EA Meta Fund Grants—July 2019

This post con­tains our al­lo­ca­tions and some ex­plana­tory rea­son­ing for the grants made by the EA Meta Fund in the July 2019 grant round, which was paid out in Au­gust. This write-up was also posted on the EA Funds page here.

Fund: Effec­tive Altru­ism Meta Fund

Pay­out date: Au­gust 24, 2019

Pay­out amount: $466,000.00

Grant au­thor(s): Luke Ding, Alex Foster, Denise Melchin, Matt Wage

Grant re­cip­i­ents:

Grant rationale

The EA Meta Fund made the fol­low­ing grant recom­men­da­tions in the July 2019 round:

  1. 80,000 Hours - $200,000

  2. EA Com­mu­nity Build­ing Grants - $120,000

  3. HIPE (High Im­pact Policy Eng­ine) - $40,000

  4. Gen­er­a­tion Pledge - $30,000

  5. EA Coach­ing - $23,000

  6. EA Sur­vey (via CEA & Re­think Char­ity) - $20,000

  7. Re­think Pri­ori­ties - $12,000

  8. RC For­ward - $11,000

  9. Effec­tive Th­e­sis - $10,000

In this round, we saw a sig­nifi­cant in­crease in the num­ber of promis­ing fund­ing ap­pli­ca­tions we re­ceived from early-stage or­ga­ni­za­tions. We de­cided to al­lo­cate over half of the to­tal amount granted to these rel­a­tively early-stage groups. We think there is sig­nifi­cant value in en­courag­ing early-stage ini­ti­a­tives, and we hope that our grant al­lo­ca­tions sig­nal that we see these groups as par­tic­u­larly promis­ing.

This said, it is im­por­tant that donors un­der­stand this is traded off with in­creased un­cer­tainty. The main value we see in mak­ing grants to promis­ing early-stage groups is ex­per­i­men­tal value: the in­for­ma­tion gained from ex­plor­ing a new idea. We ex­pect the ex­per­i­men­tal value to be greater than the di­rect im­pact of the fund­ing.

If there is a meta ini­ti­a­tive that you would like us to con­sider for a fu­ture grant, please com­plete this form. 7 of the 9 grantees in this round ap­plied through this pro­cess.

Below are some of the key con­sid­er­a­tions be­hind our grant de­ci­sions. As with the pre­vi­ous rounds, these sum­maries are by no means meant to be read as a com­plete or ex­haus­tive case for (or against) each grant. They are based on a se­ries of con­ver­sa­tions be­tween the fund man­agers, in­cor­po­rat­ing our past ex­pe­rience, knowl­edge and judge­ment. While risks and reser­va­tions for these or­ga­ni­za­tions have been taken into ac­count, we do not dis­cuss them be­low.

If you would like to dis­cuss our de­ci­sion-mak­ing pro­cess with us fur­ther, please com­plete this form and we will put you in touch with the ap­pro­pri­ate fund man­ager.

(1) 80,000 Hours - $200,000

80,000 Hours aims to solve the world’s most press­ing prob­lems by get­ting more tal­ented peo­ple work­ing on them. To do this, they carry out re­search into how tal­ented in­di­vi­d­u­als can max­i­mize the im­pact of their ca­reers, pro­duce on­line ad­vice, iden­tify read­ers who might be able to en­ter pri­or­ity ar­eas, and provide these read­ers with free in-per­son ad­vice and con­nec­tions to men­tors, job open­ings, and fund­ing.

Cat­e­gories: Ta­lent lev­er­age, scale-stage

We have made grants to 80,000 Hours in two pre­vi­ous rounds, and we con­tinue to be­lieve that they are one of the high­est im­pact-per-dol­lar meta or­ga­ni­za­tions. Please see the pre­vi­ous pay­out re­ports in­clud­ing our dis­cus­sion on 80,000 Hours here and here.

There are a num­ber of rea­sons we de­cided to write a grant to 80,000 Hours in each of three con­sec­u­tive fund­ing rounds. In ad­di­tion to the points dis­cussed in pre­vi­ous rounds:

  • We are ex­cited to see that 80,000 Hours is be­gin­ning to grow their head­count and op­er­a­tions at a greater speed. They are plan­ning to hire 5 FTE over 2019 and an­other 5 over 2020 (they cur­rently have 13 FTE equiv­a­lent). Although they have had good rea­sons for grow­ing cau­tiously over the years, we are in favour of their scal­ing up and are keen to en­sure they have suffi­cient fund­ing to ex­pand.

