research in China and India documented a substantial increase in marriage arrangements in which brides were imported into areas with adverse sex ratios (Kaur, 2016). Finally, a study of Indian states showed that states with adverse female-to-male sex ratios had greater reports of child trafficking with a 10-year lag (Prakash & Vadlamannati, 2019). However, child trafficking was also reported more frequently in states with greater women’s empowerment, raising questions about whether these states experience more trafficking or greater reporting of trafficking.

WOMEN’S EMPOWERMENT AND MIGRATION

Voluntary Migration

Ample evidence suggests that work migration can lead to large income gains for migrants and the families they leave behind (Clemens, 2011). While the proportion of women among labor migrants rose steadily between 1960 and the early 2000s (Donato et al., 2011; United Nations Population Fund & International Organization for Migration, 2006), work migration remains disproportionately male, raising the question of whether women’s empowerment programs can increase women’s ability to take advantage of migration opportunities. The answer depends on the context. Social barriers to women’s work-related migration are documented by a study in Bangladesh that examined women’s migration before and after construction of a bridge that increased communication and reduced mobility costs. Comparing areas benefiting from bridge construction with those that did not benefit, and comparing younger cohorts that came of age once the bridge opened with older cohorts using a difference-in-difference analysis, the study found that increased access improved the likelihood of women’s migration for marriage but not for work (Amirapu et al., 2022). The authors suggested that social barriers reduce women’s likelihood of taking advantage of reduced migration costs.

___________________

¹ In particular, women pioneered and dominated migration from Jamaica and the Philippines to the United States, due to employer preference for female service workers (Foner, 2009; Tyner, 1999).

appears to curtail women’s abilities to migrate (Ferrant et al., 2014). While discriminatory social practices may lead to increased migration, migration to countries with lower levels of discrimination may also lead to changes in norms in the origin countries, through what has come to be defined as social remittance (Levitt, 1998). To deal with this reciprocal relationship, OECD data were extended to assess the two-way relationship between discriminatory social institutions and migration, using instrumental variables with the date of ratification of the Convention on the Elimination of All Forms of Discrimination Against Women and religiosity as instruments for the level of discriminatory social institutions (Ferrant & Tuccio, 2015). Results showed that the unequal status of women and their low decision-making power within the family constrained their ability to choose their own development pathways, resulting in a lower probability of engaging in international migration. In addition, an undervalued status of women within the family was shown to reduce the size of the female migration flow. Therefore, high discrimination at the family level seems to affect women’s migration opportunities. A similar effect is not observed for men (Ferrant & Tuccio, 2015).

Forced Migration

Literature on forced migration (Klugman, 2022) has documented its impact on the increasing vulnerabilities of women. However, the way that women’s empowerment or lack thereof influences the likelihood of forced migration has received much less attention. One study reported results from a five-year-long intervention in Nepal, Bangladesh, and India aiming to prevent trafficking of women workers and their exploitation (Zimmerman et al., 2021). The study did not show that interventions empowering women through knowledge and attitude change resulted in reduced forced servitude. The authors argued that systemic changes and increased support systems for women workers in their destination countries, rather than simply changes in their own knowledge and attitudes, may be important to reduce women’s vulnerabilities.

WOMEN’S EMPOWERMENT AND MORTALITY

Estimating the Magnitude of Missing Women

Sen (1990, 1992) was among the first to estimate the millions of “missing” women among the population of the Global South, particularly East and South Asia. Observing that China and India had much higher ratios of men to women than other parts of the world, Sen (1990) concluded that over 100 million women are “missing” due to the neglect and maltreatment

of women and girls. These imbalances reflect gender inequalities in several outcomes, including mortality—Sen’s original focus—as well as in morbidity and the risk of induced abortion, which is also discussed by Lane et al. (1998) in the context of Egypt. The specifics of Sen’s “100 million” calculation remain a matter of controversy. Initial debates focused on whether researchers should use sex ratios from West or Sub-Saharan Africa to estimate the sex ratio that would likely be observed in Asian populations in the absence of gender discrimination (Coale, 1991; Sen, 1992). Subsequent research clarified three core issues that affected the calculation of the missing women: the age structure, the natural sex ratio at birth, and the overall level of mortality in the population. Accounting for these three factors, Bongaarts and Guilmoto (2015) estimated that 126 million women, or 4% of women globally, were missing due to discrimination in 2010. Their results were driven by excess female mortality from birth through late middle age, principally in China and India. Within those countries, they estimated that 10% and 7% of women were missing, respectively.

