The GiveWell Blog

Philanthropy vouchers

We focus on finding charities that are doing demonstrably good work already, rather than on proposals for new sorts of projects. This post is an exception: we’ve been tossing around an idea for “philanthropy vouchers” that we think could be worth trying in a broad variety of contexts, and we’re interested in others’ thoughts.

The idea is a variation of the “development vouchers” idea put forth by William Easterly in The White Man’s Burden (see page 330). Prof. Easterly proposes that official aid agencies co-create an independent “voucher fund,” and issue vouchers to people in developing countries that can be redeemed for money from the fund. The basic appeal of the idea is that, like cash handouts, it may shift the power and choice to the hands of the people we’re trying to help, rather than the hands of well-meaning outsiders at charities; but the two major concerns with cash handouts (fraud/manipulation by less poor locals and poor/irresponsible use of the money) could be mitigated by some basic regulations on what sorts of services the vouchers can be spent on.

While Prof. Easterly proposes a coordinated effort by major aid agencies, our proposal can be carried out at very small scale, unilaterally, by a single funder. The funder would simply issue a set amount in vouchers, set its own rules for how they could be redeemed, and set aside the necessary funds.

Specifically, to carry out a philanthropy vouchers program, a funder would do the following:

  1. Determine how much “money” it wanted to inject into a community in the form of vouchers.
  2. Form a definition of a “philanthropic organization,” i.e., an organization that would be eligible for collecting these vouchers from people and trading them to the funder for cash. This classification could be formed in a variety of ways: the funder might lay out a set of general criteria for “philanthropic” organizations and take applications for formal designation as “philanthropic,” with approved organizations’ getting the right to trade vouchers to the funder for cash; or it might do something as simple as accepting vouchers from any organization classified as a charity in its country of origin.
  3. Print vouchers and distribute them to the people in an area (trying to target those in need, but the targeting wouldn’t be as high-stakes as it is with cash).
  4. From there, any organization classified as “philanthropic” could offer its goods and services, and all such organizations would effectively be competing for the funds embedded in the vouchers.
  5. The funder would still be well advised to do its own monitoring and evaluation of how the program is going – in particular, spot-interviewing participants to ensure that vouchers were obtained through transparent and mutually consensual transactions

For a hypothetical example, consider an “alternative Millennium Villages” powered by philanthropy vouchers.

  • The funder would create a definition of “philanthropic organization” as any US-registered public charity, or local government agency, whose activities in the village consisted of providing or “selling” the following: vitamin and mineral supplements, health services, water, primary education, food meeting basic nutrition standards, or electricity. Organizations would apply to the funder for recognition as such an organization, a process that need not be nearly as involved as applying for direct funding. Organizations with other ideas for helping people, such as cellphones, could apply as well, and their status would be at the funder’s discretion.
  • The funder would print 5,000 vouchers for $50 each, and distribute them throughout a village of 5,000 with a rough goal of allocating one voucher per person (or N vouchers per family of N). (Assuming $50,000 in funder overhead, this would be equivalent in cost to Millennium Villages). Alternatively, the funder might allocate some of the vouchers to a “common fund” allocated through a voting procedure among villagers, in order to encourage the purchase of “public goods” such as well construction (though of course the villagers could also arrange such a “common fund” themselves, or simply choose to “pay” ala carte for water).
  • Nonprofits and government agencies would then hopefully offer services in attempts to win clients’ vouchers. If a nonprofit perceived that others were focusing excessively on farmer training as opposed to water, it could invest in providing water and hope to take in more revenue in vouchers than its costs.
  • With each voucher submitted to the funder, an organization might submit a brief description of what was provided in exchange for the voucher, and to whom; the funder would then perform “spot interviews” to see if these descriptions were confirmed by villagers.

Though the example given is for the developing world, I think the concept could as easily be used in poor communities in the U.S.

There would be many challenges involved in such a program. Tensions could arise between different “competing” organizations, and they may resort to misleading advertising or even coercion in order to win more vouchers. Vouchers wouldn’t be distributed perfectly fairly or evenly among participants. However, these issues could be monitored to some degree using spot interviews, and the concerns would be smaller than with a cash handout program. On the flip side, voucher revenues would provide strong indicators of which services people valued most and how that changed over time, and the actual services provided could adjust in real-time to these indicators. Incentives and possibilities for innovation and adaptation would likely be much greater than for a centrally planned project.

