Software engineer in Boston, parent, musician. Switched from earning to give to direct work in pandemic mitigation. Married to Julia Wise. Speaking for myself unless I say otherwise.
Full list of EA posts: jefftk.com/news/ea
Software engineer in Boston, parent, musician. Switched from earning to give to direct work in pandemic mitigation. Married to Julia Wise. Speaking for myself unless I say otherwise.
Full list of EA posts: jefftk.com/news/ea
I didn’t respond because of the “we won’t be able to continue engaging in the discussion on this here”. FarmKind can decide that they don’t want to prioritize this kind of community interaction, but it does make me a lot less interested in figuring out where we disagree and why.
It is also unclear how much FarmKind could change course based on feedback (other than giving up and shutting down)
There are lots of options helping animals (through raising money or otherwise) that don’t involve this kind of competition around the impact of donations. It’s common for startups to pivot if their first product doesn’t work out.
For what it’s worth, my first interpretation of “no-kill meat” is that you’re harvesting meat from animals in ways that don’t kill them. Like amputation of parts that grow back.
But if we agree it is not tenable, then we need a (much?) narrower community norm than ‘no donation matching’, such as ‘no donation matching without communication around counterfactuals’, or Open Phil / EAF needs to take significantly more flak than I think they did.
While I think a norm of “no donation matching” is where we should be, I think the best we’re likely to get is “no donation matching without donors understanding the counterfactual impact”. So while I’ve tried to argue for the former I’ve limited my criticism of campaigns to ones that don’t meet the latter.
As Michael says, there was discussion of it, but it was in a different thread and I did push back in one small place against what I saw as misleading phrasing by an EA fund manager. I don’t fully remember what I was thinking at the time, so anything else I say here is a bit speculative.
Overall, I would have preferred that OP + EA Funds had instead done a fixed-size exit grant. This would have required much less donor reasoning about how to balance OP having more funding available for other priorities vs these two EA funds having more to work with. How I feel about this situation (on which people can choose to put whatever weight they wish—just illustrating how I think about this) does depend a lot on who was driving the donation matching decision:
If the EA Funds managers proposed it, I would prefer they hadn’t.
If OP proposed it, I would prefer EA Funds had tried to convince them to give a fixed-size grant instead. If they did and OP was firm in wanting to do a matching approach, then I think it was on balance ok for EA Funds to accept as long as they communicated the situation well. And EA Funds was pretty transparent about key questions like how much money was left and whether they thought they would hit the total. I think their advice EA Funds gave on what would happen to the money otherwise wasn’t so good but I’m pretty sure I didn’t see it at the time (because I don’t remember it and apparently didn’t vote on it). The guidance OP gave, however, was quite clear.
One thing that would have made pushing back at the time tricky is not knowing who was driving the decision to do the match.
But I also think this donation matching situation is significantly less of an issue than ones aimed at the general public, like GiveWell’s, GivingMultiplier’s, and FarmKind’s. The EA community is relatively sophisticated about these issues, and looking back I see people asking the right questions and discussing them well, collaborating on trying to figure out what the actual impact was. I think a majority of people whose behavior changed here would still endorse their behavior change if they fully understood how it moved money between organizations. Whereas I think the general public is much more likely to take match claims at face value and assume they’re literally having a larger impact by the stated amount, and would not endorse their behavior change if they fully understood the effect.
I haven’t encountered any donors complaining that they were misled by donation matching offers
When I was at Google, I participated in an annual donation matching event. Each year, around giving Tuesday, groups of employees would get together to offer matching funds. I was conflicted on this but decided to participate while telling anyone who would listen that my match was a donor illusion. Several people were quite unhappy about this, where they told me that it was fraud for me to be claiming to match donations when my money was going to the same charity regardless of their actions. They were certainly not mollified when I told them that this was very common.
I think the main reasons you don’t see complaints about this from the general public is (a) most of the time people do not realize that they are being misled (b) people are used to all kinds of claims from charities that do not pass the smell test (“$43 can save a life!”) and do not expect complaining to make anything better.
no strong solid public evidence
In this particular case there is an RCT: Caviola and Greene, 2023.
I do think there are ways of doing donation matching that are more real than others, but at least in the case of EA organizations I would like to see the organization be public about this so donors can make informed tradeoffs. If I think $1 going to X is 70% as valuable as $1 going to Y, but X advertises 100% matching campaign, the details of the campaign matter if I’m going to make an informed decision on giving my dollar to X vs Y. For example, the organization could say what would happen with the money if it was not matched and whether past matches had been fully depleted.
