GiveWell: a case study in effective altruism, part 3

This is part of a series of blog posts examining seven arguments I laid out for limiting Good Ventures funding to the GiveWell top charities. My prior post considered the first argument, that the Open Philanthropy Project, and thus Good Ventures, has superior judgment to that of GiveWell donors. In this post, I consider the second argument:

Even if Good Ventures isn't special, it should expect that some of its favorite giving opportunities will be ones that others can't recognize as good ideas, due to different judgment, expertise, and values. If the Open Philanthropy Project does not expect to be able to persuade other future donors, but would be able to persuade Good Ventures, then these opportunities will only be funded in the future if Good Ventures holds onto its money for long enough. Thus, while Good Ventures may currently have a lower opportunity cost than individual GiveWell donors, this will quickly change if it commits to fully funding the GiveWell top charities.

This post is my most direct response to GiveWell's blog post explaining the reasoning behind its "splitting" recommendation.

Argument 2: Bargaining power

In its blog post on giving now vs later, GiveWell discusses potential policies it might have recommended to Good Ventures on funding the GiveWell top charities' funding gap. Good Ventures and individual GiveWell donors may have very different opinions on what else their money should be spent on, but still agree that the optimal allocation of resources should prioritize the GiveWell top charities.

Without holding the view that Good Ventures currently has a higher opportunity cost than individual GiveWell donors, GiveWell might still believe that committing to fully funding the GiveWell top charities' funding gaps would be a mistake on the part of Good Ventures. GiveWell might believe that this commitment would be bad because it cedes all of Good Ventures's bargaining power to other GiveWell donors.

GiveWell begins with a principled argument, asking whether Good Ventures should respond to each additional dollar given by other GiveWell donors by giving less ("funging"), more ("matching"), or the same amount ("splitting"). GiveWell recommends splitting, and in the first major section, I explore the principled case for this, assuming the conditions of symmetry laid out above. I argue that the principled case for splitting is only coherent under very pessimistic assumptions about effective altruists' ability to cooperate with one another. These assumptions may be justified, but as far as I can tell, haven't been seriously tested.

GiveWell goes on to make a specific recommendation that Good Ventures's "fair share" of the GiveWell top charities is 50% of the top charities' total room for more funding. In the second major section of this post, I see whether this recommendation seems intuitively fair, trying a couple of different simple back-of-the-envelope quantitative comparisons. I argue that the most intuitive relative allocation assigns substantially more of the funding burden to Good Ventures at present.

Funging and the prisoner's dilemma

In a comment on the 2015 top charities post, GiveWell co-founder Holden Karnofsky points out that there may be important differences in preferences about next-best options, between Good Ventures and other GiveWell donors:

I think it is a major problem if individuals are effectively supporting GV’s portfolio as a whole. This portfolio does include, and will continue to include, many gifts that are controversial and highly debatable, such that many of the donors I’ve spoken to are uncomfortable with them. By contrast, our top charities work focuses on finding opportunities whose case is largely concrete and verifiable. Donors who wish to effectively support the work of the Open Philanthropy Project can make unrestricted donations to GiveWell, but we don’t want this to be the default for people supporting top charities.

Holden framed this as a matter of upholding the trust placed in GiveWell by individual donors who would agree to support GiveWell top charities, but not to support many cases of Open Philanthropy Project giving. But I think this is better expressed as a dilemma, as implied by the post on giving now vs later:

"Funging" [...] would have the advantage of fully funding top charities, while not spending more (in the short run) than necessary to do so. However, it would have the disadvantage of creating a long-term incentive for individuals to stop supporting our top charities, since the only effect of their giving (in this scenario) would be to reduce the amount we recommend to Good Ventures. Most individuals would probably not notice this issue unprompted, but it’s very important to us to be open with our audience about the pros and cons of taking our recommendations, and we don’t want our offering to be valuable/attractive only to people who misunderstand it.

