Give Directly

Here is Tyler summarizing the principles of charitable giving:

1. Cash is often the best form of aid.

2. Give to those who are not expecting it, and,

3. Don’t require the recipients to do anything costly to get the money.

Loyal readers will recall that Tyler followed through on his advice by sending money to random people in India, as suggested by MR readers (see also here).

A new charity is formalizing Tyler’s system and reducing the transaction cost of efficient donation. GiveDirectly takes donations over the web, locates poor households in Kenya using people on the ground, and then transfers money directly to the recipient’s cell phone (even very poor households typically have cell phones but GiveDirectly provides SIM cards for those who do not.) Transactions costs are low, just 10%.

You will not be surprised to learn that the CEO and founders and are all economists (one MBA/MPA). All the founders also have extensive experience in development. A randomized control trial is under way to evaluate the program.

Transfers, following point #3, are unconditional. The founders write:

GiveDirectly intentionally provides unconditional, rather than conditional, cash transfers. We do this for three reasons. First, empowering the poor to make their own decisions advances our core value of respect. Second, it lets recipients purchase the things they need most, enhancing impact. Third, imposing conditions on the use of funds requires that costly monitoring and enforcement structures be put in place. One detailed estimate put the administrative costs of a conditional cash transfer scheme at 63% of the transfers made over the first three years of the program (Caldes & Maluccio 2005).

Points one and three are excellent. The second point is a bit disengenous, yes it lets recipients purchase they things they need but it also lets recipients purchase alcohol, cigarettes and prostitutes. Even in poor countries, a substantial amount of poverty is caused by poor choices. Still, there is no special reason to think that cash grants will increase the proportion of money spent on “bad goods.” Cash grants could even reduce bad-goods spending. Some people drink to escape depressing circumstances, for example, so if you make things less depressing, drinking can fall. Moreover, even if you give people housing, health care, or food (stamps!) it’s not so easy to get around bad-goods spending because money is fungible.  Thus, I have no problem with donating cash.

Indeed, Tyler and I wish to encourage experimentation in charity and we have therefore made a donation to GiveDirectly.

AddendumGivewell, our favorite charity evaluator, says GiveDirectly is too new to evaluate but they like the idea and they note that GiveDirectly has been unusually forthright in providing them with advance plans on evaluation.


Hey, maybe they really needed a prostitute.

Seriously, interesting concept. Seems less likely to create dependence.

No rich people have ever came up with this idea before! Money solves everything, they only need a little. I'll just mail some over there. Don't travel there, Tyler! You might get dirty!

It's like he's recreating the 20th century aid to Africa and Latin America on an individual level. Because obviously, he'd never pay taxes to a gov't doing the same thing. The government has a deficit, after all and we mustn't collect money for collective action on charity.

I don't know, maybe he should go to people with experience. People who have spent their lives helping the poor. Ask them why they bothered to establish hospitals and churches and schools and teach and heal and distribute banned material in China and Africa.

Maybe Tyler isn't an expert in everything just because he speedreads and is loved by anti-government American economists. Maybe, I don't know, maybe his perspective is an impractical rich person perspective. Just a shot in the dark there, based on a few hundreds posts.

Been reading this blog for years on my RSS reader and rarely visit the site anymore.

So when exactly did trolls start showing up on MR?

A proven way to reduce the cash spent on alcohol, cigarettes, and prostitutes is to give the money to women, who - it has been shown - spend most of it on the needs of the family.

I would be careful about saying that. I remember there was one study that came to that conclusion, but since then there have been a number of studies questioning it.

It isn't very difficult for an alcoholic, philandering husband to seize cash given to his wife in an otherwise patriarchal society.

The main benefit of microfinance (according to Munger) is that it makes money inaccessible, aiding the ability to save (even at negative rates of return).

Transaction costs are important, otherwise point number 3, regarding governance costs:

"...imposing conditions on the use of funds requires that costly monitoring and enforcement structures be put in place."

would not have been raised. The wording of this statement brings to mind both Barzel and Williamson.

