Some Organisational Changes at the Centre for Effective Altruism

This post ex­plains a num­ber of re­cent changes re­gard­ing the or­gani­sa­tional struc­ture of the Cen­tre for Effec­tive Altru­ism (CEA). Th­ese are pri­mar­ily changes to the in­ter­nal struc­ture of the or­gani­sa­tion; for that rea­son, this post should be pri­mar­ily of in­ter­est to peo­ple who fol­low CEA very closely. How­ever, I also ex­plain briefly how this might bear on the ex­ter­nally-fac­ing ac­tivi­ties CEA does in the fu­ture. 
In sum­mary:
  • Up un­til re­cently, CEA ran on a ‘fed­eral’ model, as five largely au­tonomous teams: 80,000 Hours, Giv­ing What We Can, Global Pri­ori­ties Pro­ject, Effec­tive Altru­ism Outreach, and the Cen­tral team.

  • Four of these teams (EAO, GWWC, GPP and the Cen­tral team) are merg­ing into the same man­age­ment struc­ture, and from now on will op­er­ate as a sin­gle unit. 80,000 Hours will con­tinue as an au­tonomous or­gani­sa­tion, based in the Bay, though still fis­cally spon­sored by CEA.

  • CEA will there­fore act much more like a unified or­gani­sa­tion than it has done in the past. This newly-defined CEA will be led by my­self (Will MacAskill); I will there­fore be play­ing a much more ac­tive role in the man­age­ment of CEA than I have in the past.


In the past, CEA has been the um­brella or­gani­sa­tion for a col­lec­tion of differ­ent non­prof­its. As of the end of last year, these con­sisted of: Giv­ing What We Can, 80,000 Hours, the Global Pri­ori­ties Pro­ject (in col­lab­o­ra­tion with the Fu­ture of Hu­man­ity In­sti­tute), and Effec­tive Altru­ism Outreach. There was also a Cen­tral team, which did shared op­er­a­tions for all the other pro­jects.

The ar­gu­ments for run­ning CEA as a col­lec­tion of sep­a­rate or­gani­sa­tions in­cluded the fol­low­ing:
  • By ex­per­i­ment­ing with differ­ent ideas, cul­tures, and ap­proaches, we could learn what worked best and fo­cus on that.

  • We al­low differ­ent pro­jects to pro­mote mes­sages that may differ in tone, con­tent, or em­pha­sis.

  • We could tar­get differ­ent or­gani­sa­tions to differ­ent de­mo­graph­ics.

  • Some­times, run­ning sep­a­rate pro­jects would in­crease our available re­sources. (For ex­am­ple, a donor might only be in­ter­ested in one pro­ject).

For ex­am­ple, in 2011 we de­cided to set up 80,000 Hours as a dis­tinct or­gani­sa­tion from GWWC be­cause:
  • We wor­ried that 80k might be­come a lot more con­tro­ver­sial than GWWC, be­cause of the idea of earn­ing to give, and we wanted to keep these mes­sages sep­a­rate.

  • Some peo­ple felt much more ex­cited by the po­ten­tial im­pact of 80k, whereas oth­ers felt more ex­cited by the po­ten­tial im­pact of GWWC.

  • We thought that sorts of peo­ple who would be in­ter­ested in so­cially-ori­ented ca­reer ad­vice was very differ­ent from sorts of peo­ple who were in­ter­ested in donat­ing.

  • Hav­ing sep­a­rate or­gani­sa­tions al­lowed each team to fo­cus en­tirely on their own pro­ject.

The found­ing of new pro­jects within CEA then pro­gressed on the model we’d set up with 80,000 Hours and GWWC.
Re­cently — due in sig­nifi­cant part to changes in our situ­a­tion — we started to be­come less con­vinced that the fed­eral model was the op­ti­mal way to struc­ture all the differ­ent re­la­tion­ships be­tween the differ­ent pro­jects in CEA.
  • The rea­sons in favour, though still pre­sent, in some cases be­came weaker:

  • We felt that it’s eas­ier to quickly ex­per­i­ment and scale up or scale down pro­jects within one or­gani­sa­tional struc­ture, rather than when those pro­jects are run as sep­a­rate or­gani­sa­tions.

  • We found that the differ­ent mes­sages nat­u­rally came to be closely as­so­ci­ated. Be­cause of the rise of ‘effec­tive al­tru­ism’ as a term, peo­ple would re­fer to CEA as a sin­gle or­gani­sa­tion (de­spite our ini­tial in­ten­tions of this or­gani­sa­tional la­bel not to be pub­lic-fac­ing).

  • Be­cause we now have much greater ac­cess to fund­ing and po­ten­tial em­ploy­ees than we did pre­vi­ously, the ar­gu­ment from ad­di­tional re­sources no longer has the same force.