This looks to me like it's making the following argument:

Suppose Good Ventures commits to fully fill GiveWell top charities' funding gap. GiveWell donors with preferences that diverge from those of Good Ventures will respond to this in one of two ways:

  1. They notice that they have all the bargaining power and withdraw funding, spending it on something else that they consider more valuable than Good Ventures's opportunity cost.
  2. They do not notice that their donations are being fully offset, and keep giving, under the misapprehension that the marginal effect of their giving is to fill the funding gap of GiveWell's top charities.

The problem with funging in case 2, according to GiveWell, is that it takes advantage of implicitly misleading donors. The problem with case 1 is that it may actually lead to a much worse allocation of funds, according to Good Ventures's preferences, than is potentially attainable with a tougher bargaining position.

On this model, in the long run committing to close the funding gap of GiveWell top charities means committing to fully funding them, because other donors will notice that their marginal impact is zero and stop giving. If Good Ventures fully funds all available easy-to-communicate giving opportunities such as the GiveWell top charities, GiveWell donors will switch their giving to what they see as their next-best opportunities, causing Good Ventures to spend down its endowment quickly. The Open Philanthropy Project is building up expertise in its focus areas, and might reasonably expect that in the future it will find some valuable giving opportunities which would be part of an optimal charity portfolio, the value of which is substantially harder to communicate than that of GiveWell top charities. This weighs in favor of some amount of caution.

This is closely related to the epistemic prisoner's dilemma:

Let us say you are a doctor, and you are dealing with a malaria epidemic in your village. You are faced with two problems. First, you have no access to the drugs needed for treatment. Second, you are one of two doctors in the village, and the two of you cannot agree on the nature of the disease itself. You, having carefully tested many patients, being a highly skilled, well-educated diagnostician, have proven to yourself that the disease in question is malaria. Of this you are >99% certain. Yet your colleague, the blinkered fool, insists that you are dealing with an outbreak of bird flu, and to this he assigns >99% certainty.

[...] Dr. House calls you both into his office and tells you that he knows, with certainty, which disease is afflicting the villagers. As confident as you both are in your own diagnoses, you are even more confident in House's abilities. House, however, will not tell you his diagnosis until you've played a game with him. He's going to put you in one room and your colleague in another. He's going to offer you a choice between 5,000 units of malaria medication, and 10,000 units of bird-flu medication. At the same time, he's going to offer your colleague a choice between 5,000 units of bird-flu meds, and 10,000 units of malaria meds. [...]

You know the disease in question is malaria. The bird-flu drugs are literally worthless to you, and the malaria drugs will save lives. [...] So you take the 5,000 units of malaria medication, your colleague takes the 5,000 units of bird-flu meds (reasoning in precisely the same way), and you have 5,000 units of useful drugs with which to fight the outbreak.

Had you each taken that which you supposed to be worthless, you'd be guaranteed 10,000 units. [...]

Thus I propose that the epistemic prisoner's dilemma, though it has unique features (the agents differ epistemically, not preferentially) should be treated by rational agents (or agents so boundedly rational that they can still have persistent disagreements) in the same way as the vanilla prisoner's dilemma.

The analogy here is as follows: Good Ventures and other GiveWell donors agree that the GiveWell top charities should be fully funded. Both parties also have other things they want funded, but do not agree on which other uses for the money are best. Good Ventures can "cooperate" by fully funding GiveWell top charities now, but if other donors "defect" by switching funding to something else, then at some future time, their priorities get fully funded in the future and Good Ventures's priorities don't.

If you accept the prisoner's dilemma analogy, this leads to the following formulation of the funging dilemma:

Suppose Good Ventures cooperates on the prisoner's dilemma. GiveWell donors with preferences that diverge from those of Good Ventures will respond to this in one of two ways:

  1. They defect on the prisoner's dilemma.
  2. They do not notice that defecting is the correct action, and instead cooperate on the prisoner's dilemma.

