Should effective altruism have a norm against donating to employers?
A number of people who work at effective altruist organisations also donate to their employer. We’ve had some interesting conversations recently at CEA on whether this is a good thing. We haven’t reached a consensus, but we thought the ideas were worth sharing. This is a somewhat niche topic, and is mostly relevant for people doing direct work at such organisations.
The strongest reason to have a community norm against doing this is that it’s easier to be unbiased in assessing other organisations.
Reasons why we may be biased towards believing orgs we work for are effective:
Helps us feel better about working there
Helps us convince others it’s worth giving money to or coming to work for
Extra reason we may be biased towards donating to orgs we work for:
Likely to be a popular action at office / with employer
May be hard to eliminate such bias, so better to remove the pressure for or damage from the bias (Christiano, 2016)
Norms against donating to employer help with both
Reduces pressure for bias by removing one incentive towards believing it’s effective
Reduces impact of bias by having people’s actions depend less on biased beliefs
Reduce “echo chamber” effects where everyone talking to each other all agree about what’s most effective.
The strongest reason not to have such a community norm is that it pushes against the norm of just doing what’s most effective.
This is a distinctive and important part of effective altruism.
People are particularly likely to think their employer is the most effective use of money, since a belief that it’s extremely effective could be a common cause of choosing to work there and wanting to donate.
People often have comparative advantage in being donors for their employer:
On the heuristic that people should often just fill funding gaps that they are well-placed to evaluate (and can tell it’s a good use of money), it should be fairly common for people to be best-placed to make judgements about their employer.
This applies particularly for small organisations without much of a track record.
There are often tax advantages in donating-by-drawing-a-smaller-salary.
We should note that this isn’t a binary distinction. We want to know where we should be on a spectrum:
It should be illegal to donate to one’s employer
EA orgs should not accept donations from employees
There should be a strong norm in EA of not giving to employers
There should be a weak norm in EA of not giving to employers (for example, not unless you’ve really done your homework, or not except for orgs without a track record, or not giving more than 50% of donations to employers)
There should be no norm in EA regarding this
There should be a positive norm in EA of giving to employers
EAs should be strongly encouraged or required to give to their employers
I don’t think anyone (involved in the discussion so far) supports a. or g., and perhaps not b. or f.
There are some further considerations in each direction.
Additional reasons to support stronger norms against donating to employers:
Possible conflicts of interest
Employers may be less willing to fire or discipline employees who donate significant amounts, or may give their opinions undue weight
Employee/donors may feel like they should have more say over organisational strategy than is appropriate, and may lobby to have their views implemented
Outside view suggests that a norm of donating lots of salary back to employers is weird and worth pushing against. Carl Shulman pointed out to me some specific possible harms:
Filters out the less dedicated
Filters out alternative perspectives about efficacy
Reduces efforts finding new good opportunities
Makes wages misleading
Creates the opportunity for abuse of power by the organization/employer
Creates the appearance of all of the above, even if they are not directly damaging
Probably we don’t lose too much efficiency by redirecting these donations
If we’re functioning at all reasonably as a community, the marginal value of funds at the top organisations is actually comparable (market efficiency among the small community—opportunities which are significantly better will be taken)
Of course we’ll have controversy over which are in this set … but should expect to find some other places which are in-expectation comparable, so not leaving too much value on the floor
This doesn’t push particularly against donating to an employer, but means that if there are significant considerations against, they shouldn’t get overruled by a consideration saying “Must donate to X, since X is way more effective than alternatives”
At most EA orgs, employees are making a significant de facto donation to their employer anyway, by accepting well below salary they could earn elsewhere
This reduces the relative importance of the direct effects of their actual donations, relative to other factors.
We might donation-swap to recover the tax benefits without vulnerability to bias, although there is extra hassle associated with this.
More donor-donee links across the EA community help keep it more of a community, and also mean that people are more actively keeping up with what’s happening elsewhere.
Additional reasons to oppose such norms:
They push against norms of allowing personal freedom in how to allocate resources.
EA tends to not want to dictate to people what they may or may not do. This is a serious mark against the stronger versions of the norms.
They would be somewhat weird.
We don’t think these norms exist elsewhere. There are costs to having extra and unusual bits of culture.
There’s a tradition of for example activists donating money to their parties.
Donating to an employer is an opportunity to signal confidence in the project by using your own money
Signalling is less strong than when trying to get personal gain via investment
This cuts both ways a little, since it might also increase pressure for bias
Donating to an employer by reducing salary is low-hassle
And (like payroll giving) it can also make donating easier by avoiding loss aversion.