In the case of FarmKind they are public about the details, which is great and I’m glad they do. Except that because all the money in the bonus fund is going to go to effective charities regardless of whether others donate, I think the first two considerations you give (match funders who also support less effective charities; matches sometimes don’t run out) don’t apply here.
I’m not happy about bringing this to a top-level post, but I also don’t really see other good options. I had previously raised issues with this general approach as misleading donors and the creators of FarmKind were aware of these arguments and decided to continue anyway. When Ben and I raised some of these issues in the comments on their announcement, my understanding was that FarmKind was not considering making any changes in response to our objections and considered the matter closed (ex: “That’s all the time I have to spend on this topic.”).
without the potential benefits (more money for effective charity) being considered
I’m confused why you’d say this—I mention this several times in my post as a reason for donation matching?
That’s a good question! As I said in the post, this is not my top issue in any objective sense, only what makes me most disappointed. That is, this is a place where the community had come to (what I consider) the right answer, but then backslid in response to external incentives.
I’m confused about what you mean by not being clear which cause area you support? Do you mean individuals not making their donations public? Or not disclosing how they personally prioritize cause areas?
On the question of whether organizations should have external reviews and public budgets, to me this comes down to whether organizations are raising funds from the general public. The general principle is that donors should have the information they need to make informed decisions, and if an organization is raising funds from a small group, it is sufficient to be transparent to that group, which generally has much lower costs than being transparent to the world as a whole. (I also don’t see this as a place where the EA community in general has started failing at something it used to be good at)
Hmm, then I’m still confused about why you think this is a fair way to characterize what FarmKind is doing, but not what my offer does. I’m trying to break it down by looking at a simplified case, which is why I think the hypothetical helps?
(While I did edit my comment, it was to add the final paragraph, starting with Similarly...
. I’m pretty sure the Do you think that in the simplified case it's fair for me to say...
was in there from the beginning.)
Thanks for the responses!
I’m taking your response to be saying that “part of the pot goes to your bank account” is a fair way to characterize my example offer, but if I’ve misinterpreted your response let me know?
Maybe it would be most productive to focus on this limited point:
I don’t agree that “every dollar donated to the bonus pool is one that will be paid out to effective animal charities regardless of how other donors behave” because dollars in the fund very literally get paid out to favorite charities that needn’t be super-effective or support animals.
Lets say I came to you with an offer. I have a pot of money with $100 in it. If you do nothing I’m going to donate it to ACE at the end of the day. If you’re willing to donate $50 to ACE, however, then I’m willing to give you $25 from the pot and only the remaining $75 from the pot will go to ACE.
(This offer is not intended to be fully analogous to the FarmKind situation, but to be a simplified version where your “very literally get paid out” defense still applies.)
The net flow of money if you don’t take the offer is: $100 leaves the pot, ACE gets $100. The net flow if you do take the offer is: $100 leaves the pot, $25 leaves your bank account, ACE gets $125.
Do you think that in the simplified case it’s fair for me to say that I don’t agree that “every dollar in the pot will be paid out to ACE regardless of how you behave” because $25 from the pot will very literally be transferred to your bank account?
Similarly, it seems to me that the most straightforward way to describe the flows of money if you take the offer is “The pot will transfer $100 to ACE regardless, and you will additionally transfer $25”. My original presentation of the offer, with the idea that some money goes from the pot to your bank account, is adding additional complexity that only serves to make the situation more confusing and potentially convince people to take the offer because they haven’t fully thought the situation through.
Aidan, Ben (I assume) and I aren’t commenting because we feel like we’re personally harmed by FarmKind. Instead, it’s that a core part of the EA movement is bringing clarity and transparency to charity, helping donors understand what the real impact of their donations is, and we’re concerned about an effort that seems to be going the other way.
That people are willing to donate through a platform isn’t much evidence either way on whether it’s misleading: many donation appeals are seriously misleading while bringing in large amounts of money.
Good point, thanks for flagging! Here’s what the function looks like to me (sheet):
Note that the X-axis is in terms of the fraction of the donor’s contribution actually going to the effective charity, and not in terms of the percentages displayed in the tool. For example, when you set the tool to “10%” and put in $1000 it tells you $105 will go to the effective charity and $945 will go to the regular charity. That means $55 of the donor’s initial $1000 is going to the effective charity, for a rate of 5.5%, and that is the value I use on the x-axis.
