Is it nuts to give money to the poor?

I’m breaking my blog holiday. It’s a special occasion:

Planet Money reporters David Kestenbaum and Jacob Goldstein went to Kenya to see the work of a charity called GiveDirectly in action. Instead of funding schools or wells or livestock, GiveDirectly has decided to just give money directly to the poor people who need it, and let them decide how to spend it. David and Jacob explain whether this method of charity works, and why some people think it’s a terrible idea.

That’s the blurb from my favorite radio show, This American Life. Jacob also wrote a piece in yesterday’s NY Times Magazine. Both are excellent.

What’s special? Well, in a single day it’s the fulfillment of two of my nerdiest dreams: quotes in both the Times and TAL.

Cash transfers have been my day job for quite some days now. If you’ve been following this blog, you might have seen me describe my study of a wildly successful government program in Uganda, one that sent $8000 to groups of 20 young people to help them start skilled trades. Or an even more successful charitable program in Uganda that gave some of the poorest women on the planet cash to become traders. “Dear governments,” I wrote, “Want to help the poor and transform your economy? Give people cash.”

What’s interesting is that journalists keep turning to me to rain on my own parade. That’s fair, because that’s one of the things I do best. A few days after my plea to governments, I wrote another post, “Why cash transfers are not the next big thing.”

Perhaps that’s why I appear in Goldstein’s article as the skeptical academic. GiveDirectly is very optimistic about giving $1000 to poor people in Kenya. So am I, I say, but the research doesn’t really support it. Yet. Only I didn’t get in the “yet” in that clip. It happens.

Actually, there are couple of good reasons I’m well placed to be the skeptic.

One is that the wildly successful projects I studied gave other stuff, such as training or conditions or social pressure to invest. That probably mattered a lot, and we simply don’t know if pure cash will work as well.

That brings me to the second reason: I have two other projects in the field right now that give plain cash, and the signs are not so good. One is a tough case: street youth and petty criminals in Monrovia. The early signs are that the cash ran through them (but don’t quote me yet, since I haven’t seen half the data). The other is a tough case too: a horserace between getting a factory job and getting a grant to start your own business, both compared to neither intervention. The early signs on cash transfers are not promising, but again, less than half the data are in. So maybe I, and GiveDirectly, will prove ourselves wrong.

So why am I still an optimist? I think sometimes it will work and sometimes it won’t, but that we can develop and test theory to predict that. I think these studies will give guidance about why and for whom cash works best. That’s important.

You would think.

It’s interesting: neither the government nor the charity I worked with in Uganda were willing to try just cash, if only to compare. They wouldn’t even discuss it. This might sound sensible of them, since they could be right about their “other stuff” being important. Except the “other stuff” often costs more than the cash.

This is the big “cost” no one talks about: suppose a charity could give $2000 of stuff to one person, and help them become 200% richer or healthier than they were before. Is it possible I could spend $1000 each on two people, and help get them each get 150% ahead? Wouldn’t that be better?

A lot of charities don’t like to think that way. The TAL episode talked to a woman from Heifer International, who give cows and training instead of cash. That could be the right thing to do. But she couldn’t bear the thought of finding out. She hated the idea of experimenting on poor people. They are human beings.

Let me be blunt: This is the way the Heifers of the world fool themselves. When you give stuff to some people and not to others, you are still experimenting in the world. You are still flipping a coin to decide who you help and who you don’t, it’s just an imaginary one.

You’re experimenting with your eyes closed.

Yes, flipping the coin to see who gets nothing, and measuring these poor souls over time, is different than never talking to them in the first place. But let’s not pretend that never talking to them in the first place is something we can ignore. But that’s exactly what we do. They don’t even enter our minds. Personally, I’d rather flip the coin and measure them.

Where it gets downright immoral to not measure them, I say, is if your program is so expensive it crowds out two other people who could benefit. We don’t know if that’s true or not, since Heifer (shame on them) wouldn’t share their studies or data with the journalists. But I’ve seen many, many, many projects that spend $1500 training and all the “other stuff” in order to give people $300 or a cow. Is it fair to ask, what if we’d just given them $1800? Or what if we’d given six people cows? Seriously, your one guy does six times better than that?

Why do sensible, caring people act this way? Subconsciously, I think humans feel like we owe something to the people we interact with, and pretty much nothing to those we do not. Without thinking about it, we are comfortable with that six-times moral tradeoff.

I would like to make people very uncomfortable about that.

This is actually the reason that GiveDirectly is a big deal. It’s the same reason randomized control trials in aid are a big deal. Economists can argue about whether any result from any study even applies to the village down the road, let alone the country next door. They can complain that the best young minds are answering small easy questions. Many of my colleagues bemoan exactly that. And they are largely right.