Thanks to Max Dalton, Sam Deere, Will MacAskill, Michael Page, Stefan Shubert, Carl Shulman, Pablo Stafforini, Rob Wiblin, and Julia Wise for comments and contributions to the conversation.
Opinion: Owen Cotton-Barratt
I actually used to donate to my employer, but I now think that the appropriate level is around c. on the spectrum. Probably we should allow or even encourage donation to an employer when the org is very small / just getting started, as the comparative advantage as a donor assessing the org is likely to be very large in this case. But I’d prefer to discourage donating to employers as those employers get bigger. I don’t know whether it would make sense for employers to explicitly refuse donations from employees, and I wouldn’t implement such a rule now, but I could see myself supporting one in the future.
I worry that normalising donation to employers is taking a short-sighted consequentialist view (“Where will my donation do most good?”), rather than asking which norms will lead to the best version of EA.
Opinion: Robert Wiblin
I have donated to my employer in the past and intend to continue, inasmuch as I think it’s among the best places for the money to go. The reason to give there is the same as the reason to work there in the first place—I think the intervention we are implementing is highly effective.
The amount of money employees at EA organisations can give is fairly small, so I think it’s reasonable to just choose a simple option. Forgoing salary is easy and highly tax efficient.
If you share the values of the people donating to your organisation, giving to your employer also ‘funges’ well; by shrinking the funding requirements of your org, you free up donors who would fund you to fund something else they judge to be effective. By doing this you leave it up to bigger donors to determine where the final marginal dollar goes.
- Passing Up Pay by (13 Jul 2022 14:10 UTC; 129 points)
- Where are you donating in 2020 and why? by (23 Nov 2020 8:47 UTC; 71 points)
- 2016 AI Risk Literature Review and Charity Comparison by (13 Dec 2016 4:36 UTC; 57 points)
- Why I’m donating to MIRI this year by (30 Nov 2016 22:21 UTC; 34 points)
- Passing Up Pay by (LessWrong; 13 Jul 2022 14:10 UTC; 29 points)
- CEA Staff Donation Decisions 2016 by (6 Dec 2016 16:07 UTC; 11 points)
- 's comment on Proposed: donation mechanism for people doing direct work (USA tax relevant) by (28 Nov 2022 20:12 UTC; 6 points)
- 's comment on Propose and vote on potential EA Wiki articles / tags [2022] by (10 Jun 2022 13:45 UTC; 4 points)
[ETA: I somewhat regret eliding the issue with donations to other issues re. interpersonal relationships. I take the opportunity to stress it wasn’t written with a particular individual/s ‘in mind’.]
I broadly agree with Owen. Beyond the considerations already mentioned, I prefer increasing professionalisation of those working in (at least the larger) EA orgs, and norms against salary sacrifice etc. I think would help this desirable direction of travel.
In most corporations (and most charities), including very high performing ones (e.g. Google, Deepmind, GiveDirectly, etc.) situations where employees are sacrificing considerable fractions of their salary, living in long term shared housing with their colleagues, or can name multiple fellow employees among their past or current sexual partners are rare. These arrangements are less surprising for smaller groups (e.g. the model of the start-up in the garage), but much more so when you have formal structures, a board, and turnover in the 6-7 figures.
I don’t claim expertise here, but I’d suspect there’s a reason for the ubiquity of ‘corporate’ structures, oft-maligned as they are, and I’d venture it has something to do with ensuring clarity and accuracy of decision making. It is better that, insofar as possible, corporate decisions are not interleaved with pecuniary, interpersonal, or bedroom issues. Similarly, norms that retard development of conflicts of interest like these are preferable to relying on staff to navigate them appropriately.
The failure modes are manifold. The hypothetical challenges around firing someone who is, in addition to one’s subordinate, a housemate, ex, and current partner of another staff member need not be explicated. Although I do not want to regurgitate the unhappy episode around Intentional Insights, there were records of virtual assistants offering to ‘donate’ money to Intentional Insights, the net result of which was to reduce what InIn owed them and had been late paying. I doubt any EA org has or will do anything similar, but it illustrates the risks of a parallel financial transaction alongside salary which could be used to bypass proper process (some possibilities: maybe I’m worried you might fire me for poor performance, so I increase my salary sacrifice by way of cutting my price; I fail to negotiate a raise I think I deserve from you, so I simply reduce my salary sacrifice/donation by the requisite amount (which annoys you); I don’t really take your standards re. dress, punctuality etc. seriously as although I am in principle getting paid £X, de facto you’re paying me barely minimum wage etc. etc.)
I’d generally prefer EA staff get paid close to market, and there’s a norm discouraging donations to ones employer.