(Part of why I missed this is that the screenshot Aidan provided happens to show the only case where FarmKind and GivingMultiplier function identically)
Thanks for the response!
The bonus WILL in part go their favorite charity, as you can see when you step through the donation process … This is true of Giving Multiplier as well.
I agree that the site gives the impression that part of the bonus goes to the favorite charity, but that isn’t usefully true. I explain this in detail in my GivingMultiplier post, and summarize it as “A simpler way to describe this is that it matches whatever portion you choose to give to the effective charity at 50%, and doesn’t match your donation to the other charity at all”. [EDIT: the FarmKind function is a bit more complex here, sometimes above 50% and sometimes below, but the matched portion to the non-effective charity remains 0%] To take the screenshot case, I think it would be much clearer to describe it to donors as the donor putting $150, of which $90 goes to the favorite charity and $60 to the effective charity, and then FarmKind contributes $30 to the effective charity.
It is not the case that all bonus funds would otherwise go to help animals. We fundraise for our bonus fund from the same donor pool as our regular donors …
First, the post we’re discussing this on is a counterexample. FarmKind is fundraising from EAs, and writes “If you’re reading this post, our platform isn’t aimed at you. … you may be interested in giving to them via our bonus fund”. That is, FarmKind does not want EAs (or other people who already plan to give to help animals effectively) to give through the matching platform, but does want them to give to the bonus pool.
Second, I was trying to make a more limited claim, about what happens to the money that’s already been put in the bonus pool. I think we can agree that every dollar donated to the bonus pool is one that will be paid out to effective animal charities regardless of how other donors behave?
After reading through FarmKind’s materials, I do think my criticism is accurate. FarmKind has the same issues I previously documented with GivingMultiplier. The “bonus” is presented to users as if it (a) will go in part to their favorite charity and (b) is money that would not otherwise be going to help animals, but neither of these are true. The system is fully explained on other pages, as you say, but the typical person going through the site is donating under a false impression of their effect.
I don’t think the problem is that it’s difficult to explain things transparently without putting people off with verbosity, but instead that the core thing that makes the site work (convincing people to give more by giving the impression of greater counterfactual impact) is misleading.
I appreciate that you’re trying to help people give more effectively and improve conditions for animals, but this kind of fundraising is corrosive and does not belong in the EA movement.
I agree that this is misleading. If you are telling these new donors that this is a counterfactual match, that is another way of saying that if these new donors do not put up money than the funds in the the matching pool will not go to this good cause. Which is not the case: if there was a real danger that these matching funds would go unallocated then the announcement wouldn’t be asking us not to donate!
would you say that since switching careers, your engagement measured by these kind of metrics (books read, events attended, number of EA friends, etc) has gone up, gone down, or stayed the same?
I think it’s up, but a lot of that is pretty confounded by other things going in the community. For example, my five most-upvoted EA Forum posts are since switching careers, but several are about controversial community issues, and a lot of the recency effect goes away when looking at inflation-adjusted voting. I did attended EAG in 2023 for the first time since 2016, though, which was driven by wanting to talk to people about biosecurity.
The second cohort (‘logical ceiling’) are people who basically already run their entire lives around EA principles (and I met several at EAGx). They’ve taken the 10% pledge, they work at EA orgs, they are vegan, they attend every EA event they reasonably can, they volunteer, they are active online, etc. It’s hard to imagine how people this committed could meaningfully increase their engagement with EA.
I think ‘engagement’ can be a misleading way to think about this: you can be fully engaged, but still increase your impact by changing how you spend you efforts.
Thinking back over my personal experience, three years ago I think I would probably be counted in this “fully engaged” cohort: I was donating 50%, writing publicly about EA, co-hosting our local EA group, had volunteered for EA organizations and at EA conferences, and was pretty active on the EA Forum. But since then I’ve switched careers from earning to give to direct work in biosecurity and am now leading a team at the NAO. I think my impact is significantly higher now (ex: I would likely reject an offer to resume earning to give at 5x my previous donation level), but the change here isn’t that I’m putting more of my time into EA-motivated work, but instead that (prompted by discussion with other EAs, and downstream from EA cause prioritization work) my EA-motivated work time is going into doing different things.
Seconding this! I would also be very curious about what the multiplier is if you discount giving in future years, ideally as a chart of multiplier vs annual discount rate.