  • Their fund­ing short­fall from their end-of-2018 fundraiser is $400,000, de­spite it be­ing July. This is largely due to their in­creased growth bud­get, but they have also not found suffi­cient new large donors to cover their growth. Their fund­ing gap for 2020 is $1.3m. Right now, they can­not fully com­mit to hiring in 2020 as their ex­pan­sion bud­get has not been filled. Ideally, they would already be search­ing for those hires, so they are be­ing some­what slowed down by their lack of fund­ing.

  • We think their head­hunt­ing func­tion, which has been grow­ing in re­cent months, is an ex­cit­ing new fo­cus area with sig­nifi­cant po­ten­tial for im­pact. 80,000 Hours launched this func­tion in late 2018 to help key or­ga­ni­za­tions in their pri­or­ity paths re­cruit for im­por­tant po­si­tions. The head­hunt­ing fea­ture has helped make 8 place­ments at key or­ga­ni­za­tions to date and we are keen to see it grow fur­ther. Dis­claimer: start­ing in Septem­ber, Peter McIn­tyre, who runs head­hunt­ing at 80,000 Hours, will join the EA Meta Fund team. He is re­cused from propos­ing, ad­vo­cat­ing for, or ap­prov­ing grants to 80,000 Hours.

For more in­for­ma­tion on how 80,000 Hours works and tracks their progress, please see their 2018 An­nual Re­view here.

(2) EA Com­mu­nity Build­ing Grants - $120,000

Effec­tive Altru­ism Com­mu­nity Build­ing Grants is a pro­ject run by the Cen­tre for Effec­tive Altru­ism. They write grants (typ­i­cally for $5,000-$100,000) to in­di­vi­d­u­als and groups work­ing on grow­ing effec­tive al­tru­ism within high-po­ten­tial com­mu­ni­ties. They have a par­tic­u­lar em­pha­sis on fund­ing groups aiming to tran­si­tion from be­ing run by vol­un­teers to be­ing run by full-time, paid or­ga­niz­ers.

Cat­e­gories: Ta­lent lev­er­age, early-stage

This is an early-stage grant; we ex­pect the ex­per­i­men­tal value to be greater than the di­rect im­pact of the grant.

  • In our ex­pe­rience, the ca­reer-fo­cused as­pects of the effec­tive al­tru­ism com­mu­nity play a highly valuable role in filling skill gaps in high-im­pact and tal­ent-con­strained cause ar­eas. How­ever, it is of­ten tricky to at­tribute out­comes from these com­mu­ni­ties to any par­tic­u­lar or­ga­ni­za­tion, and, po­ten­tially as a re­sult, it ap­pears there is a short­fall of fund­ing for these pro­jects.

  • We have looked into a num­ber of com­mu­nity pro­jects in our last three grant rounds. Much of the strength of these pro­jects seems to rely on highly sub­jec­tive judge­ment calls about the cal­ibre of the team and what they are aiming to achieve. There are also no clear KPIs for us to use across all pro­jects. It seems likely that de­ci­sions are best made from a top-down per­spec­tive within the con­text of other com­mu­nity efforts and op­por­tu­ni­ties.

  • We think that EA Com­mu­nity Build­ing Grants (EA CBG) is the best-placed group to eval­u­ate com­mu­nity pro­jects from this per­spec­tive. The pro­gram seems well-man­aged: they ap­pear to op­er­ate a thor­ough as­sess­ment pro­cess and to have a strong short­list for their next grant round.

  • While the pro­gram is rel­a­tively new, hav­ing been launched 18 months ago, the team has car­ried out some ini­tial im­pact as­sess­ment; so far, the re­sults seem to be pos­i­tive. In the fu­ture, if we con­tinue to sup­port EA CBG, we plan to dig deeper into their in­di­vi­d­ual grant out­comes and to fur­ther dis­cuss the eval­u­a­tion pro­cess and ev­i­dence of im­pact with the CBG team.