But GiveDirectly and randomized trials are helping drive a big, big change: those who help other people for a living are, for the first time, being forced to think about their top and their bottom lines. How much does what we do work? And is it worth the cost?

Believe it or not, these questions don’t really get asked. In fact, it’s so rare, that in my sphere of the aid world (conflict and humanitarianism) a ridiculous number of people have never even thought that way before.

I mean it. I’ve played this part dozens of times: I argue my point; they  sputter and get mad or wave me aside; then they stop for a moment, their eyes focus on the distance, and they say I might have a point. They don’t know if I’m right, they say, but no one ever suggested it to them before.

That is happening a hundred and a thousand times, as big aid donors and little taxpayers  are starting to ask this question to the people that give their dollars away for a living.

If you lead a charity at the moment, you will remember the day many years ago when you rolled your eyes and realized that every second donor was going to want to know what percentage of their money you spend on “administration”. You and I know that is a dumb fictional number, easily manipulated, that misses the end goal.

The good news for you: I predict the day is not so far off when that dumb number no longer matters.

The bad news: the day is not so far off when they ask you for a much harder set of numbers: “What impact are you having, and is it worth what you spent?” or “Couldn’t you have helped six times as many people doing it differently?”

Maybe they’ll even add, “Why didn’t you just give them the money?”

I think this will remake the charity map in my lifetime. If I play but a tiny role in this change, it will be more impactful on poverty and misery than anything else I do in my petty little academic life. So expect me to keep writing long, blathering posts on cash transfers and field experiments as long as you are a reader.

Now back to my blog holiday.

34 thoughts on “Is it nuts to give money to the poor?

  1. This might be my favorite post on this blog ever. I would love to see if this hypothesis is true: “Subconsciously, I think humans feel like we owe something to the people we interact with, and pretty much nothing to those we do not. “

  2. First of all, how is this being compared to welfare payments in developed countries?

    Second of all, how do charities decide who does – and who does not – get cash transfers, and how much each should get? Where does the host government come into this, or other local political institutions that distribute social resources?

  3. Thanks for posting this Chris! The one place I’d question your article is about making people feel uncomfortable. Economists and policy-makers are good at thinking about population-level problems and statistics, but the human brain is inherently bad at dealing with large numbers (even thinking about math imposes a cognitive tax on most people). However, most folks are very comfortable with empathy. Connect to that impulse and it will be a lot easier to have that conversation about tradeoffs and efficiency…

  4. Chris – We have been following you for some time now and are equally enthused about the recent attention to cash transfers. At The BOMA Project, we are implementing a poverty graduation model with sequenced interventions over two years that includes a cash transfer to groups of three women for an income generating business along with training, local mentoring and the establishment of savings associations. Our goal is to build the resiliency of women living in the arid lands by graduating them from extreme poverty into poverty so that they can earn a sustainable income, survive drought, feed their families, pay for school fees and medical care, and accumulate savings for long-term stability. Since 2009 we’ve launched 1380 businesses impacting the lives of 4668 women and over 23,000 dependent children. To your point, we also collect a lot of data on each participant (primarily consumption, not income) and measure for change at one year and at program exit. We are also conducting a phase-in randomized control trial to be concluded in 2014 with 585 businesses and 1755 women. We’ll have some mid-line results in about October. So far, we are seeing average graduation rates of 92% of participants at two-year program exit. Keep asking all those tough questions about cash transfers and thanks for emphasizing impact. There is a lot of pressure from donors right now to have a revenue strategy in models but we believe this is not appropriate for interventions that focus on those living in extreme poverty. The success of cash transfers, and the focus on impact, will help us make our case.

  5. We all agree that aid can be more effective and that well-formed questions and well-executed, applied research can offer many relevant clues about this. We all want to see deeper thinking behind the doing.

    Where I the practitioner often differ from randomistas is on some fundamental beliefs about what prevents this and what ails the aid industry overall. Is it a lack of information about “what works”? Or is it a lack of respect for citizen-led initiatives and a lack of understanding about complex power dynamics that impede authentic relationships among development partners? If it’s the latter, my question is: Are RCTs just a band-aid on a deeper issue? “Why didn’t you just give them the money?” is a valid question regardless.

    Whether RCTs become a part of accepted developmental practice, I do think that it’s important to have dialogues like the one started here. Here was my “how matters” advice for donors on RCTs: http://www.how-matters.org/2011/05/25/rcts-how-matters-advice-for-donors/

    As a commenter on Owen Barder’s blog once shared, “Great tools, we economists undoubtedly do have. In studying development issues, they are often used unhelpfully due to hubris and a shocking level of comfort with ignorance about the phenomenon being studied.” At the end of the day, for those of us involved in organizational learning within organizations, perhaps we should consider RCTs as one of many in the toolbox – the match of the right tool to the right job is key.

