GiveWell: a case study in effective altruism, part 5

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 third through fifth arguments, on influence, access, and independence. In this post, I consider the sixth and seventh arguments:

Argument 6: If no one else is willing to fund a program, then this is evidence that the program should not be funded. Crowding out other donors destroys this source of independent validation.

Argument 7: If Good Ventures fully funds every high-value giving opportunity it finds, this could lead to other donors preemptively abandoning programs the Open Philanthropy Project is looking into, thus substantially reducing the amount of effective giving in the Open Philanthropy Project's perceived current and potential focus areas.

Argument 6 is sometimes an important consideration, but is a poor fit for the GiveWell top charities, to the extent that most donors are already largely moved by GiveWell's recommendations. Argument 7 points to a real problem, and one that reflects poorly on the effective altruism movement, but the problem is not mainly concentrated in the area of funding.

Argument 6: Funders are evidence of effectiveness

The winner's curse is the tendency for the winner of an auction to have overbid. Consider a simplified example where all the bidders have the same preferences and resources, but different information about the item being auctioned. If their information has not been selected through a biased process, and they interpret it correctly, the one willing to bid the highest will be the one overestimating the item's value the most.

For the same reason, if you are a program's only donor, this is evidence that you are overestimating its value. The willingness of other parties to fund a program is evidence of its effectiveness, and crowding them out destroys that information. But this trades off against information scarcity.

For this reason, Good Ventures might reasonably want to avoid crowding out other donors, for programs where it is not exceptionally certain of its judgment, and has no reason to believe that its preferences uniquely favor a program like this. On the other hand, in cases where it is very confident in the value of the program, there is little reason not to crowd out other donors - and substantial reason to want to crowd out small donors, to free up their attention to evaluate smaller or less certain programs.

If other donors have already funded the program, then the confidence threshold is even lower - Good Ventures should feel comfortable crowding out other donors as long as it is confident in its ability to detect a substantial change in a program's value.

Another explanation for being a program's sole supporter is that you have very unusual preferences (or unusual ability to evaluate evidence). If this is the case, there are unlikely to be other donors to crowd out, so the issue is moot.

While these considerations do suggest that a major funder should sometimes avoid crowding out others, GiveWell's top charities seem like a poor fit for this reasoning, since they seem unusually easy to evaluate relative to other Good Ventures grants, recommended by the Open Philanthropy Project. In addition, the presence of other donors adds unusually little information in this case, because they are largely motivated by GiveWell's own recommendations.

Argument 7: Preemptive exit

GiveWell worries that if Good Ventures fully funded giving opportunities, this would lead other potential donors to avoid causes and grants they're interested in:

We do not want to be in the habit of – or gain a reputation for – recommending that Good Ventures fill the entire funding gap of every strong giving opportunity we see. In the long run, we feel this would create incentives for other donors to avoid the causes and grants we’re interested in; this, in turn, could lead to a much lower-than-optimal amount of total donor interest in the things we find most promising.

I've heard two versions of this concern. First, an after-the-fact crowding out effect - if other donors notice that their donations mainly serve to offset Good Ventures's money, then they may withdraw. Second, other donors may preemptively withdraw if they think a charity is likely to become a Good Ventures grantee. Preemptive and after the fact crowding out are different, and I think it's worth addressing the prospect of preemptive exit separately.

If you are trying to cause the most good to be done, a preemptive exit seems like it will always or almost always be counterproductive; early-stage funding is how you leverage later-stage funders. For example, if you expect that a program will motivate Good Ventures to fully fund it once it's been evaluated, this is a strong reason to support it all the way up to that point.

This seems to me like a real problem, because people are ignoring the fact that the Open Philanthropy Project is capacity constrained. I have heard people say "that's on the Open Philanthropy Project's to-do list" as a way of implying "effective altruists have already got this covered," when in fact it means the opposite. The Open Philanthropy Project has a very long to-do list. If something is sitting on their to-do list, that means that someone working at the Open Philanthropy Project believes that the thing ought to be looked into, but no one has actually done so, because they are busy with higher priorities.

This is a form of preemptive crowding out of attention. Attention, unlike money, is obviously a scarce resource on current margins for the Open Philanthropy Project. There is no good reason for effective altruists to do this. Let's do better.

The obvious response to this by the Open Philanthropy Project and Good Ventures would be to treat disclosure, not as a value or obligation, but as a strategic tool for coordination. A recent Open Philanthropy Project blog post indicates that they're moving towards a less comprehensive and more strategic attitude towards transparency, but only in general terms. Here are some specific tactical and strategic reasons for the two organizations to disclose information:

  • They are in the process of thinking something through, or rely on a particular model or fact claim, and would benefit from criticism if their current thinking can be improved.
  • They are about to make a funding decision in some area, or about to publish research on something, and other people about to consider related investments might want to focus on other things for now, to avoid unnecessary duplication of effort.
  • They have learned something that other effective altruists might find helpful.
  • They have decided not to pursue an investigation further, because they do not think the program area is promising.
  • They have decided not to pursue an investigation soon, because they are poorly positioned to evaluate it or higher priority tasks have come up, so others considering such an investigation should not worry about duplication of effort.
  • They have decided not to fund organizations in an otherwise promising area, because they do not think that the kind of funding they are well positioned to provide is the limiting factor, but would encourage EAs with different kinds of resources to contribute (e.g. smaller funders or people who might be interested in starting or joining an organization) to consider it.

Other effective altruists can help by continuing to be willing to look into things even if they're on the Open Philanthropy Project's to-do list, and not being discouraged by the idea that the Project's bigger and more capable than you. They really can't do everything worth doing. You likely have something to offer.

Conclusions from Arguments 6 and 7

The most obvious and harmful type of crowding out happening in effective altruism is that of attention, not money. Effective altruists bear substantial responsibility here for exaggerating GiveWell's and the Open Philanthropy Project's capacity to evaluate programs. (I've written about this previously in the case of GiveWell.) This blog post series is in part an attempt to alleviate that problem by evaluating them as actually existing specific institutions, and contextualizing them by contrasting them with potential alternatives.

The Open Philanthropy Project is compounding this problem by trying to do a little bit of everything. Clearer communication around what it is not doing, and where it has found important tractable problems where funding is not the bottleneck, might somewhat alleviate this.

One thought on “GiveWell: a case study in effective altruism, part 5

  1. Pingback: GiveWell and partial funding | Compass Rose

Leave a Reply

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