"Congratulations Charlie, your poverty-stricken family has just won a lifetime's supply of non-transferable chocolate!" "$%@#!"
The biggest story in the development blogosphere this week has been the 1 million t-shirts for Africa campaign and the incredibly strong reaction of the blogging community. For those new to the discussion, a full list of all the relevant posts is available here. Until now, my brain was too overloaded to manage anything but a few grunts in crayon form.
There have been various, familiar arguments against the prospect of dumping more used t-shirts into markets in poor countries. A lot of people are worried about the effect on those markets themselves, of undercutting local textile producers. I’ve made this argument myself before, but am becoming less convinced by it; I’m beginning to suspect that the damage to local producers has already been firmly done, and that other sources of cheap clothes (such as growing imports from places like China) will continue to dominated markets in the long-run. Unfortunately there hasn’t been much in terms of rigorous study on the topic, aside from an in-depth report from Oxfam, making it difficult to know what the aggregate effects are.
Let’s move on to the more important argument, which has to do with missed opportunities and meeting the needs of the recipient. This has been covered quite a bit, but I think it’s an important enough argument to be restated in a more precise way, so here we go:
Gifts in-kind are, for most recipients, strictly inferior than cash gifts of an equivalent value.
Imagine a poor person with no income. Now, give her a dollar and she will spend it on the most urgent necessity (in econospeak, the good that offers her the highest expected gain in utility). Give her a second dollar, and she may continue to spend it on that good, or she may switch to another good. As the her income rises, she will dedicate each marginal dollar to what matters most to her at that moment. If you give her another dollar and she buys a mango, that doesn’t necessarily mean she couldn’t afford the mango before, but that a mango currently offers her the greatest marginal benefit after spending her other resources on other, possibly more important things.
This is an extremely rational model of expenditure prioritization. Of course, reality is much more complex – some would argue that it’s not very easy for people to identify what they need or may ignore their own best-interests (although I don’t think the latter assumption gets us very far in life). If you can stomach the constraints of my simple model for a moment, consider the following thought experiment:
You have a population of people of different incomes and different priorities. Some of them have shirts, others don’t. You can easily identify these people and you decide to buy/obtain shirts for those lacking them (at a cost to you of $1).
What is the welfare impact of your intervention compared with just handing out dollars? For a select few – those that were poised to spend their next available dollar on a t-shirt, the welfare impact is equivalent. For everyone else that is shirtless, the welfare impact of giving a shirt is strictly less than that handing out a dollar.
Why is this the case? If you hand someone a dollar, they spend it on the good that represents the highest possible welfare benefit. So if they spend it on something other than a t-shirt, it means that the t-shirt wasn’t the best possible purchase for them*. So, if you hand them a t-shirt instead of a dollar, there is an implicit cost to that missed opportunity. The harder it is for the recipient to exchange that t-shirt for what they want, the bigger the missed opportunity.
The resulting rule of thumb is: “if someone doesn’t have a shirt, there are probably a lot of other things that they don’t have, and we have no good reason to give the shirt priority.” The important thing to take away is that the same holds for all in-kind gifts.
Of course, there are lots of other reasons why we may not want to just go around giving cash, but there are lots of alternatives to running around dispensing used goods. On a micro-scale the gold standard of cash-transfers may be impossible to beat with in-kind gifts, but interventions that move beyond the individual, such as investment in public goods or removing kinks in the system that create poverty traps, can result in larger welfare gains.
—————————-
*I am, in this simple model, mostly ignoring budget dynamics. If your income was 90 cents a day, you will buy whatever bundle of goods that is best suited for you at the prevailing prices. It could be that you really, really want a shirt, more than anything in the world, but they cost $1, and when your income goes up to $1 a day, you only spend your money on shirts, and would have loved it if someone had given you a shirt when you were still making 90 cents a day. However, these people are no easier to identify, and so giving each person an extra dollar in income is still strictly superior to handing out $1 t-shirts, because those that had always wanted a t-shirt will just go on to buy the t-shirt.
#1 by Michael Keizer on April 30, 2010 - 12:32 pm
Oooohhh, you stole that pun from me!
Otherwise a great post.