As part of our de­ci­sion to write this grant, we have referred all our high­est-po­ten­tial com­mu­nity build­ing grants to EA CBG. In pre­vi­ous grant rounds, we referred grant ap­pli­ca­tions from lo­cal groups to EA CBG, but funded some one-off pro­jects run by lo­cal groups. In this round, we re­ceived sev­eral ap­pli­ca­tions that did not clearly fall into ei­ther cat­e­gory (e.g. cowork­ing spaces and other longer-term pro­jects). We think these kinds of pro­jects are bet­ter eval­u­ated by EA CBG, and we will re­fer all such pro­jects to them go­ing for­ward. We ex­pect to still con­sider fund­ing for one-off pro­jects run by lo­cal groups in fu­ture rounds.

(3) HIPE (High Im­pact Policy Eng­ine) - $40,000

HIPE is a high-im­pact ca­reers net­work within the UK gov­ern­ment civil ser­vice. The group aims to help civil ser­vants max­i­mize the so­cial im­pact of their ca­reers, through helping them to iden­tify gov­ern­ment roles where they can do the most good, build rele­vant skills, and con­nect with like-minded and mo­ti­vated gov­ern­ment em­ploy­ees.

Cat­e­gories: Ta­lent lev­er­age, early-stage

HIPE performs re­search into how civil ser­vants can choose high-im­pact ca­reer paths, does out­reach within the civil ser­vice to dis­sem­i­nate key ideas, and pro­vides in-depth coach­ing for highly en­gaged and promis­ing civil ser­vants. Be­cause this model ap­pears to have been suc­cess­ful for 80,000 Hours, we are keen to see it be­ing ex­per­i­mented with in other ar­eas.

To date, HIPE has been run by vol­un­teers within the civil ser­vice alongside their day-to-day roles. There is now an op­por­tu­nity to cre­ate a new role within a UK gov­ern­ment de­part­ment for a gov­ern­ment em­ployee to work on HIPE full-time. We be­lieve fund­ing at this stage is par­tic­u­larly valuable to al­low HIPE to cap­i­tal­ize on this op­por­tu­nity. HIPE is fundrais­ing to cover salary and travel costs for one full-time em­ployee for 2 years.

As well as fo­cus­ing on HIPE’s ex­ist­ing re­search, out­reach, and ad­vi­sory work, the new full-time hire would sup­port a broader team of civil ser­vants to work on HIPE and mea­sure its im­pact. We be­lieve this im­pact mea­sure­ment will be highly valuable in both di­rect­ing HIPE’s fu­ture work and sup­port­ing their fu­ture fundrais­ing.

If HIPE can demon­strate value to the UK gov­ern­ment de­part­ment (e.g., through im­prov­ing policy-mak­ing, staff wellbe­ing, or staff re­ten­tion), HIPE be­lieves they would have a rea­son­ably strong chance of be­ing made a per­ma­nent pro­ject fully funded by the gov­ern­ment.

As the key hire can only be made once fund­ing is se­cured due to gov­ern­ment reg­u­la­tions, the mem­bers of the cen­tral HIPE team go­ing for­wards are still to be de­ter­mined, while the two origi­nal found­ing vol­un­teers are pri­mar­ily fo­cused on other pro­jects. This grant is made ex­per­i­men­tally, largely on the promise of the idea and the rep­u­ta­tion of the two found­ing vol­un­teers. Any fur­ther fund­ing will be sen­si­tive to the strength of their team in the fu­ture.

(4) Gen­er­a­tion Pledge - $30,000

Gen­er­a­tion Pledge aims to di­rect more philan­thropic cap­i­tal to the world’s most press­ing prob­lems, through build­ing a com­mu­nity of fu­ture heirs from ul­tra high net worth fam­i­lies (‘next gens’). They in­vite next gens to sign the ‘gen­er­a­tion pledge’, to donate a per­centage of their in­her­i­tance for so­cial im­pact, and sup­port pledgers to de­cide where to give.

Cat­e­gories: Cap­i­tal lev­er­age, early-stage

This is an early-stage grant; we ex­pect the ex­per­i­men­tal value to be greater than the di­rect im­pact of the grant.

Gen­er­a­tion Pledge mir­rors the Founders Pledge model, but work­ing with next gens rather than founders. The world’s ul­tra high net worth fam­i­lies col­lec­tively own $31.5 trillion. Sev­eral thou­sand next gens will in­herit this wealth. Gen­er­a­tion Pledge aims to sup­port those next gens to max­imise their so­cial im­pact.

