Direct Funding Between EAs—Moral Economics

News: EA Global Google­plex par­ti­ci­pants—I’ll be giv­ing an ig­nite talk in­tro­duc­ing the long vi­sion for Mo­ral Eco­nomics!
Edit: Here is the talk

More news: Giles cre­ated a new #Mo­ral­trade chan­nel on Slack, which is be­ing used dur­ing the Sun­day US UK on­line EA workathons—in per­son for those in SF. To see the workathon Sun­day event go here. And to en­ter the chan­nel for Mo­ral Trade en­ter here.

In these writ­ings we pro­pose the cre­ation of a sub-field of knowl­edge, Mo­ral Eco­nomics.

In this post I will make a de­tailed case for di­rect dona­tion be­ing more effi­cient than more com­mon meth­ods of fund­ing EA work, and cite many other benefits on top of effi­ciency. I be­lieve we can and should ad­dress its un­der­use.

Spe­cial Thanks to Eli Tyre, Dave Deken­berger, Matt Reyes, Giles Ed­kins and Ben Hoff­man for ed­its and com­ments.

Mo­ral eco­nomics series

  1. In­tro­duc­ing Mo­ral Economics

  2. Ex­am­ples of Mo­ral Eco­nomics Concepts

  3. Branches Within Mo­ral Economics

  4. Mov­ing Mo­ral Eco­nomics Forward

  5. Direct Fund­ing Between EAs

  6. Cer­tifi­cates of Im­pact, Do­ing It Right—Giles Edkins

    1. Mo­ral mar­ket failure: how COIs might help

    2. Prob­lems with COIs, and their solutions

    3. Im­ple­ment­ing COIs

  7. Agen­tial Iden­tity in Mo­ral Economics

Direct Fund­ing Between EAs—Mo­ral Economics

We will ex­am­ine the cur­rent op­tions that trans­form re­sources into la­bor within effec­tive al­tru­ism, with a spe­cial em­pha­sis in the least well-known idea of Direct Fund­ing Between EAs, that is di­rect dona­tions be­tween EAs.

Pre­vi­ously we de­scribed:

Direct Fund­ing Between EAs: A di­rect link­age be­tween an in­di­vi­d­ual who wishes to have their val­ues pro­moted and an­other who at­tempts to pro­mote these val­ues in ex­change for money.

This has been at­tempted within the Effec­tive Altru­ist com­mu­nity by cre­at­ing Grati­pay and Pa­treon dona­tion por­tals for EAs.

The Sta­tus Quo

Cur­rently there are at least four main forms of mon­e­tary ex­change be­tween Effec­tive Altruists

In­sti­tu­tional Dona­tions: The vast ma­jor­ity of dona­tions are go­ing to in­sti­tu­tions, this in­cludes both recom­mended char­i­ties and in­sti­tu­tions fa­mil­iar to the EA move­ment such as CFAR, FHI, MIRI, CEA, Lev­er­age etc…

Cer­tifi­cates of Im­pact: A very small frac­tion of re­sources is go­ing to cer­tifi­cates of im­pact, which rep­re­sent al­tru­is­tic valuable ac­tions taken in the past, and can be pur­chased, in which case the buyer be­comes re­spon­si­ble for the good deeds pur­chased.

Direct Re­quests for Events: Some events like the HPMOR par­ties were ar­ranged di­rectly by ask­ing donors to fi­nance the event.

Direct Fund­ing for EAs: A very small frac­tion of re­sources is be­ing trans­ferred via Grati­pay and Pa­treon be­tween EAs.

The ini­ti­a­tive EA Ven­tures pro­poses a model of re­source trans­fer to spe­cific pro­jects which com­bines some of the above, to fa­cil­i­tate con­nect­ing pro­jects to po­ten­tial fun­ders.