How useful is it, if the poor person your giving money to uses the money to pay off the local warlord to keep his family from being slaughtered? Wouldn't the first big hurdle to overcoming poverty is a strong rule of law? If you are going to support random charity like this program then you should also support nation building.

Seems like it'd be pretty damn useful, at least from the point of view of the potential slaughteree.

But how sustainable is it; it just incentivizes more extortion by the warlord.

locates poor households in Kenya using people on the ground

alarm bells!

very difficult to survive even if the gods live in kenya get aids

Tyler's approach does not really make sense to me. When you transfer cash to someone, they will (presumably) spend that money locally. You are not giving actual productive assets, but increasing that person's claim on local goods and services. If you pick a person at random, who is to say that person will put those local resources to better use than the next best alternative? Tyler is trying to be clever by solving some of the agency issues related to charitable giving, but seems to forget the main point. When you give cash, you are merely re-allocating local resources. That only helps if you re-allocate those resources to someone who can put them to more productive use. Choosing someone at random seems to defeat the whole point.

The initial recipient may spend the money locally, but presumably some people in the village trade with the big cities. Eventually the money from the donation should find its way into their hands, and they will use it to purchase goods from outside the village, resulting in additional resources flowing into the village.

Yes, thanks rpl for clarifying a commonly made but flawed argument.

they need to:
1. get the details of many thousands of poor households
2. use computerized random allocation within that set

it's just plumb crazy to give individuals "on the ground" the power to allocate cash payouts. Must ensure expected N is large so payout is low, so for example it's not worth buying multiple mobile phones for multiple chances of winning, or bribing whomever is collating list of households. Actually, if the average sum of money received is large enough to make any sort of difference, membership of the set of households receiving money will be an extremely valuable asset - I don't see how you can make that work without costly measures to ensure places are allocatd fairly, not being bought. Which defeats the object somewhat (but maybe only a little).

Why do you suggest the chance of getting a payment is dependent on the number of mobile phones a person owns?

From the website:

"We first select regions of Kenya with high poverty rates using census data. We then identify villages with access to an agent providing mobile-phone-based payment services. Finally, we identify the poorest households in these villages using simple, transparent criteria: we target all households living in homes made out of mud, wood, and grass... We record eligible households' phone numbers or, for those that do not have a phone, provide them with a SIM card. We follow up initial identification with a rigorous process of audits to prevent mistakes or fraud."

If anything, this has to be more transparent than most aid methods because the selection criteria is really simple (so doing an audit is really simple too). What exactly is your concern about having "people on the ground"? What aid methods exist in undeveloped areas that don't have these?

my concern was about the potential for corruption. Yes, all aid methods have people on the ground, but if the people on the ground are effectivley giving out cash (or lottery tickets) the scope for corruption looks especially high to me. You have to recruit a bunch of people on the ground who aren't going to abuse that power.

the reason I thought having multiple mobile phone would matter is that I assume the system would notice if the same mobile number is entered twice, and I assumed that corruption would take the form of the person on the ground creating a fictious household with that mobile phone number attached to it - so if you want multiple lottery tickets, you need multiple phones. They are relying on somebody going into a village and coming back with a list of households (made of mud, wood and grass) with a mobile phone number corresponding to each household

but, they do promise "a rigorous process of audits" so hopefully my concerns are unfounded.

n.b. I don't think I'm being unfair worrying about this: my view is that people in developing countries aren't idiots, and if a system of cash handouts can be gamed, some people are going to work out how to game it.

Well, I embrace some of the concepts here, but I think that there are issues being overlooked; some of these are value issues, some are unmeasured variables.

My recommendation is somewhat different, but first, I whole-heartedly embrace point 3. I also embrace 'give directly.'

Let me explain some value issues: First, I believe that there is value not only in a met material need (ie. the happiness of a full stomach, the virtue of an education, etc), but also in the exercise of generosity/caritas that occurs in the giver, and the exercise of gratitute that occurs in the receiver, and the bonds of good will that are generated between them. I also believe that the greater equality that it generates is a fifth distinct good. As a result, not only the results of the transfer matter - the means and context also matter.