We have pre­vi­ously con­sid­ered Gen­er­a­tion Pledge for a grant, but we weren’t con­fi­dent enough that their model would work suc­cess­fully. How­ever, in the past few months they have grown their pledger com­mu­nity sig­nifi­cantly. They have an ex­pected pledge value of over $300 mil­lion, with a num­ber of sen­si­ble dis­count fac­tors ap­plied.

We think that these pos­i­tive up­dates are enough to jus­tify Gen­er­a­tion Pledge be­ing funded through to a later stage, where they will have the op­por­tu­nity to prove they can turn these pledges into dona­tions to effec­tive char­i­ties.

This grant will con­tribute to­wards Gen­er­a­tion Pledge’s im­me­di­ate fund­ing gap for 2019, giv­ing them more run­way to fundraise and fur­ther grow their pledger base.

(5) EA Coach­ing - $23,000

Through EA Coach­ing, Lynette Bye works with clients at high-im­pact or­ga­ni­za­tions to help them im­prove their pri­ori­ti­za­tion, im­ple­ment bet­ter strate­gies, and in­crease fo­cused work time.

Cat­e­gories: Ta­lent lev­er­age, early-stage

This is an early-stage grant; we ex­pect the ex­per­i­men­tal value to be greater than the di­rect im­pact of the grant.

While fund­ing a pro­duc­tivity coach may not be the most in­tu­itive grant de­ci­sion for many donors, the ba­sic premise seems rea­son­ably clear. For a small in­vest­ment in time and money, pro­duc­tivity coach­ing could re­sult in long-term pos­i­tive changes to the out­put of in­di­vi­d­u­als already hav­ing a sig­nifi­cant im­pact with their ca­reer.

Lynette’s own im­pact eval­u­a­tion can be found here. Given the early-stage na­ture of her pro­ject, we found the re­sults fairly com­pel­ling. A num­ber of her clients work­ing at high-im­pact or­ga­ni­za­tions have re­ported sig­nifi­cant in­creases in their hours of pro­duc­tive work. Even if clients have achieved only a small frac­tion of the re­ported in­crease in pro­duc­tivity, this would likely be enough to ‘break even’ on the cost of coach­ing. That said, our in­tu­ition is to be cau­tious about lean­ing too heav­ily on nu­mer­i­cal rea­son­ing for pro­jects of this type.

Lynette fo­cuses on clients work­ing in AI al­ign­ment and at high-im­pact “meta” or­ga­ni­za­tions. She has pre­vi­ously worked with em­ploy­ees at FHI, the Open Philan­thropy Pro­ject, CEA, CHAI, MIRI, Deep­Mind, and the Forethought Foun­da­tion, and she ex­pects to con­tinue to do so. Given that these or­ga­ni­za­tions fo­cus on par­tic­u­larly high-im­pact ar­eas, we ex­pect that in­creas­ing their pro­duc­tivity should be very valuable.

This grant will al­low Lynette to offer coach­ing calls to peo­ple work­ing at high-im­pact or­ga­ni­za­tions at a highly sub­si­dized rate, offer free coach­ing for se­lect refer­rals from 80,000 Hours, and hire con­trac­tors to help cre­ate ma­te­ri­als to scale her coach­ing.

(6) EA Sur­vey (CEA & Re­think Char­ity) - $20,000

The EA Sur­vey pro­vides an an­nual snap­shot of the EA com­mu­nity. The sur­vey acts as a bench­mark for bet­ter un­der­stand­ing the com­mu­nity, in­clud­ing: de­mo­graph­ics and char­ac­ter­is­tics; how peo­ple first get in­volved; where peo­ple first hear about EA; in­fluences on in­volve­ment with the com­mu­nity; dona­tion data; cause se­lec­tion; and growth met­rics.

Cat­e­gories: In­for­ma­tion lev­er­age, spe­cific one-off

While it is difficult to quan­tify how use­ful the sur­vey has been, it does seem like it is valuable to gen­er­ate more em­piri­cal data to in­form move­ment-build­ing strat­egy. In par­tic­u­lar, we are aware that some high-im­pact or­ga­ni­za­tions con­sis­tently col­lab­o­rate with the re­searchers who an­a­lyze the sur­vey, in or­der to in­form their or­ga­ni­za­tional and com­mu­nity-build­ing strate­gies.