More­over, there were some cases where the differ­ent or­gani­sa­tions led to con­fu­sion:
  • Some­times differ­ent or­gani­sa­tions ended up run­ning similar pro­jects: for ex­am­ple, EA Build (run by EAO) and GWWC were both try­ing to help grow EA lo­cal groups. This cre­ated con­fu­sion for staff, donors, for EA com­mu­nity mem­bers, and for third par­ties in­ter­ested in the work of CEA.

  • In­ter­nally, there would of­ten be de­ci­sions that af­fected all the differ­ent or­gani­sa­tions. Th­ese de­ci­sions would be made by the “Se­nior Man­age­ment Team”, con­sist­ing of lead­ers of all the differ­ent or­gani­sa­tions. This de­ci­sion-mak­ing pro­cess of­ten felt slow, bu­reau­cratic and un­nec­es­sar­ily com­pli­cated.

Fi­nally, there was an un­usual op­por­tu­nity for this change to hap­pen. I had sig­nifi­cantly more free time as a re­sult of (i) the launch of the book and sub­se­quent me­dia ac­tivity dy­ing down; (ii) rene­go­ti­at­ing my con­tract with the Univer­sity, re­sult­ing in a much lower teach­ing load. This gave us the un­usual op­por­tu­nity for CEA to be led by some­one with many years of ex­pe­rience work­ing with each of the in­di­vi­d­ual or­gani­sa­tions within CEA. At the same time, lead­er­ship of the differ­ent or­gani­sa­tions within CEA felt ex­cited by the prospect of be­ing able to work to­gether and rally around a sin­gle shared vi­sion.

What changes are hap­pen­ing at CEA

We’re unify­ing the teams that com­pose GWWC, EAO, GPP and CEA Cen­tral. We’ll di­vide CEA into a Com­mu­nity & Outreach Divi­sion and a Spe­cial Pro­jects Divi­sion. The Com­mu­nity and Outreach Divi­sion will fo­cus on the ‘core’ CEA ac­tivity, which is helping to grow and strengthen the EA com­mu­nity. This in­cludes our on-line pres­ence, lo­cal groups, EA Global, EAGx, me­dia, mar­ket­ing, and the Giv­ing What We Can Trust. The Spe­cial Pro­jects Divi­sion will have three aims: high net worth philan­thropic ad­vis­ing, policy, and fun­da­men­tals (ex­plained more be­low); the former is a con­tinu­a­tion of the re­search arm of GWWC; the lat­ter two are con­tinu­a­tions of the two aims of GPP, sep­a­rated into differ­ent teams.
80k will still op­er­ate in­de­pen­dently; and we would en­courage peo­ple to re­gard 80k as sep­a­rate en­tity from CEA (though CEA will re­main as a fis­cal spon­sor of 80k). The case for also merg­ing the 80k team with the other teams seemed weaker for a num­ber of rea­sons: 80k was already by far the most au­tonomous of the or­gani­sa­tions un­der CEA’s um­brella; and it had not faced the is­sue of over­lap with other or­gani­sa­tions within CEA.

I am tak­ing on the role of CEO of CEA (in ad­di­tion to be­ing a Trus­tee). This means I will spend far more time guid­ing and man­ag­ing CEA than I have done in the past. Tara Mac Au­lay will con­tinue as COO and lead the Com­mu­nity & Outreach Divi­sion. Michael Page, a new hire, will lead the Spe­cial Pro­jects Divi­sion. Kerry Vaughan will con­tinue to work on EA com­mu­nity-build­ing and Seb Far­quhar will con­tinue to work on policy. Michelle Hutch­in­son will tran­si­tion from run­ning GWWC to helping to set up an Oxford In­sti­tute for Effec­tive Altru­ism (ex­plained more be­low).

How will that change what you see from CEA?

In the short term, not all that much. We’ll con­tinue to pro­mote effec­tive giv­ing un­der the ban­ner of ‘Giv­ing What We Can’, keep­ing its pledge, web­site etc. The main change for it is that peo­ple will work on it from within a unified team rather than in its own, siloed team. We’re rea­son­ably likely to de­pri­ori­tise or dis­con­tinue the GPP la­bel, and we will not con­tinue with the EAO la­bel.

In the mid term, our aims in­clude the fol­low­ing:
  • Devel­op­ment of and greater fo­cus on effec­tivealtru­

  • Greater fo­cus on un­der­stand­ing the de­sires of the EA com­mu­nity, and us­ing that as an in­put to de­ci­sions about what pro­jects to try or pri­ori­tise.

  • Greater fo­cus on in­creas­ing the im­pact of those who already self-iden­tify as EAs and on in­creas­ing the po­ten­tial benefits of EA as a com­mu­nity.

  • Greater fo­cus on:

  • Cause neu­tral­ity (rather than a fo­cus on global poverty)

  • Means neu­tral­ity (rather than a fo­cus on dona­tion)

  • Greater fo­cus on in­tel­lec­tual de­vel­op­ment of EA, in­clud­ing on high-level the­ory.