#2 by Matt on April 30, 2010 - 12:35 pm
Michael – oh no- and I’ve read your post! When writing it I felt like it would have been used before, but searching for unkindness and in-kind on google didn’t bring up anything I recognized.
#3 by Michael Keizer on April 30, 2010 - 12:40 pm
No worries, Matt, those things happen. And to be honest, I could only wish that you stole it from me — you know, theft and flattery and all that.
#4 by Will Seitz` on April 30, 2010 - 1:17 pm
I like how you put all of this. In the case of shipping shirts to the poor people of ‘over there,’ I have no reservation in agreeing with every point. But I would raise a question or two about more realistic situations that do not involve t-shirts.
Distributing money or some highly liquid substitute can be a really bad idea in the wrong context. Your argument assumes that the sole purpose of aid is to increase the utility of the individual from their point of view. In disaster relief however, an individual’s judgment and behavior may be severely impaired by uncertainty and anxiety. It makes sense to put limits on using aid money for alcohol for example, and it is sometimes a good idea to give blankets to all, so that the allocation of such crucial resources is not determined by a market in conditions of scarcity. This is not a small problem by any means, and the many groups (the Red Cross comes to mind) go to great lengths to ensure that people are not just trying to “maximize their utility,” even when they give out cash.
Individual utility might also ignore, say, the benefit of a microscope to detect diseases in contaminated water. Or perhaps the value of a doctor to use the microscope. The benefits of these services are dispersed throughout a group, but if everyone has an incentive to maximize their individual utility, buying supplies and filling other needs more often than not, these priorities fall by the wayside. I do not think it is a bad idea to donate the non-fungible services of a doctor armed with a microscope, mostly because I do not trust anyone (including myself) in a crisis situation to navel-gaze on economic problems of public goods.
Also, this ignores the potential to salvage sunk costs employed in some other area of the economy. In-kind donations and the like are a really good idea if the costs of disposal are positive, yet utility gains, net the impact on local production and transport costs, is positive when ownership is transferred. If you could send the machinery for an otherwise obsolete factory to Malawi for example, there is not necessarily much loss on the part of the sender, and the increased utility may (in some situations) be larger than the cost of getting everything there. Perhaps that was an overly elaborate example but the point is this: If there are sunk costs that cannot be recovered by a firm or individual, but somebody else would gain utility if allowed access, then the only real question is transferring ownership. As long as there are few and low costs on the part of the person or organization donating, we can get closer to Pareto efficiency when there is a donation.
You are right in your example of course, and I for one loved your crayon critique. We can get awfully nerdy on some of this stuff if we complicate the issue but I would love to hear your thoughts.
#5 by Matt on April 30, 2010 - 2:13 pm
Will – thanks for the comment (and for engaging with me on a particularly nerdy post). Here are some thoughts:
*On how appropriate liquid gifts are: I agree with your points here, although I’m a little suspicious about putting limits on what donations can be used for (this sort of ties in to the larger discussion about fungibility which was all the rage last week), mainly because it’s pretty difficult to enforce (if you say don’t spend this money on alcohol, or give someone a voucher, maybe they’ll take the in-kind gift and sell it, or, more likely, they’ll just compensate by shifting relative expenditure). There are, I think, other ways around this, by giving donations to family members who are more likely to “use the funds responsibly” (that’s code for women) – I think there’s enough evidence to support this.
In general, I should say that I’m actually pretty wary of cash-transfers, mainly because free cash in the long run doesn’t create good incentives. Tyler Cowen from Marginal Revolution has always maintained that the welfare-maximizing, least-distorting way of giving cash was to travel to random places in poor countries and give a load of cash to someone who looks busy, then never ever come back again.
*Your second point about whether or not individual utility is a correct measure is also true – although the examples you give are more-or-less public goods examples, which I think are typically superior to direct cash-transfers. I think my main argument was that, if you’ve decided to intervene at the individual level (as opposed to funding or providing public goods or other services), then there’s a case for cash over stuff. Once you move outside of that individual (or household) level intervention, then we have to be much more careful on deciding what makes sense to give.