There are a limited num­ber of ways of col­lect­ing mean­ingful data that tracks the broad effec­tive al­tru­ism move­ment. Main­tain­ing a suffi­ciently large sur­vey and in-depth anal­y­sis seems like one of the bet­ter meth­ods to gen­er­ate data and in­sights of this kind.

We think part of the benefit of the sur­vey is en­sur­ing data that is col­lected on a reg­u­lar ba­sis to al­low for con­sis­tent year-on-year com­par­i­sons. Fund­ing at this stage will al­low the sur­vey and anal­y­sis to be car­ried out this year in line with the an­nual pub­lish­ing cy­cle.

His­tor­i­cally, the sur­vey has been run by Re­think Char­ity. $15,000 of this grant will go to Re­think to de­sign and run an up­dated 2019 sur­vey. Re­think has been col­lab­o­rat­ing with CEA and other meta or­ga­ni­za­tions to gather in­put on the 2019 sur­vey. Like last year, Re­think will re­lease anonymized data and sum­mary sur­vey out­put. $5,000 of the grant will go to CEA, which will use the fund­ing to pay Re­think for be­spoke anal­y­sis of the re­sults.

(7) Re­think Pri­ori­ties - $12,000

Re­think Pri­ori­ties is a cause pri­ori­ti­za­tion re­search group that fo­cuses on ne­glected cause ar­eas. Their re­search agenda is cur­rently fo­cused on how to ap­ply cost-effec­tive­ness frame­works to un­cer­tain do­mains, in­ter­ven­tions aimed at an­i­mal welfare, and the growth of the EA move­ment. Re­think Pri­ori­ties is a pro­ject of Re­think Char­ity.

Cat­e­gories: In­for­ma­tion lev­er­age, early-stage

This is an early-stage grant; we ex­pect the ex­per­i­men­tal value to be greater than the di­rect im­pact of the grant.

We made a grant to Re­think Pri­ori­ties in our pre­vi­ous grant round, and we be­lieve the rea­sons be­hind this grant still stand. You can read the pre­vi­ous write-up here.

We noted in the pre­vi­ous round that we ex­pect there is par­tic­u­lar value in Re­think Pri­ori­ties un­der­tak­ing com­mis­sioned re­search into ar­eas that are ne­glected by other re­searchers. In this round, Re­think Pri­ori­ties ap­plied for fund­ing to sup­port two spe­cific re­search pro­jects, both of which were sug­gested by re­searchers at other high-im­pact or­ga­ni­za­tions, and re­quire col­lab­o­ra­tion with those re­searchers. We view this as an in­di­ca­tion that these pro­jects will provide value.

The first re­search pro­ject would fo­cus on the geopoli­ti­cal im­pli­ca­tions of cli­mate change on mass mi­gra­tion. Re­think Pri­ori­ties has an ini­tial draft, and now has the op­por­tu­nity to ex­pand the re­search un­der the men­tor­ship of Luke Kemp, a re­search as­so­ci­ate at the Cen­tre for the Study of Ex­is­ten­tial Risk at Cam­bridge. The sec­ond pro­ject would eval­u­ate the cost-effec­tive­ness of the Treaty on the Pro­hi­bi­tion of Nu­clear Weapons, at the sug­ges­tion of Carl Shul­man, a re­search as­so­ci­ate at the Fu­ture of Hu­man­ity In­sti­tute at Oxford.

We think both of these pro­jects sound po­ten­tially highly valuable and demon­strate that Re­think Pri­ori­ties is un­der­tak­ing com­mis­sioned re­search into ne­glected ar­eas. We de­cided to make this grant as un­re­stricted fund­ing, rather than di­rect­ing it to one of the pro­jects speci­fi­cally, to give Re­think Pri­ori­ties the flex­i­bil­ity to de­cide on their cur­rent high­est-pri­or­ity re­search.

(8) RC For­ward - $11,000

RC For­ward is a dona­tion plat­form through which Cana­di­ans can make tax-ad­van­taged dona­tions to high-im­pact char­i­ties lo­cated in and out­side of Canada. RC For­ward is a pro­ject of Re­think Char­ity.

