Cash Transfer Apostasy

Robin Hanson writes:

Let’s say you have been promoting some view (on some complex or fraught topic – e.g. politics, religion; or any “cause” or “-ism”) for some time…

Imagine, if you will, that the world’s destruction is at stake and the only way to save it is for you to write a one-pager that convinces a jury that your old cherished view is mistaken or at least seriously incomplete.  The more inadequate the jury thinks your old cherished view is, the greater the chances that the world is saved. The catch is that the jury consists of earlier stages of yourself (such as yourself such as you were one year ago).  Moreover, the jury believes that you have been bribed to write your apostasy; so any assurances of the form “trust me, I am older and know better” will be ineffective.  Your only hope of saving the world is by writing an apostasy that will make the jury recognize how flawed/partial/shallow/juvenile/crude/irresponsible/incomplete and generally inadequate your old cherished view is.

I studied international development at university and got a pretty uncharitable view of the field while there. My views are a bit more nuanced now but on the whole I still find much to criticize in the way that we spend money and work in other countries. For my money though, development has gotten two things as right as right can get. Global health is one. Global health successes are one of the great un(der)told success stories today – for whatever reason the New York Times doesn’t run articles titled, “122 Million Children Saved in Last 25 Years, More to Come.” Vaccines, bed nets, sanitation work, and our responses to health emergencies (e.g. Ebola) are all things of which we should be proud.

The other thing we’ve gotten right is cash transfers. Cash transfers are exactly what they sound like: an organization gives money directly to people. That’s it. The best of these programs use mobile payments so there are no middlemen to take out a portion and low overhead. Just people getting money. Recipients use the money to pay for things they normally wouldn’t be able to afford because of the nature of poverty: food enough for their families, tin roofs that don’t need to be constantly repaired, business inputs to create future income, and even savings. And despite many peoples’ misgivings, cash transfers are rarely if ever used for “sin” goods like alcohol, tobacco, or gambling.

Besides health interventions, cash transfers are the poverty intervention with the strongest evidence base of them all. I’m very much in love.

I’ve been looking for an excuse to write my hypothetical Cash Transfer Apostasy for awhile now as a way to refine my ideas and to be honest with myself about where cash transfers fall short. It’s important to note that I wrote this as a critique of GiveDirectly’s cash transfer model specifically, though many of the criticisms can be applied more broadly:

While unconditional cash transfers (UCTs) are one of the most well-studied poverty interventions and a hugely important tool in the poverty-fighting toolkit, there are still several issues with the model. These issues can be broken broadly into two categories: 1) issues with the UCT model itself and 2) issues with the UCT model in the context of other poverty interventions. In addition, because GiveDirectly uses a specific model that is itself only small subset of the larger UCT world and evidence base, there’s a third category of issue that I’ll label epistemic.

The first category of issues has to deal with general problems of the UCT model itself. First, many cash transfer programs cast as wide a net as possible and so potentially miss out on targeted benefits that could accrue to those most in need, e.g. women, youth, and other marginalized groups. This undermines to some extent the utilitarian argument in favor of UCTs over other forms of poverty alleviation; that is, increasing someone’s daily living wage from $1 to $2 seems a priori more beneficial in a purely utilitarian sense than increasing it from $10 to $11. Spillover effects on non-recipients are also not well understood and there is some evidence to suggest that there are moderate negative effects on non-recipients’ economic indicators and an increase in hostility towards recipients of cash transfers. Finally, while there is solid evidence that UCTs increase consumption in beneficiary households, recipients may not have access to supplies or other interventions which UCTs are meant to enable them to purchase. This is an argument in favor of, for example, purchasing and distributing malaria bed nets, vaccines, or other medical supplies that have dramatic long-term benefits but may not be available in recipients’ domestic markets. And while part of the benefit of UCTs is the lack of donor paternalism, it’s not clear to me where the line falls here on the value of local knowledge versus donor technical expertise.

Second are problems that relate to UCTs in the context of other poverty alleviation interventions. GiveDirectly largely markets itself towards individual donors and grassroots fundraising. As it grows however it will have to grapple with the different marginal benefits of UCTs for small donors versus wealthy philanthropists (e.g. Bill Gates) or traditional development institutions (e.g. USAID). If I’m the first of those, it may make sense for me to give all of the money set aside for donation to GiveDirectly to use for UCTs because the absolute size of that donation means I just can’t do all that much. But if I’m Bill Gates, I can leverage my giving for things like funding for research, catalyzing private sector investment, or tackling some of the most persistent development challenges. This last is really where UCTs fall short: while UCTs make for a great portion of overall poverty alleviation tools, the evidence that they address the long-term, complex, and multifactoral sorts of problems that differentiate the success of, e.g. South Korea versus the DRC, is weak. This isn’t to say that there is strong evidence for other interventions, but that generally our understanding of what creates for a successful transition from a developing economy to a more developed one is not clear.

Finally, epistemic problems are those that have to do with how little we really know about how the world works. GiveDirectly and other proponents of UCTs at least partially advocate for the model because of the wealth of evidence of the model’s efficacy. They argue often that in a world where so much is uncertain and development impact is so vigorously debated, UCTs can serve as a baseline of comparison for other interventions (e.g. the statement: “If your intervention can’t beat cash, then you should just be doing cash.”) But this framing often focuses too strongly on the question, “What intervention works in development?” instead of, “What does the world look like and how does it work?” Yes, UCTs have a huge evidence base. Yes, they address some of the fundamental issues in the traditional aid model. But between different country and local contexts; different amounts in different installments to different beneficiaries; different methods of collecting and analyzing data; etc. the research just isn’t strong enough to say specifically that UCTs really are gold standard of poverty alleviation that some of its advocates want it to be. We just don’t know enough. Questions we still need to answer include (and this is by no means an all-inclusive list): are thatched roofs a relevant indicator of household poverty? If transfers are given out in installments versus given in a bundled payment up front, how does that affect recipient consumption patterns? How does the implementation of cash transfer programs impact expectations of recipients and non-recipients into the future?