It could be argued that the absence of these women—along with the implied excesses in morbidity and mortality—result from limits in women’s agency, since premature mortality necessarily implies a loss of control over one’s life. Greater agency could possibly enable women to avoid premature mortality and unnecessary morbidity as adults, while also empowering them to shield their children, especially daughters, from the same outcomes. These influences could operate at multiple levels—by empowering women at the individual level and also by transforming communities, governments, and societies to protect women and girls.

Among adults, most excess female deaths occur in middle age rather than old age (Anderson & Ray, 2010; Bongaarts & Guilmoto, 2015). Causes of excess female deaths vary by region, but one consistent source in early adulthood and middle age is maternal mortality, defined as death during or soon after pregnancy from causes related to pregnancy or birth. Maternal mortality has declined substantially in all regions but remains very high in some regions (Alkema et al., 2016). Globally, the maternal mortality ratio nearly halved between 1990 and 2015, from 385 to 216 maternal deaths per 100,000 live births. However, maternal mortality remains more than an order of magnitude higher in Sub-Saharan Africa, where deaths stood at 546 per 100,000 live births in 2015, than in developed regions, where it stood at 12. In Sub-Saharan Africa, a woman experiencing 2015 mortality and fertility rates throughout her lifetime would have a 3% lifetime risk of dying from maternal causes (Alkema et al., 2016). Another cause of death is IPV and femicide, which has gained currency in global research and political discourse (Graham et al., 2021).

household and extended family offer insights. In India, when multiple adult brothers live in the same household, wives of younger brothers are assigned lower status. Pre-marriage characteristics are comparable between wives of younger and older brothers, so that comparisons of post-marriage outcomes shed light on the health effects on women’s intrahousehold positioning. Compared to other wives in the same household, wives of younger brothers were found to have less say in household decisions and were more likely to be underweight (Coffey et al., 2022). The weight finding is thought to be related to the practice of eating in order of social rank, such that household heads eat first, followed by their sons, then their grandchildren, and then their daughters-in-law. The youngest daughter-in-law cooks the food and eats it last (Palriwala, 1993). Consistent with this interpretation, other Indian datasets have indicated that women with less say in household decisions were more likely to eat last (Hathi et al., 2021), and women who ate last were more likely to be underweight (Coffey et al., 2018).

The birth of a girl was also shown to reduce a woman’s standing in an Indian household (Das Gupta et al., 2003). Using data from India, Milazzo (2018) documented evidence of “missing” mothers of first-born daughters relative to first-born sons, suggesting increased mortality, and found that among survivors, mothers of first-born daughters were more likely to suffer from anemia than mothers of first-born sons. Notably, Indian mothers with first-born daughters reported having more say in household decisions than those with first-born sons (Heath & Tan, 2018), implying that if women’s agency explains the health and survival differences, its mechanism of action goes deeper than everyday household decisions. The mechanism is likely related to the tendency of mothers with first-born daughters to endure more pregnancies, more abortions, and more closely spaced births, all of which carry health risks. One can interpret the increased reproductive burden as a restriction on agency, but it is different from control over everyday household decisions.

India displays well-known fissures in women’s agency at a societal level, with implications for child survival. Dyson and Moore (1983) famously noted that South India exhibited lower infant and child mortality than North India, which they attributed to the greater agency of women in the South. Along with lower overall infant and child mortality, Dyson and Moore (1983) documented that South India also had lower female-to-male ratios in infant and child mortality, which they again argued was due to greater female agency in the South. Carranza (2014) linked South India’s greater equality in survival between boys and girls to the economic value of women, which was seen to be greater where clayey soils prevented deep tillage, as in the South. She documented that women comprised a larger share of workers in districts with clayey soils, and child sex ratios were correspondingly less biased.

level, agency is most often measured as self-efficacy or through indirect measures of skills and knowledge, while intrinsic motivation and aspirations are rarely measured explicitly. Conversely, at the interpersonal level, agency is most often proxied, relying on measures of status inconsistency between partners, such as differences in age, education levels, decision-making power, and occupation. Empirical evidence measuring agency with clear constructs at the community and societal levels is missing, primarily due to the complexity, vagueness, and, at times, sensitivity of measuring social and gender norms.