Empirical data on value drift

Why It’s Im­por­tant to Know the Risk of Value Drift

The con­cept of value drift is that over time, peo­ple will be­come less mo­ti­vated to do al­tru­is­tic things. This is not to be con­fused with chang­ing cause ar­eas or meth­ods of do­ing good. Value drift has a strong prece­dent of hap­pen­ing for other re­lated con­cepts, both eth­i­cal things (such as be­ing veg­e­tar­ian) and things that gen­er­ally take willpower (such as stay­ing a healthy weight).

Value drift seems very likely to be a con­cern for many EAs, and if it were a ma­jor con­cern, it would sub­stan­tially af­fect ca­reer and dona­tion plans.

For ex­am­ple, if value drift rarely hap­pens, putting money into a sav­ings ac­count with the in­tent of donat­ing it might be ba­si­cally as good as putting it into a donor-ad­vised fund. How­ever, if the risk of value drift is higher, a dol­lar in a sav­ings ac­count is more likely to later be used for non-al­tru­is­tic rea­sons and thus not nearly as good as a dol­lar put into a donor ad­vised fund, where it’s very hard not to donate it to a reg­istered char­ity.

In a ca­reer con­text, a plan such as build­ing ca­reer cap­i­tal for 8 years and then mov­ing into an al­tru­is­tic job would be con­sid­ered a much bet­ter plan if value drift were rare than if it were com­mon. The more com­mon value drift is, the stronger near-term fo­cused im­pact plans are rel­a­tive to longer-term fo­cused im­pact plans. For ex­am­ple, you might get an en­try-level po­si­tion at a char­ity and build up ca­pac­ity by get­ting work ex­pe­rience. This has the po­ten­tial, though not always, to be slower at build­ing your CV than get­ting a de­gree or work­ing in a low-im­pact but high-pres­tige field. How­ever, it has im­pact right away, which mat­ters more if the risk of value drift is high.

The Data

De­spite the im­por­tance of value drift to im­por­tant ques­tions, it’s rarely been talked about or stud­ied. One of the rea­sons it is so un­der-stud­ied is that it would take a long time to get good data.

I have been in the EA move­ment for ~5 years. I de­cided to pool some data from con­tacts who I met in my first year of EA. I only in­cluded peo­ple who would have called them­selves EAs for 6 months or longer (I would not in­clude some­one who was only into EA for a month and then dis­ap­peared), and who and took some sort of EA ac­tion (work­ing for an EA org, tak­ing the GWWC pledge, run­ning an EA group). I also only in­cluded peo­ple who I knew and kept in touch with well enough to know what hap­pened to them (even if they left the EA move­ment). It is ul­ti­mately a con­ve­nience sam­ple, but it was based on work­ing for 4 cur­rent EA orgs and liv­ing in 4 differ­ent coun­tries over that time, so it’s not fo­cused on a sin­gle lo­ca­tion or or­ga­ni­za­tion.

I also broke the groups down into ~10% donors and ~50% donors, be­cause many times I have heard peo­ple be­ing more or less con­cerned about one of these groups vs the other. Th­ese broad groups are not just fo­cused on peo­ple do­ing earn­ing to give. Some­one who is work­ing heavy hours for an EA or­ga­ni­za­tion and mak­ing most of their life de­ci­sions with EA as their num­ber one pri­or­ity would be con­sid­ered in the 50% group. Some­one run­ning an EA chap­ter who makes de­ci­sions with EA as a fac­tor, but pri­ori­tizes other fac­tors above it, would be put in the 10% group. The per­centages are aimed at rough prox­ies of how im­por­tant EA is in these peo­ple’s lives, not strictly fi­nan­cial dona­tions. I did not count chang­ing cause ar­eas as value drift (e.g. chang­ing from donat­ing 10% to MIRI to AMF) -- only differ­ent lev­els of over­all al­tru­is­tic in­volve­ment.

The re­sults over 5 years are as fol­lows:

16 peo­ple were ~50% donors → 9/​16 stayed around 50%

22 peo­ple were ~10% donors → 822 stayed around 10%

No one moved from the 10% cat­e­gory to the 50% cat­e­gory, and I only counted fairly no­tice­able changes (if some­one changed their dona­tions from 50% to 40%, I would not have the re­s­olu­tion to no­tice).

Value drift was high across both groups, with roughly 50% of the pop­u­la­tion drift­ing over 5 years. I talked to many of those peo­ple about value drift and their thoughts on long term al­tru­ism, and most of them, like most peo­ple I talk to now, had pre­vi­ously been very con­fi­dent that they would stay al­tru­is­tic in the long term. Why peo­ple gen­er­ally value drifted was a mix of rea­sons with no clear con­sis­tent source, al­though life changes were a large fac­tor for many (e.g. peo­ple mov­ing from uni­ver­sity to the work­force, chang­ing cities or work­places, mar­ry­ing or hav­ing kids).

