No loo, No I do

A few weeks ago, a coauthor sent me a job market paper from an environmental economics student at Yale. Though in a very different department than me, we have similar interests and she thought I would find the paper interesting. Not only did I find it interesting, I found myself wishing it had been my job market paper. Apparently, so did a lot of people. The paper has been blowing up my twitter feed and was featured on the World Bank’s Development Impact Blog.

The paper evaluates the effects of a media campaign in Haryana, India designed to encourage women to make latrine presence a requirement for marriage. The project is particularly interesting because it allows for reasonable evaluation of a campaign targeting social norms without the the randomized control component so in vogue in economics right now. In addition, it provides real evidence as to the causal effect of skewed sex ratios. While we have speculated and reported on the effects of sex ratios, many of which I’ve discussed here, there is little statistical evidence. Now, we have some. It’s pretty great.

In summary, the paper shows that men of marrying age are more likely to build latrines when they live in areas with a more skewed sex ratio. Thus, a woman’s bargaining power in demanding a good that has an outsized benefit for her (privacy, sanitation, health) increases when she becomes relatively ‘scarce’ on the marriage market. While this doesn’t discount the other, more undesirable possible effects of a skewed sex ratio (bridenapping, increased violence against women, etc), it is certainly evidence that women are leveraging their bargaining power to improve their outcomes.

In addition, the means to test a social norms marketing campaign are huge. My own work on such campaigns directed at reducing gender-based violence showed the near impossibility of successfully and credibly evaluating their impact. The use of a sex ratio as a (somewhat?) exogenous measure of potential impact is novel, exciting, and I’m sure will be in use by many papers to come. There’s the obvious question of whether it’s plausibly exogenous, but perhaps we’ll save that conversation for another day.

The paper has two parts, one presents a theoretical model to explain the mechanism and the other presents empirical evidence from the program itself to show how a skewed sex ratio has increased women’s bargaining power, at least on this one dimension in Haryana, India. I have some nitpicky comments, like the theory section needs to be more thoroughly explained, or there are square brackets where there should be curly ones, but overall, I think it’s a great paper. It’s kind of wonkish, but you can download the paper here, if you’re interested. Good luck in Chicago, Yaniv!

Left at either end

Rising sex ratios in Asia and other parts of the world have been getting a lot of attention lately. Or at lest, I see them a lot. The idea that men at all socioeconomic levels are being left behind in the marriage market as women become more scarce is one that promises to have effects on everything from gender-based violence to construction inefficiencies as time goes on.

While perhaps a smaller problem, it’s clear that there are women losing out in Asia, too, when it comes to the marriage market, though not for the same reason. As much as economist might love the concept of ceteris paribus, changes in the sex ratio aren’t the only changes sweeping the world. In cities in particular, as women become more highly educated than men and begin to close the wage gap, a culture of “marrying up” means those highly educated, paycheck-earning women are having a hard time finding wives.

I have no idea how many women this might actually be affecting, but I do think it poses interesting questions of social and economic mobility. Like, is the entire distribution of women shifting towards more education and higher incomes? Or is there only movement at the top? And if there’s only movement at the top, then does that mean men are reaching “lower” into the pool, to less and less educated women? Or are the education and income differentials relatively constant?