*On your final point on sunk costs – again yes, there is a case to be made here, but two possible problems: the biggest is the transfer cost, which is one of main criticisms of most in-kind aid, which is generally very costly to ship (t-shirts, maybe less so, but food, soap, machinery, etc). So while there might still be utility gains from shipping there stuff (and, just to be clear, I do think there are utility gains), I think that other interventions, ones involving transferring cash when could then be used for some purpose in the recipient country (it doesn’t have to be used for an individual cash-transfer), start to look more appealing when transport costs are high.
Sometimes the presence or absence of markets for things gives us an indication as to what scope there is for pareto improvements. There are lots of viable second-hand markets, even internationally (as anyone who’s bought a used Japanese car in East Africa can tell you). When there’s no international market for a given used product, that’s usually a good sign that the transportation costs are too high. That actually isn’t the case with clothing – a lot of companies buy up used clothing from charities and then sell it in developing countries. With the machinery example – ideally we would expect a middle-man to recognize that obsolete capital in the developed country would make higher returns in developing countries and arrange the transfer (and possibly pre-finance the purchase)- markets often don’t work that well. The fact that no one is doing that might mean that there really aren’t any gains to be had (or, more likely, that markets don’t function well enough for that sort of arbitrage). I’m waffling now so I’ll stop!
#6 by Sam Gardner on April 30, 2010 - 2:34 pm
I am fully supportive in general, but in reality a lot of things might happen that make the argument just wrong.
Imagine the t-shirts are bought by you for 1 U$ (or the cost of transport), and through all kinds of effects, the cost on the local market is 10 US$. Not very probable for t-shirts, but for a lot of products a reasonable scenario.
In that case, for the one getting the t-shirt, it is the value of a t-shirt (10 US$) and for the others, the resale value (10 US$). in all cases, better than the 1dollar handout.
Of course, the real problem here is market inefficiencies, but the one going naked might want to get the first delivery while someone is solving the institutional problems behind this market inefficiency. Also, you t-shirt just becomes a cash gift in disguise, but so be it in the real world.
#7 by Nicolas Köhler on May 4, 2010 - 10:28 pm
“Unfortunately there hasn’t been much in terms of rigorous study on the topic, aside from an in-depth report from Oxfam, making it difficult to know what the aggregate effects are.”
Please note that there have been at least two peer-reviewed studies on the aggregate effects of used clothing exports to developing countries.
Most recently: Frazer, G. (2008). Used-clothing donations and apparel production in Africa. The Economic Journal, 118(October), 1764-1784.
“This article examines the importance of one possible explanation for the failure of African countries to step onto the bottom rung of the manufacturing sophistication ladder, that is to produce apparel. Used-clothing donations to thrift shops and other organisations in industrialised countries typically end up being sold to consumers in Africa. Since used clothing is initially provided as a donation, it shares characteristics with food aid, which always assists consumers, but at times harms African food producers. Used-clothing imports are found to have a negative impact on apparel production in Africa, explaining roughly 40% of the decline in production and 50% of the decline in employment over the period 1981–2000.”
And a little older: Haggblade, S. (1990). The flip side of fashion: Used clothing exports to the third world. Journal of Development Studies, 26(3), 505-521.
“Since the early 1970s, western countries have increasingly recycled their second-hand clothes by exporting to the Third World. In response, some LDC governments have banned used clothing importstoprotecttheirdomestictextileindustries.Thisarticle, after reviewing the structure and evolution of world used clothing trade, examines the consequences for Africa’s largest importer, Rwanda. In this low-income country, with its small domestic textile industry, displacement is minimal and imported used clothing offers a modest but rare policy leverfor directly increasing nationalincome as well as incomesof the rural poor.”
I hope this will contribute to the debate.
#8 by Matt on May 5, 2010 - 6:36 pm
Nicholas – thanks for the links, I’ll check them out.
#9 by Leah on May 11, 2010 - 10:27 pm
The other thing to wonder besides ‘why give a t-shirt to someone who’s starving?’ is ‘who is treating these t-shirts?’ Unless these t-shirts are going to be fumigated by people who really know what they’re doing, (slightly dubious, purely judging on website) then the potential for spreading smallpox, the flu, chicken pox, measles, mumps, etc., is pretty great. If there’s one thing Africa doesn’t need more of right now, it’s a giant surge in disease rates.
#10 by Michael Keizer on May 12, 2010 - 3:21 am
Smallpox? Really?