Cause profile: mental health


In this piece, I ar­gue that men­tal ill­ness may be one of the world’s most press­ing prob­lems.

Here is a sum­mary of the key points:

  • Not only does men­tal ill­ness seem to cause as much, if not more, to­tal wor­ld­wide un­hap­piness than global poverty, it also seems far more ne­glected.

  • Effec­tive men­tal health in­ter­ven­tions ex­ist cur­rently. Th­ese have been im­prov­ing over time and we can ex­pect fur­ther im­prove­ments.

  • I es­ti­mate the cost-effec­tive­ness of a par­tic­u­lar men­tal health or­gani­sa­tion, StrongMinds, and claim it is (at least) four times more effec­tive per dol­lar than GiveDirectly, a GiveWell recom­mended top char­ity. This as­sumes we un­der­stand cost-effec­tive­ness in terms of hap­piness, as mea­sured by self-re­ported life satis­fac­tion.

  • I ex­plain why it’s un­clear if StrongMinds is bet­ter than all the other GiveWell recom­mended life-im­prov­ing char­i­ties (due to in­con­sis­tent ev­i­dence re­gard­ing nega­tive spillovers from wealth in­creases) and life-sav­ing char­i­ties (due to method­olog­i­cal is­sues about where on a 0-10 life satis­fac­tion scale is the ‘neu­tral point’ equiv­a­lent to be­ing dead).

  • I make some ini­tial sug­ges­tions for the high­est-im­pact ca­reers, as well as al­ter­na­tive dona­tion op­por­tu­ni­ties. No thor­ough anal­y­sis has yet been done to com­pare these.

  • While men­tal health has the most ob­vi­ous ap­peal for those who be­lieve we ought to be max­imis­ing the hap­piness of peo­ple al­ive to­day, I ex­plain that be­lief isn’t nec­es­sary to con­clude it is of the high­est pri­or­ity: some­one could, in prin­ci­ple, value what hap­pens to all pos­si­ble sen­tient life and still rea­son­ably de­cide this cause is where they’ll do the most good. I raise, but do not seek to re­solve, the many cru­cial con­sid­er­a­tions here.

In or­der to get a sense of how im­por­tant work on this area is, I ex­am­ine (i) the scale, (ii) ne­glect­ed­ness, and (iii) tractabil­ity of the prob­lem in turn. Ul­ti­mately, tractabil­ity—which I un­der­stand as cost-effec­tive­ness—is what re­ally mat­ters and the pre­ced­ing two sec­tions should be seen as pro­vid­ing helpful back­ground. I then set out why some­one might—and might not—think this cause is their top pri­or­ity and what they could do next if they de­cided it is.

Scale (how many suffer and by how much?)

Sec­tion sum­mary: men­tal ill­ness causes more suffer­ing than poverty in de­vel­oped coun­tries, seems to cause roughly as much suffer­ing wor­ld­wide as poverty does and, un­like poverty, is not shrink­ing.

The 2013 Global Bur­den of Disease (GBD) re­port es­ti­mated that de­pres­sion af­fects ap­prox­i­mately 350m peo­ple an­nu­ally, while anx­iety af­flicts an­other 250 mil­lion.[1] By com­par­i­son, the re­port es­ti­mated that malaria af­fects 146 mil­lion peo­ple, while a 2015 World Bank re­port es­ti­mated 702 mil­lion peo­ple liv­ing on less than $1.25 a day.[2] While poverty af­fects many more peo­ple than men­tal health, the share of the world pop­u­la­tion liv­ing in ab­solute poverty is fal­ling rapidly: there were 1.76 billion in ab­solute poverty in 1999, a drop of about 1 billion peo­ple.[3] By con­trast, se­vere men­tal ill­nesses are on the rise.[4] As one ex­am­ple, in the UK the pro­por­tion of those re­port­ing se­vere symp­toms of com­mon men­tal di­s­or­ders has risen 34.7% be­tween 1993 and 2014 (from 6.9% to 9.3% of the pop­u­la­tion).[5] It’s un­likely this is solely due to in­creased re­port­ing: an Amer­i­can birth co­hort anal­y­sis run­ning from 1938 to 2010 found large in­creases in all psy­chopatholo­gies af­ter us­ing stan­dard meth­ods to con­trol for pos­si­ble in­creases in re­port­ing.[6]