  6. Your observation about the shifting importance from overhead numbers to impact numbers is spot on! The “100% of your donation goes to support…” claims are beginning to feel so inadequate when confronted with the question of the impact of that 100%. Similarly, the NY Times Magazine piece about charity:water highlighted that the charity seems more interested in the GPS coordinates of a new well, rather than the functionality and use of that well.

    Thanks for the great post and congrats on the NY Times and TAL mentions!

  7. Great comment by intldogooder. Howard White (Director of 3ie) has a thoughtful piece in New Directions for Evaluation on the use of mixed methods in RCTs (combining factual and counterfactual analysis) to make their findings more useful and relevant for practitioners.

    Journal article (gated): http://bit.ly/13AqPzA
    Slide deck: http://bit.ly/16XGnL1

    Full disclosure – I work with the Aspen Network of Development Entrepreneurs, one of the funders for the factory job study that Chris mentioned.

  8. You have a lot right about what you say but there are a lot of published research that shows our good intentions of walking into a “chosen” village and giving food and clothes collapses the economy and many more end up in poverty…I have wondered about what would happen if you walked into that same village and gave them currency if that would not be more helpful…. But I know loaning money and training women in poverty as Muhamad Yunus has done is successful….have seen for 20 years the huge change in lives and generations of families what his aid is doing…I love Mentors International and the huge successes they are having…

  9. I don’t think Heifer refused to experiment on people. Obviously, they *did* experiment on people… and now refuse to release results. I think the reason is simple: they don’t need to. Their marketing is so good, they raise their pile of cash and implement anything they want. Program implementers see examples of beneficiaries’ lives improving, and they take this as ‘proof’ that it works. I doubt the average program implementer has a sense of proportion of works/doesn’t work (however they define it). Moreover, doubt they consider the alternative of handing over cash.

    It’s the rules about cash transfer that still need so much work. In Afghanistan, cash from aid overwhelmed the economy and led to a lot of fraud. In our own American system, the EITC works well (bot giving cash and incentivizing work) whereas other systems do not. Obviously GiveDirectly is still learning. But I wish them well.

  10. I love this article – but you’re selling aid projects short with your “flipping the coin” metaphor. Cash transfer projects always start with careful assessments that ID the most vulnerable communities and households. It’s not a coin toss, it’s careful targeting.

    And let’s be clear that quality research comes at a cost. In my experience, programs are often willing to experiment and compare slightly different approaches. (Although we could and should be doing that more.) The tougher trade off is when when you have to decide between providing aid to more people, or dedicating a % of the budget to research.

  11. I think that some of Heifer’s discomfort is the language of “experimenting” on people. If I understand correctly, the’ just give money’ data collection is mostly surveying. But “experimentation” brings up images of working with animals or vulnerable human populations. Also, any GiveDirectly “control group” is not actually being controlled, right? Besides those people not getting the cash?

  12. @Sara: You’re correct that there’s targeting within a village. But NGOs programs seldom help more than a few hundred or few thousand. The losers of the “coin flip” are people who meet all the criteria but never get considered.

  13. Chris,

    I got a little confused by your saying that the evidence doesn’t support cash transfers yet. As can be seen from any review on cash transfers will say, this is one of the most rigorously studied subjects in development, if jot the most. it is estimated to have been applied to 750 million people. I don’t know what results will come out of your future studies, but whatever conclusion you take from them, I think you should bear in mind the extensive literature on the topic.

  14. Efficiency of the interventions is key to do more within the same budget. Moreover, it would be immoral not to help as much humanitarian victims as possible?

  15. Respect.
    I would love to conduct an RCT to find a cognitive bias related to this quote:
    ” I think humans feel like we owe something to the people we interact with, and pretty much nothing to those we do not.”

  16. Thanks for a great post Chris. I was wandering that it may be that the striking results of the Northern Uganda cash transfer evaluation may be, in part, driven by the relatively favorable economic prospects for that region related to peace dividends, relative strength of Ugandan economy and proximity of South Sudan with its insatiable appetite for Ugandan exports. People respond to economic incentives and it may be that gulu and the region now have these.

    Also this case it may actually be that indeed cash transfers may have different results, just down the road, in south Sudan, for example. In order to take advantage of cash transfer one may need to have at lest minimal skills and, south Sudan with, for example, 24% literacy rate, desperately lack these.

    Given your knowledge of South Sudan would you advocate cash transfers in that environment? Just in case you have too much time which I’m sure you do not

  17. @Tiago: There have been plenty of studies of conditional cash transfers in Latin America but less on unconditional transfers.

    @Rob: Indeed puzzling why Heifer wouldn’t point the journalists to this work.

