Should Givewell Rank Charities Numerically?

According to Tom Ash from Giving What We Can, if GiveWell listed their top charities numerically, his donations would be better informed. To help with this, I have aggregated the donating decisions that Givewell employees described on their blog. I have also given two reasons that ranking their top charities may be less helpful than it appears.

In his recent blog post, Tom said:

"It’s difficult to pick between the top charities recommended by GiveWell and Giving What We Can, because neither group currently ranks their recommendations, and I certainly know less about the choice than their researchers do"

I agree that in complex domains, expert opinion is often useful. So consider the opinions that Givewell employees revealed by their donations. Almost all of Givewell gave to GiveDirectly:

Note that these donations were biased toward SCI, DtW and GiveDirectly because of percieved benefits of keeping one's public donations on the object level.

Givewell staff had to make decisions of their own. Many split their donations across multiple charities and described how difficult their decisions were, despite expert knowledge. So why didn't they provide a numerical guide? Well, they said that donating effectively in 2013 came down to judgement calls - weighing the strengths and weaknesses of three charities in light of one's worldview, values, and the amount of money to be donated.

This is not entirely convincing. The values that would lead one to donate to Schistosomiasis Control Initiative are very similar to the values that would lead to Deworm the World, so if Givewell could not rank one of these as superior to the other, they should at least be able to recommend an optimal split of funding between the two. The judgement calls, where possible, are the job of Givewell. And it's often not apparent to individual donors which charities are closing their funding gaps until the season is over. However, I have two more robust reasons that ranking between their top charities for 2013 was unnecessary (these are not endorsed by Givewell):

  1. Big donors can make judgement calls. Of the money that Givewell move, GoodVentures are now giving about half. Some of it is going to each of Givewell's recommended charities and also to the Against Malaria Foundation. From their experience in business, GoodVentures' Dustin Moskovitz and Cari Tuna probably do have some information on the relative importance of transparency, scalability and so on, and so should be well equipped to decide not only between GiveDirectly and deworming, but also between Schistosomiasis Control Initiative and Deworm the World. And so if Givewell's minor donors like Tom miss some of the low-hanging fruit, then GoodVentures can pick it up anyway.

  2. Effective Altruists should consider non-top charities. Charity-evaluation can take a lot of time. However, if major projects that that are not evaluated by Givewell may offer order of magnitude more impact than the top recommendations, then it is important that we investigate them. An now Holden himself says that notwithstanding limited state of evidence, high-variance, and second-thoughts over ones own sanity, blue-sky projects may be more impactful by an order of magnitude. Saving to donate next year or better - investigating some of these blue-sky projects seems better than uncritically donating to Givewell's top charities.

If I am right about 1, then it does not matter to which of Givewell's top charities Tom donates, and if I am right about 2, then it is important for Tom to cast the net farther and wider than AMF and Givewell's top charities. And if he chooses to outsource this research, then he can do so by donating to the top the top prioritisation organisations, or even to the newly formed Global Priorities Project.