Cat­e­gories: Cap­i­tal lev­er­age, early-stage

This is an early-stage grant, we ex­pect the ex­per­i­men­tal value to be greater than the di­rect im­pact of the grant.

RC For­ward seems to fill a valuable niche for Cana­dian donors. In 2018, the first year in which the pro­ject was run, RC For­ward moved net $3.3m to 25 effec­tive char­i­ties on a bud­get of ~$50,000.

Re­think Pri­ori­ties has car­ried out a cost-effec­tive­ness anal­y­sis of RC For­ward, which is linked here. While this anal­y­sis re­lies on a num­ber of es­ti­mates from donors about how the plat­form has changed their be­havi­our, RC For­ward still looks to be a rel­a­tively strong dona­tion op­por­tu­nity even when ap­ply­ing the most pes­simistic es­ti­mates. In 2018, RC For­ward is ex­pected to have moved at least $3 for ev­ery $1 spent.

That said, RC For­ward does not ad­vise donors on where to give, in­stead pro­vid­ing a ser­vice that (in some cases greatly) in­creases tax sav­ings and con­ve­nience for Cana­dian donors who are already mo­ti­vated to give to highly effec­tive char­i­ties. This makes in­ves­ti­gat­ing the coun­ter­fac­tual much more challeng­ing for RC For­ward. While this weak­ens the case for the pro­ject, the ‘break-even point’ is still very low. Even if only 2% of the to­tal money moved through the plat­form last year was di­rectly caused by RC For­ward, they would beat the break-even point.

Although any po­ten­tial scale-up is limited to the Cana­dian mar­ket, the plat­form could pro­cess many more dona­tions within that mar­ket with­out a pro­por­tional rise in costs. We think this is a promis­ing as­pect of the plat­form and are keen to see it grow within Canada.

This grant will fill RC For­ward’s im­me­di­ate re­main­ing fund­ing gap for 2019.

(9) Effec­tive Th­e­sis - $10,000

Effec­tive Th­e­sis ad­vises stu­dents on choos­ing an im­pact­ful re­search topic for their the­sis by con­nect­ing them with ex­pe­rienced re­searchers. The goal of the pro­ject is to en­courage ju­nior re­searchers to re­search high-im­pact top­ics and con­tinue pro­duc­ing valuable, rigor­ous re­search through­out their aca­demic ca­reers.

Cat­e­gories: Ta­lent lev­er­age, in­for­ma­tion lev­er­age, early stage

This is an early-stage grant, we ex­pect the ex­per­i­men­tal value to be greater than the di­rect im­pact of the grant.

Ad­vice on a stu­dent’s choice of the­sis topic has the po­ten­tial to in­fluence where they spend hun­dreds of hours of re­search time (300 hours per stu­dent on av­er­age). Topic choice could also have some im­pact on fu­ture ca­reer steps or fu­ture re­search fo­cus, al­though this is more challeng­ing to mea­sure.

Pro­vid­ing stu­dents with the­sis topic ad­vice re­quires a small level of in­put from an ex­pe­rienced re­searcher, usu­ally 1-3 hours per stu­dent. Given the small in­put for po­ten­tially large out­puts, we think this is an in­ter­est­ing meta ini­ti­a­tive worth ex­plor­ing fur­ther.

Effec­tive Th­e­sis has a net­work of 44 coaches with the ca­pac­ity to take on at least 2x as many stu­dents as they help now. As find­ing high-cal­iber coaches with enough available time seems to be the most sig­nifi­cant hur­dle for many coach­ing pro­jects, we are im­pressed that Effec­tive Th­e­sis seems to have made good progress in this area.

Since Septem­ber 2018, Effec­tive Th­e­sis has re­ceived over 200 ap­pli­ca­tions from stu­dents who want help with figur­ing out their the­sis topic and pro­vided 60 stu­dents with coach­ing. We share their ex­pec­ta­tion that a small num­ber of cases likely ac­count for the ma­jor­ity of the pro­ject’s im­pact. The best cases so far ap­pear to have been cases where stu­dents—who already planned to be­come re­searchers in the long-term—sub­stan­tially changed their fo­cus to higher-im­pact top­ics pri­mar­ily be­cause of their work with Effec­tive Th­e­sis.