Sorry but what does this have to do with donations? In what way are these at all equivalent with donations?
Greg’s point is that the case against donating to one’s employer is part of a larger argument for increased professionalization of EA orgs. The situation he describes in the paragraph you quote illustrates what can go wrong when an organization lacks the level of professionalism he thinks orgs should have.
Yes, I see that’s what he’s trying to hint at, but there’s zero indication that donations have any of the same effects on professionalism that close interpersonal relationships have. The problem is not “I don’t understand your argument”, it’s “you’re alleging something out of the blue with no support.” It should be clear that dating a coworker and donating to the organization are completely different issues in many relevant respects. I can easily draw up examples of behavior which don’t reduce professionalism and are actually more comparable—voluntary overtime, for instance, which companies don’t forbid, or people covering expenses for the organization, which is common in nonprofits and some government agencies.
I actually regret this paragraph for the opposite reason: the risk it came across as a veiled side-swipe, or (even worse) someone might take me to be offering gratuitous commentary on their private life. (I have disclaimed accordingly).
Pablo has interpreted me correctly. I agree donations etc. are different from the others, but I aver they are similar in that they undermine ‘corporate professional’ type norms that larger EA groups are well-advised to adopt, if not to the same degree (FWIW, I think voluntary overtime is at least slightly worrisome for similar reasons).
And what evidence is there that donations undermine professionalism?
I’ve often found it strange that GiveWell staff do their meta-donations primarily to GiveWell and CEA staff do theirs primarily to CEA. I’m guessing in many cases the deciding factor about who works where is whether it’s more convenient to live in the Bay or in Oxford, rather than any serious ideological difference.
I’d feel better if cross-organizational giving were more common (hopefully not just in a “scratch my back I’ll scratch yours” way).
What’s a meta-donation? Is that a donation to a metacharity, or something else?
“donation to a metacharity” is how I interpreted it
This.
Thanks.
While previously I had felt the strength of “Your choice of where to work and where to donate will be correlated”, on reading
my System 1 immediately jumped to “Well obviously they’re being tribal”. Insofar as I should(n’t) generalise from one example, this did make me update that it has the potential to be a noticeable image problem in the future.
The claim that it’s natural to donate to one’s employer given one’s prior decision to become an employee assumes that EAs—or at least those working for EA orgs—should spend all their altruistic resources (i.e. time and money) in the same way. But this assumption is clearly false: it can be perfectly reasonable for me to believe that I should spend my time working for some organization, and that I should spend my money supporting some other organization. Obviously, this will be the case if the organization I work for, but not the one I support, lacks room for more funding. But it can also be the case in many other situations, depending on the relative funding and talent constraints of both the organization I work for and the organizations I could financially support.
And the impact of one’s skillset in different orgs/focus areas.
Sure, but your estimates of the effectiveness of sending money or your labour to a project are correlated, so it’s no surprise if people are unusually likely to want to donate to the place they work.
No, the claim is not that all employees should donate to their employers. The claim is that it’s natural, viz. that there is a pro tanto reason for it.
I think the claim should be that there is a prima facie reason for donating to one’s employer. If the reason was pro tanto, one would have reason for donating even after learning that one’s employer e.g. has no room for more funding.
I agree with the claim so interpreted. If you believe working for some organization is the best use of your time, there’s a presumption that donating to this organization is the best use of your money. So I now see that my original comment was uncharitable.
At present, I don’t have a good sense of how strong this presumption should be. So it’s unclear to me how much weight I should give to arguments that appeal to this presumption.
I agree. I have friends who work at one EA organization but I expect would donate to another. I don’t think Rob and Owen were saying there should be a norm in favour of donating to one’s employer in general. I thought we were just taking for granted as an effective altruist they would only donate to their own employer if they thought it was the best use of their money. Of course it’s often not. Givewell employees probably donate to their recommended charities rather than Givewell. I’d bet the same is true of ACE employees and ACE’s recommended charities.
ACE and GiveWell have both written blog posts about where staff donate in the past. It’s been a mix of recommended charities, the employer organization, and other charities. On skimming, it looks to me like GiveWell staff, at least in 2015, more closely followed the recommendations of their employer than ACE staff.
(Links go to 2015 staff donation posts.)
That seems like something we would expect if GiveWell and ACE researchers are doing a good job, given that animal interventions seem to have less robust evidence than global poverty ones.
Agreed. Is there any evidence employee donation is a significant problem, or that it will become one in the near future? If not, and given there is no obvious solution, I suggest focusing on higher priorities (e.g. VIP outreach).
