CGD Podcast: Tackling Child Marriage with Imran Matin and Kehinde Ajayi

Rachel Glennerster: Welcome to the CGD podcast. I'm Rachel Glennerster, President of the Center for Global Development. Today we're talking about child marriage. One in five women aged 20 to 24 were married or in a union while they were still a child. This has profound impacts on their life and their future trajectory. They have to drop out of school to get married. They often then have children quite young in countries where malnutrition is poor, and becoming a bride and a mother while you're still a child has important psychological impacts. Joining me today are two economists who have not just thought a lot about child marriage but are doing something to address it.

First, Imran Matin is Executive Director of BRAC University's Institute for Governance and Development, and formerly head of BRAC's research department, and also deputy of BRAC International. Now, if you don't know, BRAC is the world's largest non-governmental organization and, in my view, one of its most innovative and evidence-based. I'm also joined by Kehinde Ajayi, Senior Fellow at CGD and Director of CGD's Gender Equality and Inclusion program. Along with CGD colleagues, Kehinde has just completed analysis of the cost of inaction on child marriage, and the state of global funding to address it. Welcome, Imran and Kehinde.

Kehinde Ajayi: Thank you, Rachel. Great to be here.

Imran Matin: Thank you.

Rachel: Imran, you work in Bangladesh, and Bangladesh has made a lot of progress in many aspects of gender equality, and it has made progress on child marriage, but it is still the second largest country for child marriage in the world after India, with over 40 million young women who are married, or in a union before they're 18. Can you tell us why you think the rate of child marriage is still persistently high in Bangladesh, and also just give us a picture of what it means for girls to get married while they're still children?

Imran: Bangladesh now, and especially compared to even our neighboring countries in South Asia, progress have been relatively less. I think, historically, if one just looks at Bangladesh itself, there has been a reduction. It has stalled slightly during COVID, but I think it's reducing. What's really important is that though the aggregate rates remain to be not reducing at a very high rate, there is a structural shift that is happening, as you would know. Now the problem is really marriage after 15, 16 years old, and not much before 15 years old. Some of the more worst forms in some ways have been reducing.

Now, I think what this basically means is that the nature of the problem shifts from just being perhaps a simpler problem around poverty, and of parental decision-making to, one, where it is about adolescent sexual urgency, and adolescents own choices as well. We have been doing some interesting research in that area in terms of what is the effect of those perceptions, and how that contributes to child marriage, and what can we do about it. From BIGD, the institute that I run, we have been doing a very innovative program that BRAC is running right now, where they're trying to tackle essentially this more complex problem of early marriage.

In early marriage, one of the first things that happens is discontinuation of education and the formation of the human capital that would come from the kind of education from schooling. This happens primarily at the secondary school level and beyond secondary school level, where actually the human capital formation is the most important. I think that clearly has got long-term labor market outcome challenges, of course.

Then pregnancy. Childbirth happens very soon after marriage. Early marriage basically means the kind of complications that you would get with respect to reproductive health, and pregnancy complications, and low birth weight, and child malnutrition, and so on so forth. You really create much longer-term intergenerational consequences of early marriage.

Rachel: When I was working on this issue in Bangladesh and was interviewing girls, it was very striking that because of the levels of stunting in Bangladesh, you're thinking about having a pregnancy in these somewhat pretty frail and not developed bodies, which causes problems.

Kehinde, let me turn to you for the global picture, but before we do that, I think it's really worth making clear that when we talk about child marriage in different countries, it means often very different things. In Bangladesh, when you get married, you have a dowry, and that's a big financial commitment for the parents. In the majority of cases, the marriage is organized by the parents, and I think in our survey, only 1% of people had even met their future spouse before. Whereas in other parts of the world, you have unions, which is why when I defined it, I said marriage or union when women are still under the age of 18, and that can still have the same consequences of getting pregnant and dropping out of school, but it might not be a formal marriage.

Given that picture of different ways that marriage works and unions work across the world, what's happening across the world, whether places are having the same kind of improvement, but not dramatic improvement as in Bangladesh, or whether some places are doing worse than Bangladesh and seeing very little progress at all.