To prop­erly as­sess scale we also need to know how much suffer­ing each causes: if poverty makes peo­ple mis­er­able but men­tal ill­nesses are only mildly bad, poverty will be larger in scale. In a re­cent anal­y­sis of self-re­ported hap­piness scores, the World Hap­piness Re­port eval­u­ated how well poverty, lack of ed­u­ca­tion, un­em­ploy­ment, be­ing sin­gle, phys­i­cal health (N.B. not just malaria) and men­tal health ex­plain mis­ery, where ‘mis­ery’ here refers to those re­port­ing the low­est hap­piness scores, roughly the bot­tom 10%.[7] Hap­piness is stan­dardly mea­sured by ask­ing, “Over­all how satis­fied are you with your life these days?” mea­sured on a scale of 1 to 10 (from ‘ex­tremely dis­satis­fied’ to ‘ex­tremely satis­fied’).[8] For dis­cus­sion of whether such mea­sures of hap­piness are mean­ingful in gen­eral, see Dolan and White (2007) and Diener, In­gle­hart and Tay (2013) and for whether are cross-na­tion­ally com­pa­rable, see Veen­hoven (2012).[9] The au­thors of the World Hap­piness Re­port found men­tal ill­ness was the biggest cause of mis­ery over­all (i.e. it ac­counts for the largest pro­por­tion of those in the mis­er­able cat­e­gory). This is rep­re­sented in figure 1 be­low.

Figure 1. Men­tal ill­ness in­creases the like­li­hood of be­ing ‘in mis­ery’ the most. Graph­ics from The Economist, un­der­ly­ing data from World Hap­piness Re­port 2017 (ch.5)

Three of the sur­veyed coun­tries were de­vel­oped (UK, USA, Aus­tralia) and one was not (In­done­sia), but in each coun­try emo­tional prob­lems were the biggest cause of mis­ery. In­ter­est­ingly, men­tal ill­ness was still the biggest cause of lost hap­piness if we look at the non-mis­er­able part of the pop­u­la­tion too.[10] Of course, we may well ex­pect poverty causes a larger amount of mis­ery in coun­tries even poorer than In­done­sia—I’m not aware of an equiv­a­lent anal­y­sis of very poor coun­tries—so there’s room to dis­agree about whether poverty or men­tal ill­ness causes the most un­hap­piness wor­ld­wide.

For our pur­poses, it’s not es­sen­tial to work out which of those two is big­ger. The key point is that the wor­ld­wide scale of suffer­ing caused by men­tal ill­ness is huge, and at least of the same or­der of mag­ni­tude as poverty. It will al­most in­evitably in­crease, rel­a­tive to poverty, over time. I ex­pect this to be a sur­prise to many read­ers; it cer­tainly was to me, as I had as­sumed poverty caused far more global suffer­ing.[11]

Ne­glect­ed­ness (how many re­sources are go­ing to­wards this prob­lem?)

Sec­tion sum­mary: Men­tal health is ne­glected, even rel­a­tive to poverty and phys­i­cal health.

One third of Lower and Mid­dle In­come (i.e. de­vel­op­ing) Coun­tries do not have a des­ig­nated men­tal health bud­get,[12] and for those that do the av­er­age ex­pen­di­ture is 0.5% of their to­tal health bud­get.[13] In such coun­tries, the treat­ment gap for men­tal health (i.e. the num­ber who don’t get treat­ment as a per­centage of those who need it) is 76-85%.[14] A Cen­tre for Global Devel­op­ment re­port de­scribes men­tal ill­ness as a “truly ne­glected area of global health policy”.[15]

Tractabil­ity (are there cost-effec­tive solu­tions?)

