Yesterday, David argued  that “the important work on trying to raise the incomes and status of women around the world doesn’t continue to come in part by neglecting the important role you [dads] play.” While I don’t think the world of development programs is in any remote danger of ignoring men in favor of women, I do think we aren’t paying enough attention to how men and women interact, and what that means for how programs work (e.g. to increase the welfare of all).
The good news is that we know a heck of a lot more about this than we did 20 years ago, and more people seem to be paying attention to this. A couple of lessons stand out:
· Consumption: men and women, in the same household, spend their money in different ways, and on different things. And where that money comes from matters. There is a very nice paper  on this by Esther Duflo and Chris Udry where they look at the effect of positive income shocks on spending patterns in Cote d’Ivoire. First, they show that expenditures are not consistent with a pareto efficient allocation of household resources. Second, income from different sources is not fungible. Take the case of income from yams. In this context, yams are an appreciated product, i.e. “always under the control of the household head for redistribution to the entire household in the form of food.” Positive shocks to this get spent on food, education, and staples. But rainfall shocks that boost the output of crops cultivated individually by men and women are associated with shifts towards adult and prestige goods. Shocks that increase women’s crops shift expenditure towards food. Shocks to men’s crops have no effect on food expenditures. Bottom line: we definitely can’t assume that households act as a unit, and we can’t even assume that they behave in line with a collective model of the household. What we do need to do is understand how they’re interacting (which this paper, with a strong emphasis on anthropology, is a good example of). And if we understand how their consumption patterns will react, we can figure out who to target and how.
This isn’t just a developing country phenomenon. Shelley Lundberg, Robert Pollak and Terence Wales have a nice paper  in the Journal of Human Resources looking at the shift in paying a child allowance from the husband to the wife in the UK. Presto: household consumption patterns change with more being spent on women’s and children’s clothing relative to men’s clothing.
· Production: Here again, we need to understand how men and women (in the same household) interact in order to make more effective policy – even to raise income. David and coauthors Suresh de Mel and Chris Woodruff have a nice paper in AEJ Applied that makes this point – and shows us how a good survey will let you look at this in the context of experimental work. This paper follows on their earlier Sri Lanka work which showed that a cash grant had a positive effect on men’s business but didn’t do anything for female owned businesses. In this paper (ungated version here ) they unpack this difference in outcomes. It turns out women are skewing their investment towards things their husband can’t expropriate. And this is particularly pronounced for women who have lower measures of autonomy (in household and business decisions). So, efficient households? No. Add to this Chris Udry’s Journal of Political Economy paper  on the inefficiencies in agricultural household production and we have shards of evidence that suggest that in order to grow household income, we need to think about how men and women interact – within the household, as well as in society at large.
But these are just pieces. David is right when he points to the fact that we need more evidence. As he pointed out, even in the case of transfer programs, where there are a plethora of randomized evaluations, there is very, very little that looks at what the effects of transfers to men versus women look like. On the consumption side more broadly, most of the literature relies on unearned income (Duflo and Udry use shocks) to solve the endogenity problem. What would be nice is to put it all together (as David and coauthors do) and look at changes in production and then the resulting consumption impacts. This will yield insights into the gender-differentiated constraints in making income grow, and then what happens to welfare (moms, dads, sons, daughters) when we do.
So what does the ideal research taking this forward look like? First, it requires more data points. If we want to talk about men and women both, rather than some notion of “households” the sample will have to be larger as Alaka and I pointed out  a while back. Second, it is going to require a better sense of what to look for. And here, as we can see in Duflo and Udry’s paper, guidance from our non-economist friends is likely to be quite instructive. Third, it’s going to require good variables – measures of who is doing what, who “owns” what (and how secure that is), who is spending how much on different things, and better measures of the processes within households. And we need to do this in the context of both inferential research and meaningful experiments which tackle issues of growth in enterprises and in agriculture.