As a de­fault, we plan to con­tinue to grow at at least the rate that we’ve grown in the past (which has meant dou­bling in size ap­prox­i­mately ev­ery 18 months).
Some other changes in­clude the fol­low­ing:

Lo­cal Groups

We’ll en­courage peo­ple to call new groups an “EA lo­cal group” rather than a “GWWC lo­cal group”. (We cur­rently give peo­ple the choice and in most cases lo­cal group founders choose to re­fer to them­selves as EA lo­cal groups.) How­ever, if peo­ple want to run as a GWWC group and not an EA group, we won’t force the is­sue.

Char­ity Research

Our main fo­cus within char­i­ta­ble re­search will be ex­per­i­ment­ing with a new pro­ject: bou­tique philan­thropic ad­vice to ma­jor donors. We’ve found that there is no­table de­mand for ev­i­dence-based char­ity re­search from ma­jor donors who, for a va­ri­ety of rea­sons, do not want sim­ply to sup­port GiveWell’s top char­i­ties or to work with Open Philan­thropy. (Often, the donor is in­ter­ested in a par­tic­u­lar cause area, such as dis­aster pre­pared­ness, that is differ­ent from the work of GiveWell’s top char­i­ties.)
We’ve already been ex­per­i­ment­ing with this pro­ject over the last six months. Peo­ple we’ve pro­vided ad­vice for in­clude: en­trepreneurs who have taken the Founders’ Pledge and ex­ited; pri­vate ma­jor donors who con­tacted us as a re­sult of read­ing Do­ing Good Bet­ter; former Prime Minister Gor­don Brown, for his In­ter­na­tional Com­mis­sion on Fi­nanc­ing Global Ed­u­ca­tion Op­por­tu­nity; and Alwaleed Philan­thropies, a $30 billion foun­da­tion fo­cused on global hu­man­i­tar­i­anism. This pro­ject is still very much in its in­fancy and we’ll as­sess its de­vel­op­ment on an on­go­ing ba­sis. 
Within global health and de­vel­op­ment, we will move to sim­ply recom­mend­ing GiveWell’s top char­i­ties, rather than cu­rat­ing an in­de­pen­dent but over­lap­ping list of recom­mended char­i­ties based in large part on their re­search (as we do now). In the past, the ex­is­tence of two similar lists of recom­mended char­i­ties has cre­ated con­fu­sion, and we feel that the amount of value to be gained from do­ing work so similar to GiveWell’s is com­par­a­tively small rel­a­tive to our other re­search op­por­tu­ni­ties. 
It’s pos­si­ble we may, in ad­di­tion, point read­ers to char­i­ties in cause ar­eas out­side of global health and de­vel­op­ment, with the caveat that such char­i­ties will not have had the same level of as­sess­ment as GiveWell’s top char­i­ties.


Both our policy work and our fun­da­men­tals re­search will con­tinue the work done by GPP, though now split into two sep­a­rate teams.
We think that policy is an im­por­tant area for effec­tive al­tru­ism to de­velop into, and we feel we have had some sig­nifi­cant suc­cess within policy so far. Re­cent de­vel­op­ments in Bri­tish poli­tics mean that our plans re­gard­ing our policy work are cur­rently in flux; de­pend­ing on how this plays out, we could do con­sid­er­ably more or con­sid­er­ably less policy work.

Fun­da­men­tals Research

Partly due to de­mand from some mem­bers of the EA com­mu­nity, we’ll be ex­per­i­ment­ing with do­ing more the­o­ret­i­cal re­search on effec­tive al­tru­ism. This comes in two main cat­e­gories: ‘cru­cial con­sid­er­a­tions,’ or ideas that have the po­ten­tial to rad­i­cally change how we eval­u­ate our op­tions; and ‘cause pri­ori­ti­sa­tion’, or re­search on how to figure out which cause-ar­eas one ought to fo­cus on. We be­lieve that this work is both ex­tremely im­por­tant and ex­tremely hard to do, and will as­sess our progress on this front on an on­go­ing ba­sis.

Oxford In­sti­tute for Effec­tive Altruism

We have plans to set up an aca­demic in­sti­tute fo­cused on effec­tive al­tru­ism, based at Oxford Univer­sity. We hope that it can be­gin as of Oc­to­ber 2017, though this is con­tin­gent on suc­cess­ful grant ap­pli­ca­tions. The In­sti­tute will work on the­o­ret­i­cal is­sues that arise from the pro­ject of try­ing to do the most good, strad­dling philos­o­phy, eco­nomics, and other rele­vant fields, pro­duc­ing re­search that is suit­able for pub­li­ca­tion in aca­demic jour­nals. The aim is for this to be run by Hilary Greaves as Re­search Direc­tor and Michelle Hutch­in­son as Oper­a­tions Direc­tor. We be­lieve that this rep­re­sents an ex­cit­ing op­por­tu­nity to help cre­ate and shape effec­tive al­tru­ism as an aca­demic re­search field.
We’ll write more about our plans, in­clud­ing elab­o­rat­ing on some of the pro­jects listed above, in the near fu­ture.