All in all, it seems to us like a project along these lines would be worth trying, hopefully accompanied (as with any pilot) by strong monitoring and evaluation. What do you think?


  • Jonah S. on August 6, 2010 at 12:28 pm said:

    Thanks for thinking this idea through. It sounds potentially worthwhile. Two questions that occur to me are

    •With such a small scale project, would there really be sufficient incentive for “philanthropic organizations” to develop the infrastructure to accept and process vouchers?

    •I understand how this idea partially mitigates the concern of “poor/irresponsible use of the money,” but I don’t see how this idea does anything to address the concern of “fraud/manipulation by less poor locals.” In this regard, how is distributing $50 vouchers with a goal of allocating one per individual different from giving away cash with a goal of allocating $50 per individual? Is the idea that the people who have their basic necessities met are less likely to try to aggressively seek extra such vouchers than they would be to seek extra cash?

  • J. S. Greenfield on August 9, 2010 at 6:32 pm said:

    Honestly, this proposal seems ill-constructed. What you have described do not appear to me to be “philanthropy vouchers” at all. They are personal-benefit goods and services vouchers, with strings attached that are almost certain to make them highly inefficient, without achieving much or anything in the way of accountability benefits.

    In the model you describe, the vouchers are basically coupons for purchasing goods or services, but the “stores” where one can use them are significantly limited. This limitation will inherently reduce efficiency, and therefore, reduce the aggregate good/benefit generated. For example, private, for profit entities often provide goods and services better and cheaper than

    governmental or non-profit entities do. By making for-profit entities ineligible as recipients, you both reduce competition (which inherently increases prices), and eliminate what are likely to be low cost options.

    It’s likely that the authorized governmental and non-profit entities will frequently subcontract out to for-profit entities to provide goods and services people want. In this case, the governmental or non-profit entity adds an expensive layer of bureaucracy.

    The question then becomes, do you get something valuable in exchange for that layer of bureaucracy?

    I believe the answer is very unlikely to be affirmative. Presumably what you hope to get is significantly improved accountability and efficiency in directing funds to good uses. But I fail to see how anything in the proposal guarantees, or even suggests, that result. In fact, I would expect that, as a large-scale program, a program like this, without extensive, high-cost oversight, is likely to be no less subject to abuse than a program that just hands out $50 in cash to recipients. In fact, it may be more subject to abuse, because of the fact that individuals don’t have cash that they can use for any purpose.

    There’s every reason to expect that corrupt entities and corrupt individuals within entities will appear to provide voucher-holders the opportunity to exchange their voucher for (something significantly less than $50 in) cash, that they can use for any unrestricted purpose they choose. And there’s certainly no intrinsic mechanism created to ensure that the vouchers are directed only to the poorest or otherwise neediest recipients.

    So I think at best, you have pre-selected some charitable domains, and merely allowed the local population to choose among them (so it’s not clear to me that you’re really getting a significant benefit of on-the-ground insight). At worst, you’ve recreated cash distribution programs, in an even less efficient fashion.

    If the fundamental idea is to leverage the knowledge, insight and wisdom of the local population, as a substitute for outsider evaluation of the efficacy of charitable alternatives, that would seem an interesting thought, but I don’t believe the proposed approach achieves that.

    If you want to leverage that local insight, then it seems to me that what you’d want to do is truly make the vouchers philanthropy vouchers. i.e., the local individual recipients — let’s call them delegates — can give their voucher to a charitable or governmental institution, for any purpose they choose, whatsoever, EXCEPT as an exchange for goods or services or for any other direct, personal benefit. Then you are actually delegating some of your philanthropic responsibilities to the local individuals.

    In this model, a delegate may choose to give to entities that will subsequently use the proceeds in ways that benefit the delegate personally (indirectly, as a result of putting their voucher donation to work), but the decision is, at least, not biased by the inducement of a direct quid-pro-quo for personal benefit. Accordingly, in this model, it seems much more likely that a delegate’s choice will take into account the aggregate good/benefit an entity is likely to create using the voucher. And if the charities are legitimate, then a high percentage of the benefit should go to those who most need it.

    Of course, this still wouldn’t eliminate the need for oversight, because there would still be strong incentives for corruption. But your immediate oversight becomes focused purely on detecting and preventing corrupt inducement, as opposed to attempting to directly evaluate the efficacy of the charitable ventures.