More plainly: the only reason to cooperate on the prisoner's dilemma is in the hope that the other side cooperates. But cooperating is for suckers. Therefore it is unethical to cooperate, because that only pays off by taking advantage of suckers.

While I admire GiveWell's concern for not taking unfair advantage of others' misconceptions, it is not inherently unethical or exploitative to cooperate on the prisoner's dilemma, in the hope that the other party will cooperate too.

Of course, the first horn of the dilemma is still bad, and Good Ventures has a legitimate institution interest in not giving away all its bargaining power, but it seems to me like a bad idea to assume the worst-case scenario here. Other donors might continue to give anyway, not out of a misplaced naivete, but tacitly extending trust the way my mom and Charlie did. I've written about this kind of norm in very different circumstances:

[A] friend [...] mentioned that I was a good host - that I got some important things right (this on its own made me feel recognized) - and expressed some worry that I might feel unappreciated. [...]

I responded to the effect that [...] I'm only doing things if they feel worth doing with no expectation of reciprocation. (My other motives are finding it intrinsically motivating to do good for others, and empowering allies to do good things more generally.) But, I continued, I was sad that I'm not setting off a success spiral where other people perceive the public goods I'm contributing to as benefiting them, and try to reciprocate by generating more public goods of a similar kind. A sort of public goods success spiral, where people do more, not just because they have more resources & like people, but because they perceive themselves to live in an environment with prosocial norms.

My friend responded that my behavior had inspired them to be a better host, and given them affordances of things to do, that wouldn’t have occurred to them otherwise. They gave a concrete example: offering people water when they come in. And this was exactly the thing I'd wanted to happen [...].

In one possible equilibrium people expect that everyone will try and accomplish common goals whenever they have a high-leverage chance to do so, and trust that this will increase everyone else's eagerness to do the same. In another, people expect that contributing into a common pool of resources will cause others to feel that they can shirk more, so they have to manage their resources carefully and evaluate the effects of each action on their negotiating position. In the second equilibrium, common goods will more often be neglected. If there's a chance to push effective philanthropy into a common goods success spiral, this would be a very good thing. The obvious resource Good Ventures appears to have in excess of its ability to make use of it at this point is money, so a cooperative strategy would be to use that advantage, in the hope that this will inspire other donors with more time to spend per dollar available to donate to contribute their comparative advantage.

To do so would send a powerful message to other GiveWell donors: "We don't want to play games with people's lives, so we're going to make sure these charities get funded, for as long as we don't have better ways to do good with the money. We trust you to do the same, for as long as you don't have better options. If your giving to the charities we agree on means more money in our pocket, we'll use our honest best judgment to do the most good we can with it - we trust you'll be doing the same. The guarantor we can offer that we have shared values, is that we agree that these charities ought to be fully funded, and are acting on that belief."

While that sort of move isn't guaranteed to work, it seems to me to be worth trying. This is an iterated game, not a one-off. And the right strategy in an iterated prisoner's dilemma is usually tit for tat.

Why assume everyone else is eager to shirk? Why not find out whether the other players are willing to cooperate? Of course, it's a good idea to actually check this by surveying past GiveWell donors to ask what they're doing and why - and paying attention to changes in donor behavior. If they really do seem to be shirking, then we live in the world where this kind of cooperation isn't available.

Cost-benefit analyses are appropriate for some very big decisions (such as which program areas to focus on), but it's not feasible to trace out all the consequences, and most decisions can't be made that way anyway - most decisions are strongly conditioned by norms, so most of the effect of your acts is the norms your acts strengthen. Because of this, exploring prosocial norms is an extremely valuable activity, even when success is far from guaranteed and there are plausible arguments that, when taken as a one-off, some less cooperative move is more advantageous.

Splitting and fair shares

Common good success spirals might need to be worked up to gradually, and the communication difficulties and time lags involved might mean that you should target a sustainable equilibrium directly rather than playing an iterated game where you try your hardest on the first round. Even so, it does not seem to me like Good Ventures's fair share is 50%, as GiveWell proposes:

A figure of 50% seems reasonable for the split between (a) one major, “anchor” donor who has substantial resources and great conviction in a giving opportunity; (b) all other donors combined.