Kehinde: That's a great question, Rachel, and a really important point to emphasize is that first of all, the rates of child marriage across the globe are very different, and we have some places like Bangladesh where still over half of women who are aged 20 to 24 were married before they turned 18. We have places where the prevalence is much lower. The question you asked about trends that we've seen over time. There are places that have made a lot of gain. In Bangladesh, for example, the rate was over 75% in 1997, so 25 years ago, and it's dropped substantially, even though, as Imran mentioned, less than in other countries.

There are some places Central African Republic where we're actually seeing, unfortunately not just a stall but slight increases in the rates of child marriage. There is a very different distribution of progress. There are also very different drivers. As you mentioned, in some places it is more a case of parents arranging marriages, and this happening at very much younger ages. In other places, there are more economic drivers. In some places like Bangladesh, almost all childbearing is within marriage. In other places, we don't see such a tight connection. In some places, getting pregnant means you have to get married, and that's one of the drivers. In other places, about half of adolescent pregnancy occurs outside of marriage. In some places, it's a question of education. We see this as the strongest, most consistent driver of delays in child marriage. Overall, there are just very different factors that drive it across different countries, within countries, and across space.

Rachel: In some ways, the consequences are quite similar in the sense that people have to drop out of school.

Kehinde: You mentioned in the introduction. Along with Gabriela Smarrelli and Radhika Nagesh at CGD, we've just finished up an analysis of the economic cost of inaction using data from 27 countries that represent 70% of child marriages across the world. What we find is that child marriage, if we were to put an economic cost on it, costs these countries up to $175 billion a year. This is just focusing on four main drivers. One is the under-five mortality. We have increases in child mortality precisely for the reasons you and Imran just mentioned, the fact that we have these young bodies getting pregnant and childbirth is much more risky for both the mothers and the children who are born. The children have lower birth rates and lower chances of survival.

We see also increases in maternal mortality, which we estimate between about 123 million deaths a year, and $139 billion a year, coming from increases in child mortality, and maternal mortality, and then also increases in experience in intimate partner violence, and experience of miscarriage and abortion.

We have those health impacts, and then we also have impacts on education, and this leads to increases in costs from lost earnings. Girls who get married as children are more likely to drop out of school, they're less likely to participate in labor force, and when they do participate, they tend to earn less. We estimate that there's $45 billion a year that countries are losing from earnings because of girls being more likely to drop out of school. We rarely see child brides enrolling in school. Beyond the health concerns, beyond the human rights concerns, there's a real economic impact or economic consequences of child marriage, and an incentive to address it from an economics perspective.

Rachel: I think it's worth coming back to Imran's point about the difference between very early child marriage under 15 and then 16, 17-year-olds who are getting married and the shift in Bangladesh. I think there's surprisingly little research on this actually. A lot of what we see are correlations, but it does seem as though a lot of the health impacts are particularly important for those very young 14, 15-year-olds who are getting married.

The education impacts can be very severe if you drop out of school without completing secondary school, which often 18 is around the time that you finish secondary school. We found in our research that just delaying marriage until you're 18 and finish secondary school actually has a big impact because it's socially acceptable in Bangladesh to keep going to college after you're married, but you can't go to school. You can just get people up to the end of secondary school, and then they get married. They can actually continue their education in college.

Let's talk about solutions. We do have some programs that have worked. Imran, BRAC has been working with adolescent girls for many years, both in Bangladesh and across Africa. Can you tell us some of the things that you've had rigorously tested, and have found to be effective in addressing the challenges adolescent girls face in early pregnancy and marriage?

Imran: For BRAC, this whole problem of child marriage has been always linked with how to really provide interventions through which adolescents can be more empowered. Of course, at this age, economic empowerment would be really preparing for subsequent choices to be able to navigate, and to be able to negotiate. For BRAC, organizing people have always been at the heart of any development intervention. Even in the case of adolescents, been organizing them in the form of some groups, club-based activities, and providing them different types of life-scale-oriented sessions as a part of that.

We are inspired a lot, Rachel, by work that you have done in terms of the whole cash transfer dimension of it and seeing such high impact. I think there's a lot of interest in BRAC in terms of seeing, testing, whether combining the empowerment interventions that BRAC does very well along with the transfer, what are the additional impact that we basically get? The institute that I run, BIJD, we have been evaluating a very exciting program that BRAC has started on tackling child marriage primarily. It's called Swapnosarothi, which means Dream Catchers. It started in 2023. There are four arms, one pure control, and we have the adolescent club, the empowerment arm, the Swapnosarothi Pure Arm. That is what BRAC usually does.

