coauthored with Alaka Holla
So two weeks ago we talked about how we don’t know enough about economically empowering women and last week we talked about power issues when measuring this in “gender-blind” interventions. This week we’d like to make some suggestions about how, with small effort, we could make serious progress in learning meaningful things about how to increase the earning capacity of women.
One idea would be to take a hard look at the intervention you are thinking of evaluating (especially if it hasn’t had a single gender thought associated with it yet). How might women and men interact with the program differently -- from recruitment through implementation and on to effects – the whole chain. For example, if you are going to use newspaper announcements to recruit your participants, then in many countries this is going to make it less likely that you’ll attract women. And then there is implementation: women may face time constraints (like taking care of the kids) that would make it harder for them to sit through 6 hours of uninterrupted training (whereas if it were in two hour chunks, it might work better for them. And daycare might be better still). And while there is little evidence on differences in the effects of interventions, as we’ve been bemoaning for the last two weeks, programs can generate different impacts on men and women in interesting and non-obvious ways (see, for example, David’s work we talked about last week).
So how can we think about this in a bit more of an organized fashion? The argument we would make here is to target the constraints that hinder women’s productivity. And here you need to look at four things: market failures, institutional constraints, social norms, and relations within the household (an argument that will be more fully developed in the forthcoming World Development Report on Gender). In terms of market failures, you want to look for market failures that bind on men and women differentially. Take the case of sex-specific networks – in many contexts women are more likely to talk with/interact with other women rather than with men (and the same is true for men). Hence, information and the other good stuff that comes with network ties (like long-term contracts) will run along sex lines. And this will mean a lot for how interventions play out. Another example (more of a conjecture): credit market failures may affect men and women differently. The market for sex is much larger for women selling sex than males selling sex. Hence, transactional sex may be more likely an alternative source for credit or liquidity for women than for men (evidence that female entrepreneurs participate in transactional sex has been largely anecdotal, although this  paper from Burkina Faso by Nagot and co-authors presents some empirical evidence). And this can spell significantly different outcomes for an intervention in which credit plays a role.
Institutional constraints can come in a whole range of flavors. Service delivery can be biased against women – this is a common problem in agricultural extension for example (and not just in developing countries). Even with recent progress, laws in many countries still treat men and women differently. Women’s property rights - access to land, control over firm assets, and legal right to open up and/or control a business - might be less than that of men (in a couple of weeks Markus will blog on Rwanda’s attempt to change one aspect of this, and how it paid off). And don’t stop with just laws – regulations can discriminate too (e.g. restrictions on sectors and hours of work). Finally, you should look at how these laws and regulations are enforced – women can be more constrained in their access to the systems that enforce these things.
The third thing to take a look at is social norms. Women’s roles (e.g. women “are never car mechanics”) could bind in a number of places – whether or not they take up the intervention, and then how the intervention plays out among those who do participate.
Finally, intrahousehold relations are going to matter. How the household divvies up resources and how this bargaining plays out (including violence) can affect the intervention at multiple points – as we’ve seen from a number of earlier blog posts and a bunch of interesting papers.
So you have done some analysis. The trick is to see how you can tweak an intervention (and convince implementers) to address at least one of these constraints. And you may need to hit more than one at a time (or in sequence) to have an impact – but this is why you did your homework and thought about gender ex ante.
And you can get a bit more radical. This is where you shift your intervention (or come up with a new one altogether) that tries to change gender roles – getting women to act more like men and getting men to act more like women (CCTs for male provided daycare anyone?). These will be more scary – your counterparts will giggle, nervously, and avoid your calls. But, these can work. Take the string of work surrounding political reservations for women in local government in India (see for example this paper by Beaman and coauthors here  and Bhavnani here ). .
So these are some ideas. Any further thoughts, examples, and experience are most welcome.