One way to check the 50% estimate is to think about what it implies other individual fair share is.

The $10 billion Good Ventures is drawing from, even if very amortized over 20 years (consistent with giving away 5% per year)1, amounts to around $500 million per year. GiveWell estimated Good Ventures's fair share of funding the GiveWell top charities at about $50 million, about 10% of that annual figure. This suggests that a typical donor with as much information as Good Ventures should be allocating 10% of their charity budget to the GiveWell top charities, to pay their fair share just like Good Ventures is doing. However, Good Ventures has a lot more information about and influence over GiveWell's processes than GiveWell does; the typical donor should trust GiveWell less, since they don't get to work in the GiveWell office, and don't get to ask the GiveWell co-founders and other staff questions in person every workday. So 10% is an upper bound.

This means that if a GiveWell donor is earning $80,000 per year, and has taken the Giving What We Can pledge, thus promising to give $8,000 to effective charities, they should give $800 of that to the GiveWell top charities on fair share considerations, and give the remainder wherever they think does the most good on their values. My guess is that by this reasoning, other donors are already giving substantially more than their fair share.

The other way to check the 50% estimate is to try and figure out, based on other donors' behavior, what Good Ventures's fair share would be.

A good approximation for other donors' capacity to give is their actual record of giving before crowding out by Good Ventures becomes a live possibility. The capacity of current private GiveWell donors during prime giving season appeared to be around $15 million, based on behavior prior to the recommendation to Good Ventures.2 Meanwhile, the $10 billion Good Ventures is drawing from, even if very amortized over 20 years (consistent with giving away 5% per year)3, amounts to around $500 million per year. In practice it looks like GiveWell moved about $40 million in non-Good Ventures money in 2015, implying that Good Ventures's fair share of the funding gap is something like 93%. 2015 was an unusually good publicity year for GiveWell, and I think it's reasonable to assume that 2016 numbers will not be a similarly large positive surprise, so Good Ventures's "fair share" is likely to be lower, but still well in excess of 80% of the funding gap. I wouldn't be surprised if within a few years other donors further pick up the slack, and I think it's reasonable to allocate giving amounts based on some projected growth in donations by other parties.

This is a very rough estimation - you could also try surveying donors to find out their total charity budgets, and what share of that went to GiveWell top charities. This method of responsibility allocation may lead to overfunding top charities' funding gaps. This is unlikely to be a major problem in the long run, as this will likely just lead to less room for more funding in the next year, and therefore a smaller funding commitment. This could also be used to establish a cap, such that Good Ventures commits to funding the remaining gap or this amount, whichever is less.

Conclusions from Argument 2

If we assume that the situation between the Good Ventures and other GiveWell donors is symmetrical, then the plausible scenario in which a hard negotiating stance like "splitting" is worth leaving lives on the table, is one in which other donors will shirk as soon as possible. This scenario is plausible, but not obviously true. If it is not true, a more cooperative strategy becomes appealing. It seems likely that such a strategy should be tried.

In the circumstance where Good Ventures should commit to paying its "fair share" and no more, its fair share is likely substantially in excess of 50%, and should be calculated more explicitly.


1, 3 GiveWell: "Since the Open Philanthropy Project is currently at an early stage, we are setting the budget at 5% of total available capital"
2 GiveWell at the end of 2015: "We estimate that non-Good Ventures donors will give approximately $15 million between now and January 31, 2016."

3 thoughts on “GiveWell: a case study in effective altruism, part 3

  1. Pingback: GiveWell: a case study in effective altruism, part 4 | Compass Rose

  2. Pingback: GiveWell: a case study in effective altruism, part 1 | Compass Rose

  3. Pingback: GiveWell and partial funding | Compass Rose

Leave a Reply

Your email address will not be published. Required fields are marked *