We recently published a new review of GiveDirectly, a “standout” charity that gives cash directly to poor people in Kenya. As we were going through the process of discussing and vetting the new review, I found myself wondering how I would defend my preference to donate to distribute insecticide-treated bednets (ITNs) against a serious advocate for cash transfers. We’ve written before about the theoretical appeal of giving out cash, and the fact that there is a promising charity doing so renews the question of whether we should.
I continue to worry about the potential “paternalism” of giving bednets rather than cash (i.e., the implication that donors are making decisions on behalf of recipients). I believe that by default, we should assume that recipients are best positioned to make their own decisions. However, I see a few reasons to think bednets can overcome this presumption:
- The positive externalities of ITNs
- The fact that bednets protect children rather than adults
- The fact that ITNs may be unavailable in local markets or that people may reasonably expect to be given them for free.
I address each of these reasons in more depth below. Note, however, that this discussion is meant to be primarily about the theoretical question of giving cash versus giving bednets; a more practical discussion of giving to the Against Malaria Foundation versus giving to GiveDirectly would focus on the specifics of the two organizations.
The positive externalities of ITNs
We discussed the evidence that ITNs have benefits for community members other than those using the ITNs in our review of the evidence for ITNs. After speaking with several malaria scholars and reviewing the literature, we concluded:
- The evidence for the efficacy of ITNs is based on studies of universal coverage programs, not targeted programs. In particular, all five studies relevant to the impact of ITNs on mortality involved distribution of ITNs to the community at large, not targeted coverage… Thus, there is little basis available for determining how the impact of ITNs divides between individual-level effects (protection of the person sleeping under the net, due to blockage of mosquitoes) and community-level effects (protection of everyone in communities where ITN coverage is high, due to reduction in the number of infected mosquitoes, caused either by mosquitoes’ being killed by insecticide or by mosquitoes’ becoming exhausted when they have trouble finding a host).
- The people we spoke to all believe that the community-level effect of ITNs is likely to be a significant component of their effect, though none believe that this effect has been conclusively demonstrated or well quantified.
- There is some empirical evidence suggesting that the community-level impact of ITNs is significant.
In our main model of the cost-effectiveness of distributing ITNs (XLS), we assumed that 50% of the benefits of ITNs come from the total community coverage of ITNs.
To the extent that ITNs have positive externalities, private actors may underinvest in them, meaning that it may be a good idea to distribute them freely even if individuals would choose not to purchase them at the available price. More generally, since we care about helping whole populations and not any particular specific individual, providing “public goods” of this sort amplifies our impact relative to giving the same amount of money to individuals.
Although it is conceptually possible that giving a large number of individuals small cash grants also has positive externalities, e.g. by boosting the local economy, we haven’t seen any evidence of this, and we doubt that the magnitude of the externality would be as large.
Bednets protect children rather than adults
One of the central reasons that I appreciate cash transfers is that they avoid paternalism. But sometimes, especially with regard to children, paternalism seems morally justifiable. I believe this is one of those cases.
Although AMF distributes ITNs universally, not just to children, the main benefits of ITNs—averting mortality—accrue to children under the age of 5. Children under the age of 5 lack bargaining power, income, and access to credit, not to mention the cognitive faculties to make decisions about their own long-term welfare. Accordingly, purchasing something that is reasonably likely to keep young children alive, even if they don’t or can’t decide to purchase it for themselves, seems to be a justifiable form of paternalism. In general, paternalism towards such young children is unobjectionable.
By distributing bednets, we might be spending money to benefit kids in a way that their parents wouldn’t spend it if we gave it to them instead. Given the magnitude of the benefits to the children, this seems to be justified.
People may not purchase ITNs because they are unavailable in local markets or because they expect to be given them for free
This point is more anecdotal, but Natalie, Holden and I remember being told while we were in Malawi that long-lasting insecticide-treated bednets, of the sort that AMF distributes, are essentially not available for purchase in local markets. Unfortunately this is not in our published notes (DOC) from the conversation where we recollect it occurring.