Then we have two other arms. One is with low cash, which is about 500, 600 taka per month until they're 18. Another one, which we call High Cash, which is 1,000 taka, slightly more than that, 600 taka. We find a number of very interesting impacts. In the midline, we find that the Pure Swapnosarothi along with High Cash has the most impact. Swapnosarothi itself has an impact on early marriage, but moderate level of impact, but that too in the longer run, not in the midterm. Swapnosarothi Pure with low cash does not really have an additional impact. I think that is really interesting.

The way this impact is happening on early marriage, and also in education, in education, we find the impact is more in terms of dropout, reducing dropout rates substantially. Again, we find that impact in the High Cash arm, not in the Low Cash arm at all. What is happening here, we think, is that actually the attendance rate, the participation rate in the program is much higher. The engagement rate is much higher in the High Cash arm, substantially higher than it is for Low Cash, or just Swapnosarothi.

It is that higher level of regular participation and that engagement we think is having an effect. It is the cash plus the engagement that is happening in the High Cash arm that is perhaps together creating a much higher level of impact that we're seeing both in early marriage, and in terms of education dropout. 1,000 taka per month is quite comparable to the education stipend program that Bangladesh government currently provides. It's not really crazy amount in terms of real scale up.

The other very interesting thing that we found is that social media usage is quite highly correlated with this parental perception of elopement risk. This fear itself translates into early marriage decisions, if you like, with basically parents pay. It is that perceived risk. If we can somehow, to borrow a word Rachel, you use, if we can somehow create disturbance in the signals, if you like, that these perceptions create, we maybe jumble up the signal in some ways, if you like, then we may be able to shift this. This is exactly what we find. The biggest impact that we find in terms of early marriage reduction is in the subgroup where there is high social media use at baseline, and high parental perception of risk. That is the group where we find the biggest impact in terms of reduction in early marriage. This is, I think, quite a promising intervention here.

Rachel: Very exciting to hear these new results revealed on the podcast. I think just to give some background because you were referring to work that I have done, that I've mentioned various surveys in Bangladesh, but this was a finding that if you provide an incentive to not marry your daughter while she's a child, then we saw a 19% fall in child marriage. This is a regular payment that would be paid to families while they have unmarried adolescent girls and stops if they get married.

You're adding that to another program on empowerment. I should say, in this case, the empowerment program you're talking about didn't have an effect on its own, but in various studies in sub-Saharan Africa, where BRAC did these girls clubs and life skills, I think, a bit more intensively than you're currently doing in Bangladesh, it did reduce teenage pregnancy quite a lot.

Again, that was the big risk in sub-Saharan Africa was teenage pregnancy, and then leading to dropping out of school. Life skills worked for that in sub-Saharan Africa. This dowry payment and real risk that parents feel that girls might elope or might be tainted by being seen talking to boys and therefore won't get a good match is very important in places like Bangladesh and South Asia. It is very much a financial decision. Dowries actually go up. The amount you have to pay to marry your daughter goes up as she gets older. Offsetting that financial burden so that it's not more costly to marry a daughter when she's older actually seems to really move the needle. It's great that you're finding that in your case. Kehinde, can you tell us about what approaches have been tried elsewhere?

Kehinde: Yes. As you were talking about dowry, this is one thing that really varies whether marriage is associated with dowry or bride price. Whether you get paid for providing a daughter or whether you have to pay when your daughter gets married, that really does inform what an effective solution is. You can imagine if the incentive to marry your daughter is because you might get a payment for them getting married, that's going to lead to a different set of solutions than if in order to marry your daughter, you do need to have some additional funding.

We have seen that the effectiveness of cash transfers and conditionality around cash transfers really does depend on what the dynamics of the marriage market are. Another thing that people often think about intuitively as being something that would address child marriage but has had mixed success is laws. The key takeaway is that legal reform on its own often can be ineffective because we do see either limited enforcement or a shift from formal marriage to marriage that is not yet formalized but a de facto marriage basically where people are in these unions and are living as married but not going through the formal thing until it's the legal age. That in and of itself has had limited impact.

