Originally written 2017-11-24; crossposted here after discussion on GiveWell Donation Matching
Sometimes people will describe a donation as "counterfactually valid" or just "counterfactual". For example, you might offer to donate a counterfactual dollar for every push-up your team does. [1] The high-level interpretation is that you're doing something you wouldn't have done otherwise.
What does "wouldn't have done otherwise" mean?
- If you hire a mason to repoint your wall it's not something that would have just done on their own.
- If you donate to a charity matching drive, the matching funds were very likely going to the charity regardless.
The former is fully counterfactually valid (you caused impact) while the latter isn't counterfactually valid at all (the impact of the matching funds was unchanged by your donation).
Say I offer to make a counterfactual donation of $50 to the Against Malaria Foundation (AMF) if you do a thing; which of the following are ok for me to do if you don't?
- Donate $50 to the AMF.
- Donate $49 to the AMF.
- Donate $50 to the AMF tomorrow.
- Donate $50 to another long-lasting insecticide treated anti-malaria net distribution charity.
- Donate $50 to another of GiveWell's top charities.
- Donate $50 to another group that is commonly supported by EAs.
- Donate an extra $50 to the AMF next year.
- Donate an extra $50 to the AMF next year, not because of intentional dishonesty, but just because not having given $50 this year I happen to have more money available next year when it comes time for me to figure out how much to donate and at that point I still think the AMF is a good choice.
- Spend an extra $50 on myself (go out to eat when I wouldn't otherwise, etc).
- Light a $50 bill on fire. [2]
The first example is exactly what counterfactual doesn't mean here: I'm just going ahead and doing my half of the deal whether you do your half or not. The last example is pretty clearly counterfactual. Which of the ones in between are ok?
I would draw the line as allowing only the last two. The goal of clarifying that something is counterfactual is to allow the other person to reason as if they're causing the thing to happen. On the other hand, maybe that's an unreasonably high barrier, and if we decide that's what "counterfactual" means no one will be able to use the term for anything, so we should adopt something weaker?
[1] This is philosphy-inspired EA-jargon, and as jargon I'm mixed on it but I think it's helpful to think about what we've been using it to mean and what it should mean.
[2] We could add a final one here, something like I donate $50 to a malaria promotion organization, but that's extortion. (For some reason this is commonly referred to as 'blackmail', even though it doesn't involve threats to reveal information.)
I'm probably a bit unusual in this regard, but I have different budgets for different things, so a counterfactual donation means spending $50 from my personal luxuries budget on a donation to that charity, which is in addition to the 10% of my net income that I donate otherwise. That keeps everything simple.
If you spend your personal luxuries budget in full every year, this sounds like #9, and I agree it's fine to call it counterfactual.
A minor variant on 9) which is still perhaps worth making explicit would be if you donated the $50 to a different charity that the other person did not think was very valuable. I think this maintains counterfactual validity if it is credible.
I really like how you've laid this out 😀
Personally, I think that there's a spectrum with many more points between #8 and #9. Even many employer matching programs aren't entirely counterfactual, they likely have a budget of how much they're willing to spend on charity matching which would be adjusted down on a per employee basis if they all used it, and the counterfactual portion is the difference in impact between different charities it might be donated to.
Per my comment I think there are different uses of counterfactual that are getting tied up, particularly when it comes to donor matching: impact and actions:
In the case of #2 it is not misleading to donor A to say that their donation was matched IMHO. But it isn't the full story for impact, that is only as counterfactual as the difference between the impact of the actions that are taken or not.
I think this misses out on an important question, which is "What would you have done with the money if you hadn't offered the counterfactual donation?"
If you were planning to donate to AMF, but then realised that you could make me do X by commiting to burn the money if I don't do X, I think that's not ok, in two senses:
In fact, to make nitpicky distinctions... If I didn't do X, I feel reluctant to say that it's "not ok" for you to donate to AMF. I want to say that it is ok for you to donate to AMF at that point, but that doing so is strong evidence that you were behaving dishonestly when initially promising a counterfactual donation, and that said offering was not ok.
I think I agree that only the last two should qualify, but presently I would assume a weaker definition is most common.
I suppose this can create a bad incentive where someone offering a counterfactual donation then has to make sure to do something not charitable with that money later on? I guess in my view a ‘counterfactual donation’ really only ever makes sense when you have a strong prior the money would not otherwise be put to similar use.