In­ter­est­ingly, this data also sheds a lit­tle light on con­cerns of “push­ing your­self too hard” and vs “tak­ing it too easy on your­self”, with gen­er­ally more in­volved or ded­i­cated peo­ple value drift­ing no­tice­ably less (~30% vs ~60%).


Over­all, these re­sults seem pretty scary to me, es­pe­cially since there’s a nat­u­ral se­lec­tion effect where I tend to make friends who are more ded­i­cated and drift apart from the ones who leave the move­ment. It’s also worth not­ing that there have not been par­tic­u­larly ma­jor con­tro­ver­sies or prob­lems in the EA move­ment that would cause a lot of value drift over this pe­riod of time vs. any other. His­tor­i­cally, many EAs have been young non-fam­ily-starters, so we’ve ar­guably been see­ing a pe­riod of ar­tifi­cially low value drift that is not sus­tain­able as the move­ment gets older and goes through stan­dard life changes.

Of course, the data could be a lot bet­ter qual­ity, and I wish it were mea­sured in a more rigor­ous way. I would be keen to see any more data that any­one else has along these lines. De­spite qual­ity con­cerns, I still think we can draw some con­clu­sions from it, par­tic­u­larly given that peo­ple are already effec­tively draw­ing con­clu­sions about the like­li­hood of value drift from no data at all. I also spoke to a few EAs who have been around the move­ment awhile, and this data broadly fit their in­tu­itions, which gives me more con­fi­dence that it’s not 100% off the mark.

The im­pli­ca­tions of this data are that peo­ple should be cau­tious of defer­ring im­pact to later, and should set up com­mit­ment de­vices to help them stick to what they care about.

One ex­am­ple: be wary of build­ing ca­pac­ity for very long pe­ri­ods of time, par­tic­u­larly if the built ca­pac­ity is broad and leaves open ap­peal­ing non-al­tru­ist paths. In­stead, see if you can build ca­pac­ity in such a way that it also does good at the mo­ment, such as get­ting work ex­pe­rience with non-prof­its.

For in­stance, if you want to do di­rect im­pact, vol­un­teer­ing for an or­ga­ni­za­tion and show­ing how good your work is is of­ten bet­ter than hav­ing a de­gree, es­pe­cially an un­re­lated one. It’s also sub­stan­tially faster and does good di­rectly. De­grees are largely a way to sig­nal that you’re a hard worker with a de­cent amount of in­tel­li­gence, and if all you are is a CV in some­body’s in­box, that’s very im­por­tant. How­ever, if you’ve been work­ing alongside them for months, they’ll already know these traits of yours.

Another way to build ca­reer ca­pac­ity with value drift in mind is to get ex­pe­rience and cre­den­tials that make it harder to work in a non-al­tru­is­tic area. This could be get­ting a de­gree in de­vel­op­ment eco­nomics in­stead of eco­nomics gen­er­ally, or work­ing for pres­ti­gious non­prof­its in­stead of other pres­ti­gious or­ga­ni­za­tions. Op­tion value is great if the risk of value drift is low, but if it’s high, it makes it eas­ier for you to slip. It’s like only hav­ing healthy food in the house. If the only easy op­tions are also al­tru­is­tic, you’re much more likely to stick it out in the long haul.

If your pri­mary path to im­pact is dona­tions and you want to keep value drift in mind, but you don’t know where you want to give yet, don’t save those dona­tions. Put them into a donor-ad­vised fund. That way, even if you be­come less al­tru­is­tic in the fu­ture, you can’t back out on the pledged dona­tions and spend it on a fancier wed­ding or a big­ger house. You can also set up monthly dona­tions, or ask your em­ployer to au­to­mat­i­cally donate a pre­set por­tion of your in­come to char­ity be­fore you even see it in your bank ac­count.

Over­all, if 50% of the EAs I met 5 years ago have value drifted, this should fac­tor into your plans. No­body thinks they’ll value drift, just like no teenager with a fast metabolism thinks they’ll be the one who gains weight when they hit mid­dle age. By all means, in­dulge in junk food ev­ery once in awhile and don’t con­stantly stress about calories, but put some time into set­ting up your life to make it eas­ier for you to reach for a ba­nana in­stead of the ice cream, or in this case, the al­tru­is­tic path in­stead of the less al­tru­is­tic one.

For a deeper dive into con­crete ways to re­duce value drift, check out this post.