Direct Fund­ing—De­tails and Advantages

There are sev­eral im­por­tant differ­ences be­tween di­rect dona­tion, hiring some­one in an Effec­tive Altru­ist in­sti­tu­tion, and pur­chas­ing peo­ple’s cer­tifi­cates of im­pact. Direct dona­tions are based on trust­ing what the per­son has done be­fore, and what she claims she will do now. They are more sta­ble than cer­tifi­cates—which are as sta­ble as bounty hunt­ing—and thus could be the en­tire source of in­come for in­di­vi­d­u­als funded by donors who share their val­ues, in­stead of in­sti­tu­tional dona­tions. Direct dona­tions are also much cheaper per worker hour than in­sti­tu­tional dona­tions. As a back of the en­velope calcu­la­tion I es­ti­mated the differ­ence as about three­fold, and even af­ter fac­tor­ing 501c, char­ity sta­tus, Pa­treon’s fee, and all other dis­counts, it re­mains two times more ex­pen­sive to donate to in­sti­tu­tions than di­rectly.

The cur­rent cli­mate within the Effec­tive Altru­ism com­mu­nity is one of skep­ti­cism of di­rect fund­ing which re­ceives a small frac­tion of re­sources, skep­ti­cism of cer­tifi­cates of im­pact, which also re­ceive a small frac­tion of re­sources, and profound trust in in­sti­tu­tions. This trust at a re­cent EA Think Tank meet­ing was called “sanc­tifi­ca­tion” of in­sti­tu­tions—which re­ceive the vast ma­jor­ity of fi­nan­cial and vol­un­teer­ing re­sources from Effec­tive Altru­ist donors. This is un­jus­tified and could be taken as a tem­po­rary failure in the moral mar­ket. I ex­pect this trend to change over time as the ideas of di­rect fund­ing and cur­ren­cies such as cer­tifi­cates of im­pact be­come more in­te­grated into the strate­gic think­ing of in­di­vi­d­ual donors, and the snow­ball of in­creas­ing trust in long-term com­mit­ted in­di­vi­d­u­als gen­er­ates a self-re­in­forc­ing in­cen­tive.

Direct dona­tions would be a great ad­di­tion to the moral mar­ket, since they offer many ad­van­tages, some of which are sim­ply un­available to in­sti­tu­tions or cer­tifi­cates of im­pact. Let us look at some of these ad­van­tages:

Direct fund­ing per­mits in­ves­ti­ga­tion of ne­glected causes: It re­quires merely two in­di­vi­d­u­als for a cause to be in­ves­ti­gated. If, for ex­am­ple, a sin­gle donor con­sid­ers an in­ves­ti­ga­tion on the dis­rup­tive po­ten­tial of Emu­la­tion Eco­nomics to the global econ­omy, and one in­di­vi­d­ual wants to un­der­take this re­search, this suffices for at least some at­ten­tion to be al­lo­cated to that prob­lem. Even causes ne­glected within EA can be eval­u­ated.

Direct fund­ing is sta­ble over time: Cer­tifi­cates of im­pact pay back­ward in time. An in­di­vi­d­ual can­not an­ti­ci­pate to re­ceive them be­fore­hand be­cause other pro­jects may be judged more valuable. Direct dona­tions done monthly could guaran­tee se­cu­rity for EAs that in­tend to zig-zag be­tween in­sti­tu­tions, re­searchers who want to re­search a spe­cific cause, and other helpers who would un­der­take a func­tion that isn’t defined by a job-role. Direct dona­tions are a con­tract of trust be­tween donor and donee, and the donee could es­tab­lish ways for the donors to keep track of the work they are pay­ing for, such as by send­ing pe­ri­odic re­ports or hav­ing pe­ri­odic meet­ings.

Direct fund­ing per­mits un­usual work: Perfor­mance is hard to mea­sure. In­di­vi­d­u­als who trav­el­led be­tween EA in­sti­tu­tions and in­fluenced the EA move­ment via hard to as­sess means (e.g. Michael Vas­sar, Justin Sho­eve­lain, Jasen Mur­ray, Scott Aaron­son...) are fre­quently vol­un­teer­ing their time and effort to EA causes. Some of these in­di­vi­d­u­als have built a rep­u­ta­tion of do­ing great work as EAs though they were always be­hind differ­ent pro­jects. In the­ory, in­sti­tu­tions could hire such peo­ple, but in prac­tice, di­rect dona­tions would be a much more effi­cient way of in­cen­tiviz­ing more peo­ple to be EA gen­er­al­ists. Once a thread of trust is built, these EAs could have free­dom to work in what seems more ap­pro­pri­ate to them, as long as their back­ers still thought they were on the right track. More im­por­tantly, many more peo­ple could be­gin to work as EA gen­er­al­ists who cur­rently don’t even con­sider that pos­si­bil­ity, the labour mar­ket would be­come more fluid.

