EA Survey 2018 Series: Donation Data

Summary

  • Me­dian dona­tions were slightly higher than in 2016 and to­tal dona­tions much higher

  • A small num­ber of very large donors ac­count for the ma­jor­ity of the to­tals donated

  • A ma­jor­ity of EAs re­port donat­ing less than they would like due to fi­nan­cial constraints

This post ex­plores dona­tion data in the 2018 EA Sur­vey, in­ves­ti­gat­ing how much peo­ple are donat­ing, where they are donat­ing and what in­fluences their dona­tions.

1891 out of 2607 (73%) self-iden­ti­fied EAs in our sam­ple offered data about their dona­tions. This is a sig­nifi­cant in­crease from the 2017 Sur­vey where we had dona­tion data from 1019 EAs out of 1853 (54.9%).

To­tals Donated

As in pre­vi­ous years, there was a very wide range in amounts donated (note the 2018 Sur­vey col­lects re­ports on amounts donated in 2017). All amounts are in USD ($).

This marks a sig­nifi­cant in­crease in to­tal amounts donated and me­dian dona­tions com­pared to 2015 and 2016 data.

As the graph be­low shows, a very small num­ber of very large dona­tions dwarf the size of most oth­ers.

The fol­low­ing his­togram shows the num­ber of dona­tions of differ­ent sizes be­ing made by EAs in our sam­ple (note the log x axis).

A com­menter on a pre­vi­ous EA Sur­vey post asked about the pro­por­tion of to­tal EA dona­tions that came from dona­tions of cer­tain sizes and re­quested a cu­mu­la­tive dona­tions graph (be­low). Per­haps, even if the largest dona­tions are many times larger than the rest, much of the to­tal might still be com­ing from a very large num­ber of smaller donors?

As the graph above shows, how­ever, a rel­a­tively small por­tion of to­tal dona­tions recorded in our sam­ple comes from those donat­ing smaller amounts. In­di­vi­d­u­als donat­ing $1000 or less in 2017 (47.8% of donors in our sam­ple) ac­counted for $267,942.5 of dona­tions or about 1.5% of the to­tal. Those more in the ‘mid­dle’, donat­ing be­tween $1000 and $10,000 in­clude 40% of donors, and the sub­stan­tial sum of $2,569,359.59, which is large rel­a­tive to most EA pro­jects, while still only be­ing 14% of to­tal dona­tions. Con­versely donors giv­ing more than $100,000 (0.7% of donors) ac­counted for 57% of dona­tions.

As the figures in the table above show, a dona­tion of $1000 per year (or 5% of a $20,000 salary) would place one in the top half of EA donors (speci­fi­cally, the 55th per­centile), whereas be­ing in the top 10% of donors would re­quire donat­ing $10,000 and the top 1% >$75,000.

Per­centages of In­come Donated

We also looked at the per­centages of in­come that EAs were donat­ing, based on the 1798 EAs who dis­closed both in­come and dona­tion data.[1] As in pre­vi­ous years, most EAs were donat­ing sig­nifi­cantly less than the 10% Giv­ing What We Can Pledge. How­ever, as the graph be­low shows, there is a marked ‘bump’ in the donors giv­ing at around the 10% figure, per­haps due to the Giv­ing What We Can Pledge tar­get around this amount, or due to the figure’s wider pop­u­lar­ity as a tar­get (e.g. in tithing).

The above graph shows a small num­ber of re­spon­dents re­port donat­ing more than 100% of their in­come in 2017. In all but three of these cases the sums re­ported for dona­tions and in­come were very low (less than a cou­ple of thou­sands dol­lars), so we do not view them as par­tic­u­larly im­plau­si­ble, though they may well rep­re­sent un­usual situ­a­tions.

What Ex­plains Low Dona­tions?

In pre­vi­ous years, read­ers have com­mented on the large num­bers of EAs who ap­pear to be donat­ing $0 or close to $0 and at their sur­prise at low dona­tion lev­els over­all. In ear­lier years, where dona­tions and the Giv­ing What We Can Pledge were em­pha­sised more, com­pared to now where (di­rectly) im­pact­ful ca­reer choice and up­skil­ling EAs are em­pha­sised more, this was per­haps a more a cause for sur­prise and con­cern. How­ever, it is still worth in­ves­ti­gat­ing what ex­plains low lev­els of dona­tion.

Students

It goes with­out say­ing that many EAs are stu­dents (45.2% in our sam­ple this year), who might there­fore be ex­pected to be donat­ing lower amounts.

True to ex­pec­ta­tions, we found that non-stu­dents donated sig­nifi­cantly more than stu­dents:

Employment

Similarly, many EAs in our sam­ple, though not stu­dents, may not be em­ployed or fully em­ployed. When we ex­clude both stu­dents and those who are not em­ployed full-time, me­dian dona­tions are, again, sub­stan­tially higher:

De­spite this, the per­cent donated even among full-time em­ployed non-stu­dents was very low, with a me­dian of 3.7% in­come donated. In­deed, only the 80th per­centile of full-time em­ployed non-stu­dents was donat­ing more than 10% (10.591%) with the 90th per­centile donat­ing 14.52%.

