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:

  1. It should be ille­gal to donate to one’s employer

  2. EA orgs should not ac­cept dona­tions from employees

  3. There should be a strong norm in EA of not giv­ing to employers

  4. 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)

  5. There should be no norm in EA re­gard­ing this

  6. There should be a pos­i­tive norm in EA of giv­ing to employers

  7. 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.