I think too many (brain power x hours) have been expended here.
Sorry to be a downer, just trying to help optimize.
“I think too many (brain power x hours) have been expended here.”
I agree that more thought went into this than was really called for, including by me. Law of triviality in action.
Evaluated just on the near-term impacts I’d agree with you (even if total time was, I guess, around a day, which is not huge).
The reason this seems more important to me is that it’s part of the community culture. Building a community with the right culture seems both:
easier to get done while everything is relatively small;
an important determinant of the long-term impact of effective altruism.
With a quick Fermi estimate from my cycle ride I think this might be worthwhile even on short-term impacts.
There are maybe 50 employees at EA orgs. Perhaps they plan to donate $200,000 in the next year. Suppose that a correct answer to this question adds +20% to the value of a donation (roughly the amount of the tax benefit, which people mostly weren’t claiming to be the biggest effect). Suppose further that this much-more-in-depth-consideration than previously existed has a net 5% chance of nudging people towards making the better decision, whichever that is. Then that would be worth ~~$2,000 even on a 1-year timescale. Which is plausibly worth a day.
Those numbers are close and messy enough I don’t feel happy supporting it just on short-term impact, though.
Many people won’t donate to their own org either way, and many of those who do won’t care about norms. This will probably include a large majority of donations.
This came out of my pleasure budget.
I feel as though building a good culture is really quite important, and like this sort of specific proposal & discussion is how, bit by bit, one does that. It seems to me that the default for large groups of would-be collaborators is to waste almost all the available resource due basically to “insufficiently ethical/principled social fabric”.
(My thoughts here are perhaps redundant with Owen’s reply to your comment, but it seems important enough that I wanted to add a separate voice and take.)
Re: how much this matters (or how much is wasted without this), I like the examples in Eliezer’s article on lost purposes or in Scott Alexander’s review of house of god.
The larger EA gets, the easier it is for standard failure modes by effort becomes untethered from real progress, or some homegrown analog, to eat almost all our impact as well. And so the more necessary it is that we really seriously try to figure out what principles can keep our collective epistemology truth-tracking.
I agree with Owen. I don’t have anything to add to what’s been said, other than a response to the strongest reason against having that norm: It only conflicts with the norm of “do what’s most effective” if it truly is more effective to donate to one’s own employer. But because of the signaling/weirdness reasons (and, yes, the bias) that doesn’t seem to be true. We’re sophisticated enough that we can have a hierarchy of norms, with “do what’s most effective” at the top and “don’t donate to your employer unless there’s a special circumstance” as a lower norm—as a helpful heuristic/guideline.
How much money is saved from taxes by foregoing salary? If it’s at least 20% of the donation then I might change my mind.
+1 for the statement of what would change your mind.
In the UK I think it’s normally about 7% (assuming you’d get tax deductability / Gift Aid on another donation).
Does that include National Insurance? As you can’t claim back NI from Gift Aid, but you never pay it if you forego salary, the saving looks like it would be 12% on employee NI and 13.8% on employer NI (if I’m interpreting the taxes properly).
(source: https://www.gov.uk/national-insurance-rates-letters/contribution-rates)
Wow, thanks. It was supposed to represent National Insurance, but it was based on a remembered figure. I think I must have worked out an average rate and assumed it was marginal at some point; perhaps this was also just on employee contributions.
Anyhow, I think that makes it 22.5% lost to taxes (so +29% on your donation) if you’re a basic rate taxpayer, and 13% lost (+15% on donation) if you’re a higher rate taxpayer.
Adding another mark in the “Look factual information up even when you think you know the answer” ledger.
Were I working for an EA org this would be the decisive factor that would swing me, so it would be really good if we could work this out. Giving to another org adds Gift Aid to your donation. +20% Forgoing salary saves you and your employer National Insurance. +29%
So if you’re basic rate, is giving to your employer better value?
The only reason affecting giving is doing what’s most effective. So the norm you’re proposing for employer donations should never come into effect. Every rational person who donates to their employer will claim that it’s a most effective use of their money.
“They push against norms of allowing personal freedom in how to allocate resources. EA tends to not want to dictate to people what they may or may not do.”
I can confirm that someone trying to stop me giving wherever I thought the money would do the most good—whether my employer or elsewhere—would infuriate me.
Another possible intermediate norm would be to discourage talking about donating to one’s employer. This would reduce some of the pressure for bias, some of the conflict of interest, etc., by keeping it secret from colleagues.
It might even be possible to keep it secret from the employer—or visible only to payroll, who would report only how much was donated in total by employees, not by whom.