    Still, I don’t have a high level of confidence that this variation would produce efficient and effective results, either. I suspect that just as most educated donors have difficulty evaluating charitable efficiency and effectiveness, poorly educated local delegates aren’t likely to do much better, despite their on-the-ground perspective.

  • David Henderson on August 10, 2010 at 1:50 pm said:

    Interesting idea, but the reality is that social services, at least in the US, are provided on an as funded basis. What this means is that funders provide the startup capital to provide a service (staff, supplies, etc.). Your voucher model assumes all non-profit participants would already be providing said service, this does not really reflect the way social programs are funded and implemented.

  • Holden on August 11, 2010 at 12:44 pm said:

    David, the proposal does not require that nonprofits already be providing said service. If the program were done with enough size, philanthropy vouchers could create enough demand and upside to make it reasonable for nonprofits to invest in startup costs and take the associated risks. In practice, at a smaller/experimental phase it may be necessary to take other proactive measures to get participants (such as subsidizing startup costs). But I believe this idea has promise as a way to get new services to the underserved, not just a way to make existing service providers more accountable.

    The idea this proposal is closest to is cash transfers. When people have money, organizations willingly take on risk and startup costs in order to serve them.

  • Holden on August 27, 2010 at 7:18 am said:

    Jonah, responses to your questions:

    1. Possibly not. As I mentioned in my response to David, if the project is done at a small scale it may be necessary to subsidize startup costs for something so experimental.
    2. Vouchers are less fungible and flexible than cash. Therefore, the incentives to stockpile them – especially for those in less need – are reduced (not eliminated). (In an extreme case – vouchers that can be used only for tuition – incentives to stockpile are very low. As the options go up, the incentives to stockpile go up, but they never reach what they are for cash.)


    • I don’t understand your proposal to use vouchers that can’t be used for “direct, personal benefit.” The idea here is to replace/improve current modes of distributing things like food, water, agricultural supplies, tuition, etc., all of which are currently often distributed by nonprofits with no feedback mechanism from recipients that is nearly as direct. What sort of services would you rather see people purchase with their vouchers?
    • Your comment mostly seems to be comparing the vouchers idea to cash transfers (unfavorably). But the proposal is mostly intended to replace/change services currently provided entirely by nonprofits, and therefore subject to all the issues (lack of for-profit participants; bureaucracy) you describe. The advantage of this program over cash transfers is that it may reduce the likelihood of vouchers’ being used on unhelpful things like alcohol, and also of vouchers’ being stockpiled/stolen (since they are less valuable to a random person than cash).
    • A couple of details: the “nonprofit” criterion is just one example of a criterion the funder could use – it isn’t central to the idea. For-profits providing appropriate services could participate. And enforcement is an issue, but I think that spot interviews regarding what was purchased could go a long way toward addressing this.
  • J. S. Greenfield on August 28, 2010 at 11:13 am said:

    Holden, I wouldn’t call what you’re describing “philanthropy vouchers.” To my view, the term “philanthropy voucher” implies giving somebody a voucher to engage in philanthropic activities: delegation of philanthropic power. When Warren Buffet gives each of his kids a billion dollars to give away, that’s a (really big) philanthropy voucher.

    Because of the name, and because a few comments within your post, I thought your idea was that local individuals (many, most or all needy, themselves) would effectively prioritize where philanthropic dollars were directed, by choosing the non-profits to receive their dollars.

    That, I think, is an interesting idea. Not sure that it would actually improve outcomes, but it’s at least an interesting idea to think about.

    However, I expect any such benefit is undermined if the vouchers are effectively used as coupons to purchase goods or service for personal benefit. Then I think it is similar to cash (and we can argue over the pros and cons in terms of efficiency at achieving intended purposes).

    What you are describing sounds to me like a broader version of food stamps. In practice, I don’t think food stamps have proven terribly effective at preventing people from using welfare grants for unintended/prohibited goods, like tobacco and alcohol. So I’m a bit skeptical that they would prove more successful in this context. Perhaps they can be marginally more successful than food stamps, if the volume of vouchers is large relative to the local economy, but I wouldn’t be incredibly optimistic that they wouldn’t become tradable for any purpose, at some discount that people are willing to give up, to get what they desire.

  • Holden on August 31, 2010 at 3:19 am said:

    The analogy with food stamps is relevant, but not exact. The biggest difference is that our proposal calls for a single funder, without the practical and political restrictions of government, to set the ground rules. A smaller number of “vendors” with a higher level of accountability would be possible than with food stamps.

Comments are closed.