The other thing that people often think about is social norms change. That has proved to be also quite difficult to systematically scale up and implement things that will address social norms. What we have seen, as I mentioned earlier, having broad impacts are things related to education. Scholarships for girls, free secondary education, things to support transport for girls, making schools more girl-friendly where you have more female teachers, especially in rural areas where there are concerns about girls being in environments where there are a lot of men, not just rural areas, but places where you have fewer female teachers. Some of the things that come into these, ILA, Empowerment, Livelihood, and Travelessence type of programs, which are about having mentors for girls, and giving them role models, those type of things related to education, and incentives for education really do tend to have strong impacts. Then this economic strengthening types of interventions that you mentioned, Rachel. Something that often comes up that's associated with this and I do want to highlight is the questions of funding this.

In working on this report, we collaborated with Girls Not Brides, Girls First Fund, and Publish What You Fund to integrate some analysis on the current funding landscape. It's really striking that over the last several years, less than 0.025% of official development assistance was going towards projects with a primary objective of addressing child marriage. Similarly, if we look at a broader definition of funding to projects that have ending child marriage as one of multiple objectives or as one of their result indicators, only 0.08% of total ODA goes to that. There's a very small amount of funding that is dedicated towards this issue, and a lot of scope for thinking about how can we better design education projects, social protection projects, reproductive health projects tailored to address the particular needs of adolescent girls, and leverage these type of investments, which tend to take up quite a bit of both government domestic funding as well as ODA. How can we leverage these pockets of funding to better address the issue of child marriage, and really integrate some of the insights from these evaluations from the large body of evidence that we have on what is effective of addressing child marriage.

We're seeing a lot of the countries that have historically been the largest funders of interventions to address child marriage cutting down their funding even more than the overall funding cuts that we're seeing. It's really a crucial time to think about domestic funding to address child marriage. One example that comes from Kenya is the Kenya Social and Economic Inclusion Project, which is a project that's jointly financed by the Kenyan government and the World Bank, and is integrating a focus on adolescent girls, so using the social protection system to support adolescent girls at scale through an intervention that is like the project that BRAC has really been a leader in developing and demonstrating the effectiveness of.

Rachel: Just to follow up on this funding point, increasingly, middle-income countries, including lower middle-income countries, are having to fund their own solutions to development problems. One of the things they have prioritized, interestingly, is education. As you saw democratization go across Africa, you saw free education follow with it, because that's something people really care about. Politicians were responding to pressure from the population. What of these solutions do you think will be taken on by governments, and what won't be? I'll ask Kehinde first, and then end up with Imran.

Kehinde: I will start off by referencing some of my CDG colleague Dave Evans' work, which is finding that a lot of things that are good for improving education overall can be effective at improving girls' education at the moment. Within the education system, we do find, and there's a lot of work demonstrating this, that when you do make education free, it often does support increased enrollment for girls, especially from vulnerable areas, but complementing that as much as possible with targeted interventions also can be helpful to offset costs.

In terms of what I think governments are going to adopt, certainly free education, as you mentioned, partly for these political economy reasons, but also thinking about teaching. This is one that I think can be lower cost than providing free secondary education, is thinking about how can you get more women as teachers and how can you get them in areas where there are limited representation of girls. That is something that I think governments could take up.

Also, I do think within the social protection system, we do see that, especially after COVID, there's a lot of infrastructure that was strengthened in terms of targeting and identifying vulnerable households and linking that with complementary interventions to support adolescent girls. I think that is also an opportunity where we can see that being used. For example, if you're giving cash transfers to households already, attaching some of these insights that you mentioned, like conditionality that is linked to child marriage or girls' enrollment, depending on the context and what the drivers are, or even self-conditionality with messaging, even if it's not enforced, but using an opportunity of having connections to these households to say, "Girls, education is really important," communicating what the benefits are of delaying child marriage, and using these entry points to support vulnerable households to address these issues.

Rachel: Yes, I'm slightly less optimistic on trying to persuade people because we found in our Bangladesh surveys that people thought that girls should stay in school, but these financial pressures really went against it.