A disadvantage of this is that it would push against the EA norm of being open about giving. It could even make it hard for employees to talk openly about their external giving, because a conspicuous absence of discussing external giving might mean giving to the employer instead. (That might be circumvented if everyone were giving something to external orgs and could honestly talk about where and why, and they didn’t talk about the amounts they were giving to specific orgs.)
Don’t talk about donating to your employer… how do you think that will play out? “Shh, don’t tell us about where you donated.” I’m not sure how to explain this gently, but ordinary people who aren’t immersed 24⁄7 in EA organizations are likely to view this as an extremely strange and weird thing to do, and it’s worth having a sanity check in cases like this to make sure that your ideas are healthy and robust. You’ll need to make sure that your claims are well grounded in reality and evidence rather than speculation, and to ensure that reasonable people (incl. people other than the average readers on this forum, who have already displayed a penchant for elaborate solutions to nonexistent problems) could see them as reasonable points of view, if you want to propose anything which is successfully picked up by the broader community.
I’d guess that a number of the people working at organizations take jobs there precisely because they believe they are the most effective organizations. In those cases there is no room for alleging bias. Moreover, there’s still room for bias with external organizations. If I work at ACE I’m biased in favor of the charities I reviewed. If I work at FHI I’m biased in favor of MIRI because we do some stuff together. And so on.
Also note that people in an organization have inside information that outsiders lack. Are corporate employees banned from trading in their own stock because there is worry over them being optimistically biased about their own company’s prospects? No, it’s because such positions provide an unusual advantage in financial markets.
It might help to be more specific about what we mean by ‘norms’. Obviously no one is going to ban people from donating to their employer. The first thing I would do as an organization or as a donor would be to get a third party to circumvent the ban. So what’s left… social stigma? How? Why? If there is worry about bias in donations, just tell people—hey, here’s some reasons to be cautious before you donate to your employer. Just give them the reasons you are using as an argument for the norm. Norms just supervene on people’s opinions. If people see this idea and think it’s powerful and share it with their friends, you have the norm. If they don’t, then you don’t have the norm.
You can argue for all kinds of norms pushing in various directions based on allegations of cognitive bias (e.g. people who grew up poor should not donate to Givewell charities because their life experiences will bias them, vegans should not donate for animal relief because their diet will bias them, etc), but actually demonstrating the problem is another matter. We don’t know if biases significantly apply to this case. Donating money is a serious, occasional, major financial decision being conducted by people who are already trying to be very rational, so it’s a very strange domain for one to introduce paternalism.
Finally, there’s more personal value in being aligned with your workplace. Employees who donate are likely to experience having a greater stake and greater motivation for the organization.
I agree with this, but I’m a bit less concerned about it because I think that there is less pressure for bias than from the daily interactions in the workplace, and less possibility of setting bad precedent by acting on such bias.
One thing that hasn’t been mentioned is that it’s not clear overall you are biased in favour of the org you work for based on experience—you also know everything that’s bad which isn’t obvious to outsiders. I usually have a more negative view of things I’m more familiar with (though try to debias by realising everything I don’t know well is probably bad in ways I can’t see as well).
That’s clearly true as well. There is strong conventional wisdom to essentially ‘never meet your heroes’. My personal experience is that going to actually work with an organization tends to lower one’s level of attraction and enamoring for that organization.
There’s still room for bias, so why not take action? Why draw the line there and not here? Why not make up a norm against donating to related charities? And a norm against people who grew up in poverty donating to poverty charities? A norm against people donating to charities within their city?
Why not make a norm against volunteering? Lots of EA organizations have volunteers. Perhaps volunteering is bad because this will create a conflict of interest when former volunteers seek employment.
I think I’m coming at this from a basically different perspective. I’m really not trying to make allegations of anyone being biased. I’m in favour of removing a conflict of interest, for basically the same reasons this comes up in other domains (e.g. recusing someone from a selection panel when they are related to a candidate): there might be bias, there might be pressure towards bias, and there might be the appearance of bias.
The particular point of expecting it to be somewhat frequent that people want to work for and donate to the same org for totally legitimate reasons I agree with and made in the post.
Sure, there MIGHT be bias. So what? This type of ‘conflict of interest’ is totally speculative. It’s not existed for other nonprofit organizations where people have made donations to their employer. It’s not existed for employees working overtime. It doesn’t exist for the military (fairly commonplace for individuals to cover unit expenses there). I’ve never heard anything about this worry from inside or outside EA. You need to have a solid case before deciding that something is a problem. Right now you’re trying to find a solution for a problem that doesn’t exist, and what’s worse is you think that it should determine the long run culture of the movement.