Sec­tion sum­mary: ex­ist­ing psy­cholog­i­cal treat­ments for men­tal ill­ness are effec­tive; more effec­tive in­ter­ven­tions seem to be on the hori­zon.

Even if men­tal health is a large-scale, ne­glected prob­lem, we shouldn’t con­sider it a pos­si­ble moral pri­or­ity if there aren’t effec­tive treat­ments. For­tu­nately, there are. Broadly, we can di­vide the treat­ments into three types. (Dis­cus­sion of the cost-effec­tive­ness of men­tal health treat­ments, rel­a­tive to GiveWell top recom­men­da­tions, oc­curs in the next sec­tion.)

First, there are psy­cholog­i­cal treat­ments. While many peo­ple’s men­tal image of this is still Freudian psy­cho­anal­y­sis (“lie down on the couch; tell me about your child­hood”), such ther­apy has been shown to be in­effec­tive and mod­ern medicine has moved on. In fact, there are no ran­domised tri­als which have demon­strated psy­cho­anal­y­sis does bet­ter than the nat­u­ral rate of re­cov­ery.[16] The UK’s Na­tional In­sti­tute for Health and Care Ex­cel­lence (NICE) recom­mends Cog­ni­tive Be­havi­oural Ther­apy (CBT), as the first line treat­ment for all men­tal health di­s­or­ders.[17] CBT in­volves teach­ing peo­ple how to un­der­stand their thoughts and emo­tions and pro­cess them differ­ently. The UK’s Im­prov­ing Ac­cess to Psy­cholog­i­cal Ther­a­pies (IAPT) pro­gramme treats 560,000 pa­tients a year, of whom 50% re­cover and two-thirds show worth­while benefits.[18]

CBT is best un­der­stood as a fam­ily of ther­a­pies, rather than a sin­gle treat­ment: the CBT pro­ce­dures for de­pres­sion and, say, so­cial pho­bia, share al­most no over­lap.[19] The lat­est forms of CBT are a sub­stan­tial im­prove­ment on ear­lier meth­ods. For ex­am­ple, 78% of pa­tients with so­cial anx­iety re­cov­ered from cog­ni­tive treat­ments, com­pared with 38% from ex­po­sure ther­apy, which was the origi­nal (non-psy­cho­an­a­lytic) pro­ce­dure used to treat it. [20] The effects of CBT can be long-last­ing, too. A study on CBT re­cip­i­ents found the effects of treat­ment (com­pared to usual care with an­tide­pres­sants), mea­sured in stan­dard men­tal health scores, were pre­sent 4 years later with­out ob­vi­ously ap­pear­ing to re­duce over time, as shown in figure 2. [21] This can be ex­plained by the fact that CBT teaches the pa­tient a skill; so as long as they re­mem­ber what they’ve learnt, the effect should con­tinue.

Figure 2. The effect of Cog­ni­tive Be­havi­oural Ther­apy (CBT) re­mains for at least 4 years, as in­di­cated by the lower mean scores for the treat­ment group on the Beck De­pres­sion In­ven­tory (BDI-II), com­pared to the con­trol group.

CBT is not the only effec­tive psy­cholog­i­cal ther­apy: oth­ers in­clude mind­ful­ness-based cog­ni­tive ther­apy, in­ter­per­sonal ther­apy, and coun­sel­ling.[22]

Psy­cholog­i­cal treat­ments are con­ven­tion­ally de­liv­ered face-to-face, ei­ther one-on-one or in a group set­ting. A lead­ing ex­am­ple of a char­ity in the de­vel­op­ing world pro­vid­ing these types of in­ter­ven­tion is StrongMinds, which offers in­ter­per­sonal group ther­apy to women in Uganda.[23] Such treat­ments can also be effec­tively de­liv­ered elec­tron­i­cally, ei­ther via com­put­ers or smart­phones.[24] As this doesn’t re­quire hu­man in­ter­ac­tion, it could po­ten­tially be very cheap and scal­able.

