About a year ago, GiveWell announced they were going to be offering donation matching. I missed this at the time, but now that I see it I'm disappointed. They intend to offer a "true" match, where donors who take advantage of donation matching can trust that the matching funder would not have given otherwise. They write:

We plan to verify that the donors who provide matching funds for GiveWell campaigns would not have otherwise donated. We are taking the following steps to do so:

• Approaching donors who have shown interest in increasing GiveWell's reach.

• Asking if they would be interested in making an additional gift this year to underwrite our matching campaigns.

• Assessing their giving history and our expectation of their likely giving in 2020 so that we can see if matching funds appear additive.

• Confirming with potential matching donors that we are only interested in donations they would not have otherwise made.

• Communicating to potential matching donors that we will only accept their gift in the amount we are able to match from other supporters. This might mean asking the donor to wait to give until the matching campaign is complete so that we only receive the correct amount, or returning unused funds to the matching donor.

While this is a lot better than the more common practice of matches that are entirely illusory, it is below the standard I expect from GiveWell. Critically assessing impact is GiveWell's core strength, but on this question they are essentially taking the donor's word that it's a "true" match. GiveWell does have options if they wanted to more carefully validate these claims, such as refusing a fraction of donations and verifying that the money was not spent on anything similarly positive.

(I shared a draft version of this post with GiveWell, and they described the vetting that they currently do. While it was better than I had guessed from their website, I still don't think it's sufficient to support their strong claim of counterfactuality. I would encourage GiveWell to write publicly about the steps they take here.)

As with many donor illusions, however, I think it would probably be worse if GiveWell really did have a pool of money that would be wasted if people did not take them up on their match offer. As GiveWell explained ten years ago, that "creates incentives for [funders] to take gifts they would have made anyway, and structure them in a way that gets you to give more to the program of their choice."

I think GiveWell does a great job overall, and I really appreciate having their recommendations available when I'm deciding where to donate. I'm glad GiveWell is trying to expand its reach and move more money, and I understand how offering matching can drive donations. But I don't think donors understand how much weaker GiveWell's match vetting is than their charity vetting, and either way we should not be incentivizing setting up situations where funders leverage others into increased giving by threatening to spend their funds poorly.

(Disclosure: my wife is a GiveWell board member, but she had no input into this post.)