Direct fund­ing pre­vents group­think: In­sti­tu­tions need to have their fo­cus well un­der­stood, and re­quire some level of in­ter­nal goal al­ign­ment. Bold new differ­ent ideas of what needs to be done by EAs, or what needs to be re­searched, or new meth­ods of re­search won’t come eas­ily out of in­sti­tu­tions—this is less of a prob­lem within the EA com­mu­nity than out­side it, but the prob­lem is not com­pletely elimi­nated. Direct fund­ing would at­ten­u­ate it even fur­ther; If a per­son with a good track record of EA ideas in the past wants to pur­sue a very differ­ent pro­ject, their di­rect donors could guaran­tee their fi­nan­cial sta­bil­ity with­out jeop­ar­diz­ing ei­ther the per­son’s ac­tion flex­i­bil­ity, nor the rep­u­ta­tion of the in­sti­tu­tion to which the per­son works for, since there would be none.

Direct fund­ing is in­stan­ta­neous: Hiring an in­di­vi­d­ual can take a long time, es­pe­cially a for­eigner. Direct dona­tions can be done as soon as there is an agree­ment on con­di­tions and trust be­tween donor and donee.

Direct fund­ing fa­cil­i­tates in­ter­na­tional EA work: The rea­sons here are clear. It is bu­reau­crat­i­cally less costly to trans­fer money in­ter­na­tion­ally than to do it via in­sti­tu­tions.

Direct fund­ing makes mov­ing to EA hubs eas­ier: Many EAs move to lo­ca­tions with more EAs. It is much eas­ier to do this if there are sta­ble in­come sources that are re­gion in­de­pen­dent and provide a fal­lback dur­ing the pe­riod of tran­si­tion. More im­por­tantly, dona­tions are con­sid­ered pas­sive in­come, and as far as I know dona­tions within a bounded range can be re­ceived by non-US cit­i­zens in US soil, even in non-fa­vor­able Visa con­di­tions. Similar con­straints are re­laxed for dona­tions to other coun­tries as well. To move to Oxford, Berkeley, Switzer­land or Melbourne would be much eas­ier at the mo­ment for EAs who are re­ceiv­ing the ma­jor­ity or en­tirety of in­come through di­rect dona­tions.

Direct fund­ing al­ters the mind­set of the re­cip­i­ent: If you are hired by an EA in­sti­tu­tion to do a par­tic­u­lar set of tasks, you are be­ing paid to do those tasks, and help that or­ga­ni­za­tion do well in its func­tion. If you are di­rectly be­ing paid to be an EA how­ever, this is much more in­grained in your iden­tity. There are peo­ple pay­ing you to do the most good you can do. If you see an op­por­tu­nity for do­ing good, you will be more likely to take it, since you are never done with the obli­ga­tion of be­ing a great EA.

Direct fund­ing can be the en­tire source of in­come of an in­di­vi­d­ual: This seems to be the un­tapped se­cret of di­rect fund­ing. If some­one wants to do EA work all the time, they can be fully funded by donors so as to not need any in­sti­tu­tional af­fili­a­tion, or have a to­ken one. This thresh­old will be what de­ter­mines the ul­ti­mate value of di­rect dona­tions, since the point at which there are suffi­cient in­cen­tives for in­di­vi­d­u­als to try do­ing this as their main in­come source will be the tip­ping point for most of the ad­van­tages above to man­i­fest in full. When peo­ple can con­fi­dently un­der­take an EA ca­reer or a paid year on the ba­sis of dona­tion, then the la­bor mar­ket will have the nec­es­sary fluidity to ac­crue all of the ad­van­tages that can be had via di­rect fund­ing.

Even though they are im­por­tant all these ad­van­tages and de­tails are com­pletely dom­i­nated by the cru­cial fac­tor when it comes to effec­tive dona­tion.