Kehinde: Really, the social protection system, leveraging that as a way to scale up transfers, I think is really a promising avenue.

Rachel: Yes, although that's expensive.

Kehinde: It is expensive, yes.

Rachel: Imran, what hope do you have for you found this effective program in Bangladesh? I know BRAC was seeking to do a program that was cheaper than its previous program that was effective, but very much relied on donor support, which ended up drying up. What's your hope for how you get this new program funded?

Imran: The education stipend program that we currently have can be a very good platform for this, and there's a lot of political--

Rachel: When you say you have, the Bangladesh government?

Imran: Yes, Bangladesh government. I'm talking as a citizen here. [laughs] Yes, so I think because there's a lot of political, and this is something that successive governments have been scaling, continuing, and the current government started this, so they are going to be really excited about using the education stipend program more innovatively, more effectively, scaling it up further, and so on and so forth. That gives me a lot of hope, with a few caveats.

One is, I think because the problem much more right now is around the secondary education part, and our transfer is to adolescent girls, and 50% of the transfer is given to adolescent girls every month, and the rest 50% is turned into a lumpsum, and provided after they're 18-year-old to the adolescent girl, but, obviously, parents get also quite interested in that lumpsum.

Now, how you design the cash transfer is really important, and we don't have any experimental evidence on that, but our qualitative work, and our process evaluation work that we're doing suggest that this is really important, because it is renegotiating the bargain, and trust between adolescent girls who are slightly older and their parents. This is where you need to really reward both. It's not about the amount of money, so much. It is how we design this, and I think it is absolutely possible. This will require innovation. It will require partnership between NGOs and the government to be able to really design this, and basically deliver this because the other part is, of course, the social empowerment bit, which we believe is really important, because it is the social empowerment bit, along with the higher cash together that has this big impact. I think that's also very important, and that would be how we embed this within either curriculum or extracurricular activities within the school.

Then the final problem in the policy scaling up is that this education stipend is only for government schools, but we have large majority, especially in the secondary level where this problem is, in private and other types of schools. That needs to be scaled up as well. There are these type of challenges, but I think the policy platform for this, and the instrument for this is going to be the education stipend.

Rachel: At least there's something there to work with. When you were talking, it just reminded me that there are other things that have been found to work on child marriage that we haven't talked about, which is job opportunities. Again, in Bangladesh, but some evidence from India, too, of something similar, which is when a garment factory arrives in Bangladesh, in a particular location, you see a reduction in child marriage. Which I think is a little bit counter to what garment factories and women working in them are seen as negative in many parts of the West. Actually, it's an incentive for girls to invest in education because they think they can get a job. We've seen similar things in India.

It's great to talk to you both about all the different things that feed into this complicated issue of child marriage, which differs around the world. It's different at different ages, but has large consequences on the future of women, and countries where they are, because it often means dropping out of school. That means less productivity, and particularly at earlier ages, means potentially big health costs. It's really an urgent issue that we need to address. As we wrap up, if there was one policy you could change in the world, it could be about child marriage, it could be about something else, what would it be? I'll start with you, Imran.

Imran: I think it will be a meta-level policy that we don't always have to look for new innovations. I think there are many innovations that already we know work. We just need to find the resources and adapt it in our context to scale it up. I think it is really so wasteful that there's so much of really good proven models that are not being scaled up. I think that would be a matter level policy that I'll really push for.

Rachel: Okay, more evidence-based policy making-up. Music to my ears. Kehinde?

Kehinde: I'll pick up on what Imran had mentioned about the foundation for BRAC's work being about agency. For this, I would have a policy where adolescent girls, and whoever in society is seen as the people who would primarily benefit, or be impacted by the policy get to be part of the policy-making process. Having adolescent girls at the forefront, sit at the table in terms of defining education policy, sexual reproductive health policy, training, livelihoods, those types of policies, the labor market policy. They could say, "How would I shape this policy to make sure that it supported me and my interests," rather than being the recipient of whatever grownups are deciding, whoever the powerholders are. Really, having adolescent girls, and other marginalized groups have a seat at the table in defining policies.

Rachel: Okay, ambitious but hopeful suggestion from Kehinde. Thank you both for joining us. This has been a great discussion, and thank you, listeners. Join us again shortly for our next session.