Second, I believe that every study to date demonstrates that people operate on a relatively small spatial and social scale. Poverty is related not only to the inability to meet basic biological needs (hunger, shelter) but also to meet basic social needs (companionship, stable associations, freedom of action in the home). Poverty is largely experienced in the context of close association: a millionaire among billionaires feels poor as an extreme case. So, given that basic biological needs are met, in general, poverty will decline most precipitously if wealth is distributed evenly at local levels and unevenly at the global scale. Some of this is bound up in the phrase 'cost of living.' When everybody has less, one can feel equally well off with less.

Now, how do I translate these observations and values into policy recommendations?

First, giving from the richest on one continent to the poorest on another through a professional charity in a predictable way is pretty much the worst solution. It generates the lowest amount of personal contact, all that contact is through people who create entitlement. The least poverty is eradicated and it threatens the social system in both places because the scales are readily mismatched. Neither the giver nor the receiver really understands the context of the other.

The best solution is that in local arenas, people assist each other freely. People on the block should understand the needs of each other, communicate needs, and fill needs. They should do so directly; in a way that isn't hardened or atrophied into a routine or obligation; and this should increase equality, generosity, gratitude, and good will on that block. At limnal borders, people may be receiving in one context and giving in another. That's great; it is like a cascading chemical or physical reaction - the most good will and equality is generated from the transfer by doing it step-wise rather than all at once, between people who don't have much contact.

From block to block, neighborhood to neighborhoood, town to town, there should be lower volumes and less regular giving, but in a way that decreases gross inequality. At each heirarchical level, the transfers will be less regular, of lower volume, and probably more formalized. It is a necessary friction in the system, unavoidable.

I would like to see how Tyler responds to this proposal, informed as much by biology, chemistry and physics as by economics; and certainly informed by psychology, much as is his.

How about looking at the loss of money in the US or whatever home country the giver sends money from? The assumption seems to be that giving money to people you don't know who will spend it in unpredictable ways is somehow better than spending it locally, for goods that earn your money by being better. Why not just tear the money up or burn it? The effect on our economy would be the same, a negative. But, oh, that's right, then the giver would not be able to feel superior and full of "caritas". And those who buy into such a scheme probably also feel that developed economies need to be deprived of money to the maximum extent possible. Bilge.

The money will come back, one way or another. If not, everyone's purchasing power in the US will be increased.

Is there much empirical evidence that this: "People drink to escape depressing circumstances," is true?

Those that I have known with drinking problems did not seem to be in depressing circumstances (their heavy drinking some times put those around them into depressing circumstances).

This explains the binge drinking so common in frat houses. Can you imagine the horror -- 12 hours of class to attend per week, no real responsibilities or cares, can't shake a stick without hitting an 18-22 year old blonde. No wonder they drink to escape.

Yeah, damn. I can't believe the U.S. government didn't throw me hard-earned taxpayer money during my frat boy days in college. We were but poor victims of U.S. international hegemony and were drinking to escape the rigors of our lives.

Like the author concludes, I'm glad to see experimentation with giving. However, disagree that providing sporadic donations to random people is the best way to go about this. Providing a steady stream of small donations would help the recipient have more confidence in investing in productive assets that required accumulated savings to purchase. However, if these handouts are intermittent, and unreliable, isn't it much more likely that it will be used for smaller, consumable purchases? If that's the case, then the effectiveness of such a program would be severely undermined.

There are many inefficiencies in giving aid, so we have to be careful to minimize all of them, not just a subset. For instance, I expect this system to produce "aid finders" (the people who designate households to receive money) who require recipients to kick back a portion of the aid (in cash or by using other means). If this type of aid flourishes, we should try to measure this inefficiency.

>Transactions costs are low, just 10%.

The word "low" does not mean what you think it means.