Se­cond, there are chem­i­cal treat­ments, the best known be­ing an­tide­pres­sants.[25] De­spite con­tro­versy, ev­i­dence shows they are effec­tive (more so for se­vere than mild or mod­er­ate de­pres­sion), al­though they seem to func­tion like painkil­lers, treat­ing the symp­toms with­out re­mov­ing the un­der­ly­ing cause(s). This ex­plains why cog­ni­tive treat­ments, un­like an­tide­pres­sants, re­duce the rates of re­lapse.[26]

There are also some highly promis­ing treat­ments for men­tal ill­nesses that rely on cur­rently ille­gal recre­ational drugs. Ke­tamine may provide sub­stan­tial short-term re­lief from de­pres­sion.[27] MDMA (‘ec­stasy’) has re­cently been la­bel­led a ‘break­through drug’ by the FDA in the USA be­cause of its re­mark­able effec­tive­ness in treat­ing post-trau­matic stress di­s­or­der.[28] Per­haps most promis­ing of all is the po­ten­tial of psychedelics, such as LSD and psilo­cy­bin (the ac­tive in­gre­di­ent in ‘magic mush­rooms’) for treat­ing men­tal ill­ness. Carhart-Har­ris et al. (2016) gave a sin­gle dose of psilo­cy­bin to 12 peo­ple with mod­er­ate to se­vere treat­ment-re­sis­tant de­pres­sion.[29] The sub­jects had been de­pressed for a mean av­er­age of 17.8 years. After psychedelic-as­sisted ther­apy, 67% were classed as non-de­pressed af­ter 1 week and 42% as non-de­pressed af­ter 3 months with­out any fur­ther treat­ment. This is dis­played in figure 3.

Figure 3. De­pres­sion sever­ity dropped sub­stan­tially af­ter psilo­cy­bin treat­ment and shows an effect 3 months later with­out fur­ther treat­ment.

Stage-2 tri­als started in Septem­ber 2018 to test the effec­tive­ness of psilo­cy­bin on a larger scale. In Oc­to­ber 2018 the FDA granted the ‘break­through ther­apy’ des­ig­na­tion for psilo­cy­bin ther­apy for Treat­ment-re­sis­tant De­pres­sion. If psychedelics turn out to be even par­tially as effec­tive as they first ap­pear, they could still be­come the most-effec­tive treat­ment for de­pres­sion.[30]

Third, there are di­rect, or elec­tri­cal, treat­ments, such as deep brain stim­u­la­tion (DBS)[31] and repet­i­tive tran­scra­nial mag­netic stim­u­la­tion (rTMS).[32] I am un­cer­tain of how effec­tive these are com­pared to other treat­ments, but note that NICE do not recom­mend these as a first-line treat­ment. Pos­si­bly, fur­ther re­search will im­prove them or demon­strate their po­ten­tial.

Of course, we shouldn’t only be think­ing about treat­ing peo­ple who are already suffer­ing, but also about pre­ven­tion. One promis­ing (psy­cholog­i­cal) method of pre­ven­tion this would be ‘pos­i­tive ed­u­ca­tion’: teach­ing re­silience and life skills in schools. A se­ries of tri­als in Bhutan, Peru and Mex­ico in­volv­ing over 750,000 stu­dents found that pos­i­tive ed­u­ca­tion not only im­proved mea­sures of child well-be­ing, but also in­creased stan­dard­ised aca­demic test scores.[33]

Over­all, there are a num­ber of ways to tackle men­tal ill­ness and it re­mains to be seen which will be the best.

