There aren't that many development initiatives I know that have their own slickly produced video , sponsored by a major corporation, let alone a parody . But the "girl effect," which makes the argument that investing in adolescent girls is a great thing, is one.
I have to admit, I was skeptical (maybe it was that video). But, in a new paper  written with my colleagues Oriana Bandiera, Niklas Buehren, Robin Burgess, Selim Gulesci, Imran Rasul, and Munshi Sulaiman, I found a reason to rethink my skepticism.
We take a look at the Empowerment and Livelihoods for Adolescent (ELA) program administered by the NGO BRAC in Uganda. The intervention consists of three main activities for girls aged 14-20: 1) a social space to hang out, 2) life skills training, which covers everything from sexual and reproductive health to negotiation, and 3) local market-informed vocational training (later work will look at a fourth piece: microfinance). All of these take place within clubs, usually in a small room in the village (rural) or neighborhood (urban). Finally, given that a lot of these girls may still be in school, the club is designed to have its main activities outside of school hours.
To look at the impacts, we worked with BRAC to randomize the intervention across 150 communities in urban and rural Uganda. We conducted a baseline survey in 2008 with 4800 adolescent girls and their parents, and a follow-up survey about two years later.
So what do we find? First of all, take up is pretty decent -- 21% of girls in treated communities join these clubs. And about half of them show some pretty time intensive commitment -- coming to club meetings 1-2 times per week over the two years for which we follow them. Using ITT estimates, we look at outcomes in a number of dimensions.
In terms of sexual behavior, the girls who participate in the clubs show significantly better HIV and pregnancy knowledge than the control group. They are also 12.6 percentage points more likely to report always using a condom when they have sex (which matches up with a reduction in those reporting often or occasional use of a condom). They also experience a striking reduction in fertility – at follow up, treatment girls are 2.7 percentage points less likely to have a kid (26 percent of the baseline mean). Now since they also report no increase in use of other forms of contraception, these things taken together strongly suggest that they are markedly reducing their risk of exposure to HIV.
The program also emphasizes economic activities. Here we find that, relative to the baseline, girls are 35 percent more likely to be engaged in an income generating activity, almost all of which is coming from self employed entrepreneurial activities as opposed to wage labor. They are spending more hours working, and (conditional on employment) are earning more money. Interestingly, we find that this also results in a 33% increase over the baseline in the amount of money they spend on things they consume (jewelry, clothes, cosmetics, air-time, etc). And they don't drop out of school to do this (heck, the ones who have already dropped out say they are more seriously considering going back -- but they haven't).
Does this come with more "empowerment"? It seems so. For starters, the percent of girls reporting having sex against their will drops 17.1 percentage points, or a whopping 83 percent when benchmarked against the baseline mean. We also see some changes in their attitudes on gender roles. One particularly strong result is that they are significantly more likely (+18.2 percentage points against a baseline of 37 percent) to think that females should earn money for the family. We see some mixed results on life satisfaction -- the positive impacts come on satisfaction with income/earnings and less concern about finding a good job in adulthood.
So, these seem like pretty significant effects -- especially when compared with most HIV education or vocational training programs. But what do they cost? Given that what I've discussed above are our ITT estimates, we can also take a look at ITT costs. Using the number of potentially eligible girls in the village, the per girl cost was $28.1 in the first year and $17.9 in the second year. To put these costs in context, we can look at the value of the additional consumption the girls have in the second year. The second year cost of the program is about 21 percent of these expenditures. And that's not including some valuation of the pregnancy, empowerment, and other benefits. Now I am still not sure that this will save humanity, but it looks to me that putting money into a program like this, in Uganda (with one of the lowest median ages in the world), is a darn good investment for girls during the critical transition from childhood to adulthood.