In another case, an RCT in Kenya in which researchers experimentally subsidized the cost of bednets (PDF), even very small increases in prices led to substantial reductions in bednet purchases by mothers (e.g. charging $0.60 led to a 60% reduction in take-up). Two different people told us in off-the-record conversations that they thought that this occurred because the mothers offered subsidized bednets believed that they would be able to acquire free nets at some other point. There have been periodic free ITN distributions in many sub-Saharan African countries over the last decade, and the international consensus seems to be that governments should distribute ITNs free of charge in malaria-endemic areas. Accordingly, it should not be especially surprising that citizens may expect bednets to be provided free of charge, and may not move to purchase them even if they are available at subsidized prices in the marketplace. If we reasonably expect to be given something for free in a relatively short time window, why buy it now?
This wouldn’t necessarily have been the case if philanthropy had never funded bednets, but having started down this path, I think it provides another consideration in favor of continuing. If we could credibly and cheaply communicate that no more bednets would be forthcoming, this consideration wouldn’t matter, but there is no obvious way to do so.
This is something to keep in mind in the future: philanthropic funding decisions may create an unanticipated a form of “lock-in,” in which future philanthropists become effectively committed to continued funding, even if it would not have been necessary in a counterfactual world of no philanthropic support. Although unlikely to be crucial, this consideration may counsel against certain undertaking some marginal philanthropic activities.
Conclusion
I think that in order to avoid paternalism, philanthropists working to improve the lives of the global poor should have fairly strong presumption in favor of cash transfers, and that those who advocate other strategies should have a convincing story to tell about why they beat cash. Above, I’ve tried to justify my view that bednet distributions are one of those philanthropic strategies that may beat cash. In searching for future top charities, I’d like to see a similarly strong case.
Comments
This comes from a 1st world view that assumes that money will let you buy anything because there is a mall just 15 minutes away. That may not be true in parts of Africa. They may not be able to purchase bednets easily. I would also be concerned what happens when there is rampant inflation and the cash help you gave them decreases in value overnight and conversely, what happens when the day the cash comes out is made known and theft, mugging and burglary increases on that day, because that the day everyone has a bunch of cash.
Chris: thanks for the feedback. I agree that bednets may not always be available for purchase in the developing world–that’s the first point I made in the section titled “People may not purchase ITNs because they are unavailable in local markets or because they expect to be given them for free.”
We’ve written about the potential for inflation related to cash transfers here: https://www.givewell.org/international/technical/programs/cash-transfers#Whatarethepotentialdownsidesoftheintervention. In short, the only empirical study we’ve seen on the issue did not find that cash transfers caused inflation.
Theft strikes us a legitimate concern, but the M-Pesa mobile transfer system in use by GiveDirectly seems to me to make the scenario you describe less likely. GiveDirectly is currently running an RCT in which they’re measuring both whether families are receiving the full values of their transfers and whether they report any crime or violence related to the transfers. We discuss both of these issues in more detail at our review of GiveDirectly: https://www.givewell.org/charities/give-directly
The real question should be :
do we give bednets or do we sell them at a subsidized prize ?
We favor bednets over cash, because this action is meant to avoid malaria, and that is a no brainer
We favor selling these bednets at a cheap prize so people take ownership of them and use them properly, otherwise , once sold or damaged they don’t have any use!
Just basic food for thought
Some studies find that many Africans are by our standards mentally retarded. Do you find these results credible or relevant?
Antoine: we haven’t seen any evidence indicating that selling bednets at a subsidized price leads to higher usage rates. We discuss the evidence on that issue here: https://www.givewell.org/international/technical/programs/insecticide-treated-nets#Freevs.costrecoveringdistributions
Samuel: we’re not aware of any such studies. If you found them credible, you’re welcome to send them to me at alexander@givewell.org so I can take a look. I’d be surprised if I found them credible.
Alexander, thanks for this thoughtful post. One of our main goals in creating GiveDirectly was to start more conversations like this one.