    @Tomek: I think cash transfers will probably stimulate investment wherever people are below their potential because of good opportunities and limited access to finance. This describes young people in rural areas of any African economy that not moribund. You don’t need skills to be a petty trader, which is what some of my other evidence shows delivers high returns to some of the poorest women in Acholiland in Uganda.

  18. The question “Why didn’t you just give them the money?” can be applied to many things ( as it has been already). Any form of human transaction that takes a fee for activity. Why not give students 40,000 and a few economic textbooks and tell them to study rather than have them attend Columbia. The 40 thou goes along way towards creature comforts that would make studying easier. A randomized trial comparing attenders of class at Columbia with non attenders given 40 thou may show that the non attenders have a greater grasp of economics, and perhaps better outcomes. This is the online world in a nutshell.
    Mental health counseling in particular may fall prey to this paradigm — talk with somebody for a few sessions or get $500. The money may alleviate some anxiety.

  19. This was a good blog, and I agree with a lot of it. But I still am disturbed with some ideas of the “experimentation” i.e. deciding which people get money, which don’t, and then interviewing them and comparing results, as was explained on This American Life. I guess my analogy would be since Harvard graduate students were behind this very just cause, how would they feel if when looking for a job American industry decided that they were doing an experiment, and half of Harvard graduates would get work next year but half wouldn’t, and then we’d compare the results to see how successful jobs were. Isn’t this more or less the same “experiment?” I completely recognize that doing something to help people is important, and that you have to evaluate how successful your work is. But choosing one “poor” family over another as a random factor just strikes me as playing god. Yes, it’s better to help someone than no one. But something about it just feels wrong to me.

    Sorry if I didn’t articulate my position clearly enough. I just heard “This American Life,” read this blog, and wanted to comment on my reactions.

  20. Great post, although I am afraid someone not familiar with the topic may become a little confused with the mixing of cash transfers, experimentation, measurement, randomization, cost-effectiveness, crowding out, etc.

    Since you have at times in fact been successful at overcoming this resistance to rigorous evaluation, it would be interesting to hear about what made the difference then, and in general what you have found to be effective responses to the most common objections.

  21. The most important thing that the Planet Money segment highlighted, in my view, wasn’t the case for cash transfers per se. Rather it was the need for a culture change in the field of economic development. That culture change involves 1) actually letting recipients of aid determine how best to utilize resources, and 2) actually assessing the impact and efficiency of development efforts. Most development projects, from my vantage point, are both donor-directed and refuse to assess themselves with any scientific rigor, which I think are both reflective of a self-serving approach among a great many of those working in development. Cash transfers are reflective of this much-needed cultural change, but shouldn’t be confused with the bigger shift that needs to occur.

  22. BAGS

    A part from exporting RMG, We Develop a Decent Fashion able ladies Bag the Martial of Which are plastic and stone.
    hello there please you help me I am a poor man. thanks

  23. I am starting a new business in vegetables export business, but my family is very poor family, I want a Rs.5lckhs and me ready start the new business.i wait for the last 5 year,not ready to the money in my mistake.so please helping mind genius VIPS and TRUST organization…please help me…..

  24. Dear Sir,
    My self D.K.Verma 53 years old from Ghaziabad near Delhi staying with my wife and one daughter. Last two years back I have lost my job and earning. and since then I am leaving in very pathetic condition. I am hard working person and beg your help to survive with my family. I am requesting you since I understand that you people are so kind to help the poor.
    If you wish please help me with money or any job to do for you.
    Position have reached that I will finish our life with family. Your help is needed please assist me.
    My bank details are as under for your needful done.
    Name – Dhananjay Kumar Verma
    Bank Name – Oriental Bank Of Commerce
    Branch – Vaishali
    Account no. – 51512010025740
    IFS Code –ORBC0105151
    Contact no.-07428745315

  25. Food stamps, welfare, cash, the giro. It is all much the same.

    How do you get able bodied people off welfare.

    I think the posts would be more useful if people stopped using the general catch all description – the poor. Who are they. How are they means tested. I remember a guy saying to me that it was good that there were rich people in Europe to send things to poor people in Africa. He had 50 cattle. It is an abused term and I am not sure audits on ngo spend on people described as poor would pass, if poverty was means tested properly. Label a continent – Africa – as poor and then you avoid the hard work of means testing.

  26. In my experience the so called poor are often resource rich, but cash poor. Lacking a grip onto the global economy. In this case cash transfer is a great waybof helping them unlock the value in their pre-existing resources, instead they tend to be stripped of their resources for cash by profiteers. Consider a farmer selling his land in Cambodia to go work in a textile/clothing factory to earn cash and survive in the modern society. Some direct cash transfer is being done by Willie Uy. There should be some statistics on their recepient selection and outcomes.