Income

The fact that dona­tions are higher among full-time em­ployed non-stu­dents may well be ex­pected to be largely due to higher in­come. In­deed, within the sam­ple as a whole, we found that dona­tions in 2017 were strongly cor­re­lated with the av­er­age of house­hold and in­di­vi­d­ual in­come in 2017 ( 0.9232, p<0.001) in the raw data. How­ever, we should note that there are some large out­liers driv­ing this very strong re­la­tion­ship.

To ac­count for full-time em­ployed non-stu­dents who were nev­er­the­less on a low in­come, we also looked at full-time em­ployed non-stu­dents who re­ported earn­ing more than $20,000 USD in 2017 (762 out of 1050 full-time em­ployed non-stu­dents, of which 747 re­ported dona­tion data).

While the me­dian dona­tion among this group was sub­stan­tially higher than for the sam­ple as a whole, the per­cent donated re­mains quite far short of the level of the Giv­ing What We Can Pledge.

Self-re­ported rea­sons for donat­ing less than desired

This year, for the first time, we asked re­spon­dents whether they were donat­ing as much or less than they wanted to.

Across the sam­ple as a whole, a ma­jor­ity (56.99%) re­ported donat­ing less than they would like to due to fi­nan­cial con­straints or some other rea­son (17.36%), whereas less than a quar­ter (24.42%) re­ported donat­ing as much as they would like to.

We also hand-coded the qual­i­ta­tive re­sponses in the “please spec­ify why” box. Re­sponses could be clas­sified as mul­ti­ple differ­ent cat­e­gories at once. Many peo­ple who se­lected this op­tion also speci­fied fi­nan­cial con­straints as a rea­son, mean­ing that the to­tal for this cat­e­gory was even higher. Most of the coded cat­e­gories are fairly self ex­plana­tory. How­ever, “So­cial” refers to so­cial in­fluence or so­cial norms and most re­sponses in this cat­e­gory ex­plic­itly referred to a spouse or part­ner. “Effi­cacy” refers to con­cerns about lo­cat­ing an effec­tive char­ity to donate to or con­cern than any char­i­ties were effec­tive dona­tion tar­gets. The graph be­low shows the fre­quency of these other rea­sons.

Other In­fluences on Giving

Career

EAs work in a va­ri­ety of ca­reers and per­haps peo­ple aren’t donat­ing money, but rather donat­ing their time or tak­ing a lower salary to work at a di­rectly im­pact­ful org. In­deed, those work­ing in re­search roles donate sub­stan­tially less than their peers and have a lower me­dian in­come.

Giv­ing What We Can Pledge

The ob­ser­va­tion that so many EAs in our sur­vey were donat­ing much less than 10% of their in­come to char­ity sug­gests the ques­tion: how much are self-re­ported GWWC Pledge tak­ers donat­ing?

Ac­cord­ing to this anal­y­sis, the me­dian per­cent of in­come donated by some­one who had taken the GWWC Pledge, in our sam­ple, was 7.22%, short of the 10% tar­get. Nev­er­the­less, this could be in­fluenced by GWWC Pledge tak­ers be­ing stu­dents, not em­ployed or only re­cently hav­ing taken the Pledge. We will be ex­plor­ing these is­sues in more de­tail in a ded­i­cated post.

Year First Heard of EA

Me­dian dona­tions, in­come and per­centage of in­come donated all in­crease fairly steadily the longer re­spon­dents have been in EA. 2010 bucks the trend slightly ow­ing to one very high in­come-high dona­tion in­di­vi­d­ual and small num­ber of EAs within that year. One in­tu­itive rea­son why this might be is that those who have more re­cently heard of EA tend to be younger, more likely to be stu­dents (or may be ear­lier in their ca­reers) and so have a lower in­come. In ad­di­tion, new EAs might also be ex­pected to be less likely to be ex­tremely ded­i­cated or will­ing to donate large sums right away. There may also be some sur­vivor­ship bias, where those EAs who re­port first hear­ing of EA in ear­lier years, and are still tak­ing the sur­vey in 2018, may be more likely to ded­i­cated, highly in­volved (and per­haps higher donat­ing) EAs.

This vet­eran ver­sus new­comer differ­ence is ap­par­ent within non-stu­dents, pledge tak­ers, those earn­ing to give, and es­pe­cially among high-in­come in­di­vi­d­u­als. Without lon­gi­tu­di­nal data it is difficult to know if this is due to a time effect, or group-switch­ing over time. There is very lit­tle differ­ence be­tween new­comer and vet­eran non-pledge tak­ers, stu­dents, other ca­reer types, and other in­come-level in­di­vi­d­u­als. This makes in­tu­itive sense as this group are likely to now ei­ther have low in­comes or not be ex­pressly com­mit­ted to donat­ing a sig­nifi­cant por­tion of in­come, ir­re­spec­tive of the length of time they have been in EA.