As Holden has discussed elsewhere the development field is crowded with “white knights” arguing that their approach produces good results for the poor. Sadly, few are asking the tough question: are the results good enough to justify taking money that could have been given to the poor and using it on someone else’s idea? At GiveDirectly we share your view that there should be a presumption in favor of letting the poor set priorities. This leaves space for other interventions if a convincing case can be made that they will do more good for the poor than the poor could do for themselves given the same funds.
On bednets versus cash, I personally find the externalities argument (#1) most compelling, though apparently better numbers here would help. The argument about children (#2) seems to hinge on the idea that parents care too little about their kids – I don’t know if this is true or not, but are there data that inform your view here? The argument about availability (#3) seems a bit too broad to me; after all, there are plenty of products that are not available in low-income areas for the very good reason that poor people don’t see them as a high priority and don’t want to spend precious pennies buying them.
Full disclosure: I have given to bednet charities in the past because I believed that the external benefits were large.
Is paternalism necessarily a bad thing though, particularly when you take into account behavioral implication and as you point out, the principle-agent dynamics at play with respect to children? I wrote a blog post on this recently, I think it’s worth a deeper examination: http://www.developingjen.com/blog/paternalism-friend-or-foe-
Here’s one argument, that is perhaps closer to the personal experiences of some here in the U.S.
Imagine a wealthy individual (the benefactor) who wishes to enhance the life of a young niece or nephew (the child). The recipient has a single parent, but that parent has various problems in their own life (perhaps drugs or alcohol or other issues) that have put them in a poor position economically to support their own child, and furthermore call into serious question their ability to make good choices for their child.
The child is perhaps 12 or perhaps 15 or perhaps 18 or 21. I’m not sure the age makes a huge difference. But the child is immature (not surprisingly) and lacks for good role models.
So, how should the benefactor extend support to the child?
1) Send the child a one time check for $50,000
2) Send the child ongoing payments of $500/mo for several years
3) Do either 1 or 2 by proxy, sending the money to the parent
4) Pay for something specific with durable value to the child – a suitable education at college or a trade school
===
I think most of us, given a choice, instinctively would lean towards #4. Call it paternalistic. Call it a distrust of the wisdom and intentions of either the child or the parent.
GiveDirectly sounds a bit like option 1 or 2. Direct (or nearly so) financial support to questionably competent 3rd world governments sounds a bit like option 3. The more traditional charity/aid model of attempting to find a specific need (perhaps ITNs) in the target community and addressing it seems like option 4.
“Here’s one argument, that is perhaps closer to the personal experiences of some here in the U.S.”
I guess I shouldn’t have implied that those interested in GiveWell style charitable analysis and donation are necessarily sugar daddies to their relatives and relatives’ children.
But I do hope that by bringing the analogy of how to help someone in need closer to home, that those who are fans of the GiveDirectly approach can see why I, and probably others, are more skeptical about it.
Are economies of scale not also a factor, in that a large scale distribution of a product (e.g. bednets) may be achieved at a lower cost per item than if the item were bought by individuals at market stalls, where the purchases of individuals are not coordinated so bulk distribution doesn’t happen?
I don’t have such a strong bias against funding what you want with your money. I understand it can be a bit uncomfortable if you equate doing so with paternalistic attitude to poor foreigners. I like freedom. I like economic freedom. But I don’t see much evidence for how brilliant people are in making decisions. This isn’t about thinking poor people make bad decisions. Rich, college graduates in the USA make tons of lousy decisions for their long term interests.
If I want to fund libraries rather than giving them cash that is fine with me. If I want to fund a park so kids and adults can play and get exercise I think that is fine.
I do like how you all evaluate the effectiveness of various giving strategies. If giving cash has substantially more measurable benefit than giving school nets or funding health care or education or entrepreneurship wonderful. That is a good piece of data to evaluate.
