Chris Blattman

Invest in women? The effects of aid on female entrepreneurship

There is a lot of talk about “investing in poor women”. They will grow businesses, get richer, be empowered, and invest in children. So give them a cow or a grant or a microloan.

It’s been a very effective marketing message for aid agencies, but it’s not clear it’s true. Especially the part about women entrepreneurs and women empowered. A lot of the experimental evidence has been pretty pessimistic on this front–to the surprise of many.

This article talks about how established female entrepreneurs don’t have the same returns to cash as men. This one discusses the weak effects of microloans on business growth. This book on how conditional cash transfers do many good things, but starting female enterprises and empowering women is not among them. There’s some evidence giving poor women cows helps, but a host of similar trials (not quite out) are suggesting more tepid results.

Before you get upset or pessimistic, read on.

There’s a big hole in this evidence. These programs don’t necessarily do the simplest thing for the most needy people: find very poor women who don’t already have a business, give them cash, and let them decide what they want to do themselves. Get rid of the conditions and the cows and the ridiculous microloan interest rates and let them do their thing.

Now things start looking up. That’s what I and several coauthors did with AVSI Uganda, working with some of the poorest women in the world. Here is the policy report and brief.

The short story: in 18 months, they become petty traders, incomes double, with a big boost to savings and poverty reduction. (Did I mention that income doubles?)

We also randomly evaluated the paternalistic part–whether social workers held the women accountable for investing the money in business, and provided follow-up visits and advice. This actually boosted incomes (by some measures at least), but not so much that it was worth the expensive follow-up. Better, we argue, to just give more women more cash. Or find a way to deliver the support and accountability cheaply.

Why did this intervention build new female businesses where others have seen tepid results? Hard to say. It’s a different country than the others, and women in northern Uganda may be further behind, and more constrained. So perhaps we should expect more potent results. The more coiled the spring, the bigger the bounce on release.

My hunch, though, is that it mattered that these women weren’t already entrepreneurs, that they were given cash, and that they weren’t confined to cows or tailoring training or other things we think are good for them.

Two other big findings. One is that the cash transfers were large enough that they shifted the whole economy, transforming prices, agricultural wages, and the distribution of income. There were winners and losers. More on this in a future paper and post, but the immediate lesson is this: big transfer programs usually ignore the distributional effects, and (more importantly) the distributional conflicts.

The other big finding: for all the increase in business and incomes, we don’t see any evidence these women feel more “empowered”. No more decision-making power in the household, no more independence, and no less domestic violence (among other measures). And they are no less depressed or stressed. This is a common-enough finding in the psych literature that there’s a name for it–“the impact paradox”. But you don’t hear that very often in the program designs or sales pitches. Aid may have to make the case for investing women on the economic case alone. Forthunately that may be a strong case to make.

The story, of course, is more nuanced. Read the full report or see the policy brief here.

160 Responses

  1. Most findings in that direction really have a limited geographical validity, no?. Or at least the underlying conclusions that would be true anywhere are still very poorly developed. I would love to be able to take findings from Uganda and compare them in some way or another to rural Punjab, but fail to see to yet how that makes sense in any way. We ran a vocational training centre in Pakistani Kashmir and looked at the ‘entrepreneurial and economic impacts’ later on ( Women earned good money with their skills afterwards, but what did they invest that in – in many cases they had the ‘honour’ to finance their own dowry or they were then financing their brothers education. That made them proud and raised their standing. But it left me wondering whether that made sense. The seemingly more sensible impact was that women rejoined educational institutions to find a job not necessarily in what they were trained here – the fact that they managed to earn with that training just gave them the feeling that it was actually possible to earn money for them, which school education alone did not manage.

    Looking forward to your follow up research into this.

  2. Hi Chris, I seem to remember that the first 2yr follow up of this programme found large impacts on men, but no impact on women. What happened?

  3. Thanks for the response. You get at a very important question here — where is the value add for an NGO beyond the transfer. My assumption is that for a particular segment of the extreme poor (i.e., the “ultrapoor”) there are two value adds specific to the non-capital constraints facing this population:
    o identifying and engaging the “ultrapoor” (e.g., people with disabilities)
    o facilitating new connections with economic and social opportunities and resources (“coaching”)

    To the former, I would love to see more research into the downreach of cash transfer programs. Certainly from our experience government and many NGO programs fail to even reach the poorest. (What percentage of a given programs’ population are people with a disability? Globally under <1% despite being ~20% of world's poor.)

    To the latter, I agree we need to justify using cash for coaching. While, like your NGO partner, we do see it as critical to our success, we are looking for an opportunity to test out the relative cost-effectiveness of different levels of coaching and follow-up to try to see where we might gain efficiencies.

    Thanks again for this helpful contribution to the space.

  4. @ChrisProttas: Thanks for the comment.

    I think the NGO we worked with would have said (and still says) the same thing: we see coaching working, and we see empowerment.

    One option is that the numbers lie. Sometimes they do, especially with hard to capture concepts like empowerment.

    The other option is that anecdotal evidence lies. We know from the behavioral decision-making literature that the human brain is really bad at making correct inferences from qualitative experience. We overestimate salient observations for one–giving far too much weight to a handful of experiences. Especially when they agree with our hopes or preconceived notions.

    In the case of coaching, there was certainly an impact. the problem is that the cost of coaching is as much as or greater than the grant, usually because it means sending well-paid people out in vehicles. The relevant question is: are the returns to coaching higher than the alternative, which are the returns to using that aid money to give more cash to more women. Our analysis said, in this case, no.

    To the extent that sending cash becomes cheaper and cheaper through mobile technology, it is going to be increasingly difficult for NGOs to add value in other respects. They need to be innovating and paying attention to the cost of some of their services–trying to bring similar impact at a fraction of the cost. Unfortunately that seems all too rare.

Comments are closed.

Why We Fight - Book Cover
Subscribe to Blog