Direct Fund­ing is Very Cheap

In­sti­tu­tions are made of peo­ple. When you pay an in­sti­tu­tion to do some­thing, you are ul­ti­mately pay­ing the peo­ple par­ti­ci­pat­ing in it to do some­thing. If you are an effec­tive al­tru­ist donat­ing to an EA char­ity, then you are pay­ing some­one be­cause you want them to do some­thing. This thing is what has value for you. Hav­ing in­sti­tu­tions and char­ity sta­tus are very valuable for the EA move­ment, and it is one of the ways in which it causes a lot of peo­ple to do a lot of things.

The ques­tion then is, is there a niche or a space for caus­ing peo­ple to do what needs to be done also out­side in­sti­tu­tional bound­aries?

The re­sponse is a re­sound­ing yes!

Why would we want to com­ple­ment in­sti­tu­tional roles with in­di­vi­d­ual roles? Ba­si­cally be­cause in­di­vi­d­u­als are cheap. It costs a lot in dona­tions to hire a new worker at an in­sti­tu­tion, whereas there are many EAs who would (and do) work for much less money. In­di­vi­d­u­als who work in EA in­sti­tu­tions fre­quently cost 75.000USD or more per year to their donors. Some EAs on the other hand would be happy and will­ing to work on their pro­ject for a third of that, 25.000USD. In fact for a cou­ple years at least some EA or­ga­ni­za­tion paid less than that. The differ­ence in cost to donors is strik­ing. Even af­ter ac­count­ing for the fees that Pa­treon and Grati­pay charges (~3%), tax ex­emp­tions for char­i­ties and other ad­van­tages, the differ­ence is still very sub­stan­tial. In­sti­tu­tions have to bear many costs that in­di­vi­d­u­als don’t have to, and if I re­call cor­rectly, di­rect dona­tions are not sub­ject to in­come tax on the donee’s part, which makes it even cheaper for the donors. Some­times in­sti­tu­tions also have to pay ac­cord­ing to a table that de­ter­mines how much peo­ple in differ­ent jobs should make, ir­re­spec­tive of whether both par­ties would pre­fer oth­er­wise. No such con­straint ex­ists for dona­tions.

For an al­tru­is­tic donor, there are two num­bers that mat­ter: num­ber of dol­lars that they have spent on the thing they value, and num­ber of things that got done to­wards that value be­ing ac­com­plished. The cheap­est way to do so seems to me hands down to be di­rect dona­tion to in­di­vi­d­u­als. It is at least twice as cheap.

Why Are In­sti­tu­tions Sacral­ized?

When dis­cussing this strange phe­nomenon with other EAs, some­one (who works for a high pay­ing in­sti­tu­tion, so I’ll pro­tect their name) sug­gested this in­ter­est­ing ques­tion. It seems to me that the rea­son why in­sti­tu­tions are sacral­ized is that it feels like an in­sti­tu­tion is much closer to a cause or goal in men­tal space. The as­so­ci­a­tion be­tween “sav­ing the en­vi­ron­ment” and “Green­peace” is much stronger than that of “sav­ing the en­vi­ron­ment” and “Al Gore”. Al Gore may have strongly ad­vo­cated for en­vi­ron­men­tal causes, but a per­son seems less like a cause than an in­sti­tu­tion does. The for­est, how­ever, is made of trees, and Green­peace is made, in great part, of peo­ple work­ing in en­vi­ron­men­tal causes.

EAs are no­to­ri­ous for their abil­ity to rea­son through their felt sense, and do what they think is valuable even if it doesn’t nec­es­sar­ily feel im­me­di­ately right. EAs are prob­a­bly good then at com­bat­ing the felt sense that donat­ing to in­sti­tu­tions is donat­ing di­rectly to a cause, in­stead of for peo­ple who work on/​ad­vo­cate/​re­search on a cause. Donat­ing to in­sti­tu­tions is pay­ing for the peo­ple who work there, plus taxes, rent, re­tire­ment benefits, le­gal fees, in­sti­tu­tional costs, and many other col­lat­er­als. Maybe 70% or even 90% of EA paid work should in­deed be done through in­sti­tu­tions and wages, but I am very skep­ti­cal that not even 10 EAs should be be­ing paid di­rectly to work on EA pro­jects, re­search, ad­vo­cacy, etc… When we con­sider that this could also be done to pay in­di­vi­d­u­als tem­porar­ily while they re­ar­range Visa situ­a­tions, while their le­gal con­tract is pro­cessed, or that it can be used to fi­nance EAs who live in coun­tries with ad­van­ta­geous cur­ren­cies, I find it im­pres­sive that this was not cho­sen by any­one yet as the prob­lem they will solve.