Men­tal health vs global de­vel­op­ment: com­par­ing char­ity cost-effectiveness

Sec­tion sum­mary: a men­tal health char­ity, StrongMinds, looks more cost-effec­tive at im­prov­ing lives than one of GiveWell’s top char­i­ties, GiveDirectly. This as­sumes we un­der­stand cost-effec­tive­ness in terms of hap­piness, as mea­sured by self-re­ported life satis­fac­tion. It’s un­clear if StrongMinds is bet­ter than all GiveWell’s life-im­prov­ing char­i­ties (due to gaps in ev­i­dence re­gard­ing the size of nega­tive spillover effects) or GiveWell’s life-sav­ing char­i­ties (due to method­olog­i­cal is­sues about where on a 0-10 life satis­fac­tion is the ‘neu­tral point’ equiv­a­lent to be­ing dead).

Even if there are men­tal health in­ter­ven­tions that work, effec­tive al­tru­ists shouldn’t donate to or­gani­sa­tions that com­bat men­tal ill­ness if they can find some­thing even bet­ter. As the global de­vel­op­ment char­i­ties recom­mended by GiveWell are widely ac­cepted as the most effec­tive or­gani­sa­tions to donate to at pre­sent, they are the nat­u­ral point of com­par­i­son.

I claim that, on the pre­sent ev­i­dence, a men­tal health char­ity, StrongMinds, is at least four times more cost-effec­tive than at least one of GiveWell’s life-im­prov­ing char­i­ties, GiveDirectly, which pro­vides un­con­di­tional cash trans­fers to poor Kenyan farm­ers. I am not able to say whether it is bet­ter than GiveWell’s other life-im­prov­ing char­i­ties—which all, ul­ti­mately, aim to alle­vi­ate poverty—or GiveWell’s life-sav­ing char­i­ties. I will com­pare StrongMinds first to the life-im­prov­ing, then to the life-sav­ing, GiveWell recom­men­da­tions, set­ting out the key un­cer­tainty in each case.

StrongMinds vs GiveWell’s recom­mended life-im­prov­ing charities

To eval­u­ate StrongMinds and GiveDirectly, we need to con­vert their im­pact into a com­mon unit. I’ll use self-re­ported life satis­fac­tion scores, as men­tioned ear­lier, to make this com­par­i­son. In effect, the ques­tion be­comes: which char­ity is more cost-effec­tive at in­creas­ing life satis­fac­tion? A longer treat­ment of how and why life satis­fac­tion scores should be used to mea­sure hap­piness can be found here.

A study of GiveDirectly in­di­cates cash trans­fers in­crease life satis­fac­tion by about 0.3 life satis­fac­tion points—here­after ‘LSPs’ - on a 10 point scale.[34] To ex­plain, this effect would be equiv­a­lent to in­creas­ing my life satis­fac­tion from 6 to 6.3 out of 10. This was mea­sured af­ter 4.3 months on av­er­age, but let’s as­sume this effect lasts a whole year. I as­sume it is the same for ev­ery­one in the re­cip­i­ent house­hold, and there are 5 peo­ple per house­hold on av­er­age. Hence the an­nual LSP im­pact is 0.3 (LS/​per­son) x 1 (year) x 5 (per­sons) = 1.5 LSPs. The av­er­age cash trans­fer is $750, im­ply­ing a cost-effec­tive­ness of 2 LSPs/​$1000.

How­ever, this may well over­state the effec­tive­ness of cash trans­fers. It ac­counts only for the life satis­fac­tion in­crease of re­cip­i­ents. Re­search into GiveDirectly has sug­gested that their cash trans­fers, while mak­ing some peo­ple wealthier (and so more satis­fied with life) have nega­tive spillovers: it makes non-re­cip­i­ents less satis­fied with their lives. As Haushofer, Reis­inger and Shapiro (2015, p1) state:

“The de­crease in life satis­fac­tion in­duced by trans­fers to neigh­bors more than offsets the di­rect pos­i­tive effect of trans­fers, and is largest for in­di­vi­d­u­als who did not re­ceive a di­rect trans­fer them­selves.”[35]