(Right click and open in new tab to ex­pand image)

Pre­dic­tors of Donation

We ex­am­ined the effect of in­come, stu­dent sta­tus, num­ber of years in­volved with EA, and mem­ber­ship in GWWC as po­ten­tial pre­dic­tors of dona­tions. In­di­vi­d­ual in­come was the strongest pre­dic­tor of dona­tion amount, fol­lowed by a pos­i­tive im­pact of mem­ber­ship in GWWC in­ter­act­ing with in­come (link to re­gres­sion table). Be­fore anal­y­sis, two out­liers with large in­fluence were re­moved (in­di­vi­d­ual in­comes >$5 mil­lion), and the data were log trans­formed, cen­tered and scaled to im­prove nor­mal­ity. House­hold in­come was strongly cor­re­lated with in­di­vi­d­ual in­come (~80%) and was there­fore ex­cluded from the model. This model (AIC=3114), was slightly preferred over a model with no in­ter­ac­tion be­tween GWWC and in­come (AIC=3121), and ex­plained about 45% of the vari­a­tion in the data. The model was also preferred over a sim­ple re­gres­sion of dona­tion vs in­come (AIC=3498) which ac­counted for 37% of the vari­a­tion in the nor­mal­ized data.

As pre­vi­ously noted, there was a small nega­tive im­pact of be­ing a stu­dent on dona­tion amount. As seen in the figure be­low, this effect was slightly miti­gated by be­ing a mem­ber of GWWC.

There was also a very small nega­tive effect of be­ing aware of EA for a shorter time. There was lit­tle cor­re­la­tion be­tween be­tween years of EA in­volve­ment and in­di­vi­d­ual in­come (16%), so the ex­pla­na­tion for this effect may have more to do with years of com­mit­ment to char­i­ta­ble giv­ing than in­come-re­lated fac­tors such as stu­dent sta­tus.

Which Char­i­ties are EAs Donat­ing to?

We re­ceived in­for­ma­tion about which spe­cific char­i­ties re­spon­dents donated to from 494 out of the 1363 EAs who re­ported donat­ing any­thing in 2017. Given this, in­for­ma­tion about to­tals of dona­tions to spe­cific char­i­ties should be treated with cau­tion.

As in pre­vi­ous years, GiveWell char­i­ties, led by GiveDirectly, re­ceived among the most re­ported dona­tions. How­ever, un­like pre­vi­ous years, CEA tops the list for to­tal re­ported dona­tions. The Hu­mane League at­tracted sub­stan­tially more fund­ing (mostly ex­plained by one large donor) and donors than in 2016, among those who re­ported dona­tions. In ad­di­tion, AMF, GiveDirectly, GiveWell, and MIRI re­ceived the largest num­ber of in­di­vi­d­u­als donat­ing to them.

Conclusions

To­tal EA dona­tions within our sam­ple are dom­i­nated by a fairly small num­ber of very large donors. Nev­er­the­less, me­dian dona­tions do seem to be slowly in­creas­ing, com­pared to ear­lier years. Fur­ther­more, me­dian dona­tions and per­centages of in­come donated are sub­stan­tially higher when ex­clud­ing stu­dents, those not fully em­ployed or those on a low in­come.

[1] Un­less oth­er­wise stated “in­come” refers to the av­er­age of in­di­vi­d­ual and house­hold in­come, which were re­ported sep­a­rately. Th­ese mea­sures were ex­tremely highly cor­re­lated and us­ing ei­ther of the in­di­vi­d­ual mea­sures made no differ­ence to our analy­ses where we tested this.

Up­dates and corrections

Cor­rected per­centage donated values

Added column la­bels to per­centiles ta­bles at request

Credits

This post was writ­ten by David Moss, with con­tri­bu­tions from Neil Dul­laghan and Kim Cud­ding­ton.

Anal­y­sis con­ducted by Re­think Pri­ori­ties staff, David Moss, Neil Dul­laghan, Kim Cud­ding­ton and Peter Hur­ford.

Thanks to Peter Hur­ford and Tee Bar­nett for edit­ing.

The an­nual EA Sur­vey is a pro­ject of Re­think Char­ity with anal­y­sis and com­men­tary from re­searchers at Re­think Pri­ori­ties.

Sup­port­ing Documents

Ar­ti­cles in the 2018 EA Sur­vey Series

I—Com­mu­nity De­mo­graph­ics & Characteristics

II—Distri­bu­tion & Anal­y­sis Methodology

III—How do peo­ple get in­volved in EA?

IV—Sub­scribers and Identifiers

V—Dona­tion Data

VI—Cause Selection

Prior EA Sur­veys con­ducted by Re­think Charity

The 2017 Sur­vey of Effec­tive Altruists

The 2015 Sur­vey of Effec­tive Altru­ists: Re­sults and Analysis

The 2014 Sur­vey of Effec­tive Altru­ists: Re­sults and Analysis