Creat­ing an EA dona­tion net­work is one of the most valuable things we can do. Ozzie Gooen, Pa­trick Brinich-Lan­glois, and many oth­ers have be­gun to do this, but more needs to be done, and more donors should con­sider di­rect dona­tions their pri­mary place to donate. Also, we should con­cen­trate our dona­tions to­ward EAs with good port­fo­lios of EA work com­pleted, un­til dona­tions amount to a sig­nifi­cant frac­tion of their in­come. Hav­ing 10 in­di­vi­d­u­als re­ceiv­ing 1/​10th of the in­sti­tu­tional cost for 1 worker won’t give any of them the se­cu­rity and trust needed to pur­sue an EA ca­reer and EA pro­jects, but hav­ing 3 re­ceive 1/​3rd of that in­stead would likely cause all three of them to feel con­fi­dent in this op­por­tu­nity. At least two times more work would be done. If this were life-sav­ing work, then twice the lives would be saved.

As the un­named per­son above pointed out: It is as if EA in­sti­tu­tions have be­come sa­cred, but there is no rea­son for this.

Im­prov­ing Direct Dona­tions Between EAs

Pa­treon ver­sus Gratipay

Ozzie—who put up the Grati­pay net­work—and I ad­vo­cate a tran­si­tion to Pa­treon as quickly as pos­si­ble, since it is much more likely to be stand­ing in the near fu­ture, Grati­pay has gone down once already and has lower trans­ac­tion vol­ume.


One is­sue that comes up when dis­cussing di­rect dona­tions is “how to make sure the donors trust their donees to be do­ing good work?” This is one more rea­son to re­duce the num­ber of bas­kets in which donors should put their eggs. It is much eas­ier to have one or two in­di­vi­d­u­als you trust do­ing work you value, and keep track of what they are do­ing, than to in­vest at once in many peo­ple. Espe­cially for ne­glected causes, it is valuable to have a set­ting where only one large or two smaller donors suffice to fi­nance some­one they trust.


It is of course up to the dyad to es­tab­lish how they will de­ter­mine for how long the dona­tions con­tinue, how much free­dom the donee has to change pro­jects, and which types of re­port they will give. To make this eas­ier, donees could have a stan­dard way to re­port to their donors, and make con­tex­tual ad­just­ments when re­quested.


EA donors have es­pe­cially high re­gard for ex­ter­nal eval­u­a­tors such as GiveWell, GWWC and An­i­mal Char­ity Eval­u­a­tors. If in­di­vi­d­u­als are com­pet­ing with or­ga­ni­za­tions for funds, it is likely that donors will place a similar bur­den of proof on each. As such, an eval­u­a­tor for di­rectly funded in­di­vi­d­u­als might be­come nec­es­sary (al­though note that ex­plic­itly EA-al­igned or­ga­ni­za­tions which rely on dona­tions, such as CEA, also fall out­side the scope of cur­rent ex­ter­nal eval­u­a­tors).


Even­tu­ally, it would also be valuable to have fi­nan­cial guaran­tees for donees be­ing funded. If a donor de­cides to stop fund­ing a per­son re­spon­si­ble for a pro­ject (say be­cause they had a fi­nan­cial emer­gency), the per­son still gets funded for 1 or 2 months by a guaran­tee fund, and gets some ex­tra ex­po­si­tion to other prospec­tive donors, who may choose to save the pro­ject from go­ing un­der and the per­son from fi­nan­cial in­se­cu­rity. Even if there were a guaran­tee bank, the whole pro­cess would still be much cheaper than hiring peo­ple in­sti­tu­tion­ally.

In the next post on Mo­ral Eco­nomics Giles Ed­kins will dis­cuss sev­eral new ideas re­lated to Cer­tifi­cates of Im­pact, and sug­ges­tions on how to cre­ate a moral mar­ket.

Don’t for­get to check the ig­nite talk on Mo­ral Eco­nomics if you are go­ing to EA Global at the Google­plex on Satur­day.