The guys who cash your check and take 10% are called vicious monsters.
The people who offer mortgages at 6% are called inhuman predators.

But if you take 10% for handing money to a Kenyan, this is "low," eh?

It's considerably more difficult to get money to a Kenyan.

I can get money to Nigeria very cheaply.

What TallDave said. I'd be thrilled to drop a $10 bill in a slot and find out a Kenyan got $9 of it.

It's a sad commentary on the inefficiency of contemporary charities that we have come to the point where a 10% overhead is considered good.

I guess you always have the option to start your own organization at 9%

"People drink to escape depressing circumstances."
In most other cultures, from east to west, moderate drinking is considered to be a fulfilling individual or social experience. Idem for sex and smoking. All three habits are for sure more dignifying than moderate overeating of tasteless food or moderate watching of reality TV shows. File under the culture that is America.

Like most economists they miss the point entirely. Evolutionarily (as in why it evolved and sticks around, not why people think they are doing it), the point of giving to foreign charity, is to create dependence and harm the foreigner. Its a way for the first world to maintain its power over everyone else.

You say that a great deal of poverty is caused by poor choices, and you cite an anecdotal article by Nick Kristof? Really?

Do read the Kristof piece. It’s an excellent intro and, contra your dismissal, he provides data and citations.

“Two M.I.T. economists, Abhijit Banerjee and Esther Duflo, found that the world’s poor typically spend about 2 percent of their income educating their children, and often larger percentages on alcohol and tobacco: 4 percent in rural Papua New Guinea, 6 percent in Indonesia, 8 percent in Mexico. The indigent also spend significant sums on soft drinks, prostitution and extravagant festivals.”

Here is the link

A lot more could be said but that was not the point of my post.

Hi Alex,

Thanks for the reply. I already had read the piece - yes he cites Banerjee & Duflo - but there's some selective citing here. The very next sentence from the same paper: "though in Guatemala, Nicaragua, and Peru, no more than 1 percent of the budget gets spent on these goods (possibly because they prefer other intoxicants)". Furthermore, Kristof is clearly cherry-picking his results: if you actually go to B&D's tables, you'll see that 5/10 of the rural surveys of under $1 households they looked at (and 8/10 of the urban ones) show higher expenditure on education than alcohol and tobacco.

I don't think it is particularly controversial to suggest that poor decisions keep some people in poverty, but there's more nuanced evidence out there than budget shares (which, in themselves, tell us next to nothing about whether or not households are in self-created poverty traps).

But I'm harping on one small point in your post - which, on the whole, I completely agreed with.

Playing God and increasing entropy on the cheap.
Do you enclose a "Gift from ..." card???

I disagree with the notion that "overhead costs" are always a sign of a bad charity. Planning is important, and there are costs to figuring out the most effective type of investment, and ensuring it is not mis-used, etc. The random recipient may not be in the best position to know this. There is also a strong case for the provision of public goods that many individuals in a village could benefit from, and individual families may not have an incentive to give to.

This scheme is terrifying. What is to prevent thousands of dollars from going to Mukesh Ambani, for example? I'm sure he has lots of "plans for helping India," including building overseas mansions that will employ no Indian labor.

I do not think these cash transfers will result in a sustainable path to prosperity. Higher incomes and increasing wealth in a poor country usually take off when productivity increases.
I have not seen a lot of evidence that cash transfers increase the productivity of recipient nations, despite the fact that countries like the Philippines get more than 12bn$ of the stuff yearly, and most African nations receive a relatively big share (>10% of GDP) of remittances.
The downside of this money flow is that it props up the local currencies and hence deprives local farmers (& industries) of export opportunities for their commodities and thereby reducing their trading incentives and income.
Unfortunately, this is an "unseen" consequence and therefore most economists choose to ignore it. Once in a while it pops up, when a stream of cash dries up and the currency drops to a less artificially inflated level (see Uganda over the last weeks).

However, it ought to be a piece of cake for Mr. Tabarok to provide the evidence and to prove me wrong.

Comments for this post are closed