I accept from capitalism that letting people decide what to do with their cash instead of having a state 5 year plans is wise. That is far different from saying people are brilliant with deciding how to spend their cash. I believe overall people are less than than state 5 year plans, but I don’t think they effectively optimize their spending for their desires. And that isn’t about treating them like kids. It is that many people make lousy decisions with their money. Thinking giving people free rain by just giving them cash isn’t a sensible goal in my opinion.
On top of all that, many times individuals can’t buy a solution that can be put together realistically. You could fund a library. Giving out a bunch of cash and then while those people theoretically could fund a library practically it is likely unrealistic (the logistics and even vision are just not likely to come together even if it is what would give the greatest good from their collective individual interests).
On the one hand, it seems unlikely that direct cash would be spent optimally for their own benefit. On the other hand, we have a wealth of examples where our so-called experts made purchasing decisions that were clearly mistakes. Locals do tend to know better.
Before we can say anything substantive on this issue, we need more data on the effectiveness of direct cash. However, if we assume that direct cash does turn out to be effective, then I feel we should err on the side of direct cash, unless there is a compelling reason not to, such as with malarial nets, which may not be purchased at all because they are used to free nets.
A somewhat analogous case that might be helpful to think about for fleshing out intuitions is welfare in e.g. the US: unrestricted cash grants vs. restricted cash grants (with e.g. work requirements) or food stamps.
(Of course, there are multiple differences. For one, the US case is only for a portion of the population, not the median family. Secondly, the US case is more purely paternalistic, since e.g. anything you can buy with foodstamps you can buy with cash, but it might not be possible to buy to bednets. Etc., etc.)
From what I’ve read of Heather Beth Johnson’s “The American Dream and the Power of Wealth”, I suspect that wealth – in terms of nonliquid physical assets – is mentally placed in a very different category from money; wealth gives a sense of having financial security and a “safety net”, while money is viewed more in terms of what it can buy right now. Giving physical objects is giving wealth, while giving cash is giving money. I think that both likely have their merits, but that the effects of giving wealth are more likely to have a more stable, long-term effect, while giving cash will probably give a short-term boost.
Thanks for all the comments, and sorry I’ve been slow to respond.
Paul: I don’t think the argument about kids necessarily hinges on the view that their parents don’t care enough about them. I think you only need to believe that parents don’t care as much about their kids as their kids care about themselves, which seems more plausible to me.
I think you’re mostly right about the market availability issue, but that the “reasonable expectation of public provision” point remains and is fairly important.
Jenny: thanks for sharing your blog post. I think that well-documented cognitive biases are a big part of the right way to make the case for paternalism, but I have a couple additional worries:
Phil and Colin: I think you’re saying similar things, and it gives some insight into why people are more skeptical of cash transfers than I am. I think Colin’s caveats explain part of the reason that I feel differently about this case than the domestic welfare case.
Peter W: economies of scale could be a factor on the other side of the ledger, but my guess is that they play a relatively small role. On the other side, private provision might enable more competition to drive down prices.
Alexander writes:
Bednets save lives, a pretty universal value, so I don’t see a real concern about paternalism here. Putting aside efficiency issues, what do you think about charities that promote family planning or educating girls or tolerance for minorities? All of those would surely involve donor values rubbing against conventional values in various societies in a much more dramatic way that I can imagine bednets ever would; would you be highly concerned about such interventions on “paternalism” grounds? They seem to me like positive interventions notwithstanding.
Colin: I’m not necessarily especially concerned about “rubbing against conventional values.” I think the way I’d put my paternalism concern now, a year later, is that spending philanthropic money to help people in a way different than they would spend the money themselves demands justification. Sometimes that justification may take the form of a difference in values (e.g. if donors value a girls’ education more than she does), and other times it may take the form of different fact bases or inferences.
As I wrote above in reply to Jenny:
A difference of values is probably somewhat more problematic in the international development case because philanthropists aren’t democratically accountable. I don’t think that’s a remotely insurmountable objection, but I do think it’s a consideration in the other direction.
Bottom line, though, is that I think these are efficiency issues, as opposed to some kind of moral principle that should never be violated.
Comments are closed.