While this re­sult may sur­prise the reader, as Clark (2017) notes ‘there is con­sid­er­able ev­i­dence from a va­ri­ety of sources to sug­gest that well-be­ing is a func­tion of rel­a­tive in­come’.[36] In other words, your ab­solute level of wealth mat­ters much less than whether you are richer or poorer than your peers. Hence the find­ing about GiveDirectly is con­sis­tent with the other re­search on hap­piness. In­deed, Clark et al. (2018), us­ing de­vel­oped coun­try data, finds dou­bling one per­son’s in­come causes a re­duc­tion in life-satis­fac­tion in that in­di­vi­d­ual’s peer group that is nearly iden­ti­cal in size to the life satis­fac­tion gain to the in­di­vi­d­ual them­self [37]. What is novel about the 2015 GiveDirectly re­sult is that it finds this rel­a­tive effect at such a low level of ab­solute wealth.

We might hope these nega­tive spillovers would dis­si­pate even­tu­ally and, over the long run, cash trans­fers would be effec­tive in in­creas­ing life satis­fac­tion. A 2018 study on the long-term (3 year) effects of GiveDirectly by Haushofer and Shapiro (2018, p. 22) finds re­cip­i­ents, com­pared to non-re­cip­i­ents in dis­tant villages, have 40% more as­sets but that re­cip­i­ents do no bet­ter on a psy­cholog­i­cal well-be­ing in­dex.[38] Hence, the effect seems to be in the short-term.

Con­fus­ingly how­ever, GiveWell claim a more re­cent ‘gen­eral equilbrium’ (GE) study of GiveDirectly, to which GiveWell have been given pri­vate ac­cess to a draft, shows there is:

“No ev­i­dence of across village spillover effects on house­hold as­set own­er­ship or sub­jec­tive well-be­ing. They find a pos­i­tive spillover effect on sub­jec­tive well-be­ing of in­el­i­gible house­holds within treat­ment villages”.[39]

I note this (more re­cent) find­ing is in­con­sis­tent both with the 2015 pa­per, and the more gen­eral find­ing of rel­a­tive in­come noted by Clark (et al.) above. Given this in­con­sis­tency, and with­out ac­cess to the draft, I’m un­clear how to up­date my views ac­cord­ingly. Let’s re­turn to this con­cern in a mo­ment.

There isn’t re­search on StrongMinds which has di­rectly mea­sured its im­pact in terms of life satis­fac­tion, so I es­ti­mate this us­ing other available in­for­ma­tion, ex­plained this end­note.[41] I in­fer that the treat­ment effect is an in­crease 0.8 LSPs per per­son (to 1 d.p.; mod­el­led ei­ther as 0.2 LSPs per per­son per year for 4 years, or 0.2 LSP in the first year with a 75% an­nual re­ten­tion there­after). StrongMinds say their per-par­ti­ci­pant costs are $102 (StrongMinds Q1.2018 re­port). That sug­gests the im­pact is 8 LSPs/​$1000 (to 1 d.p.).

The cost-effec­tive­ness eval­u­a­tion be­tween StrongMinds and GiveDirectly is not sen­si­tive to in­clu­sion of the nega­tive spillovers as, even if we as­sume GiveDirectly has no nega­tive spillover effects, then StrongMinds would still seem to be around four times more cost-effec­tive (2 vs 8 LSPs/​$1,000). Note that, if we were take the view that its nega­tive spillovers were just as big as its pos­i­tive im­pacts, then GiveDirectly would not in­crease ag­gre­gate life satis­fac­tion at all and any­thing with a pos­i­tive im­pact would be more cost-effec­tive.

In con­trast, the anal­y­sis of whether StrongMinds is cost-effec­tive than GiveWell’s other life-im­prov­ing char­i­ties—namely SCI, De­worm the World, SightSavers and END—is sen­si­tive to how strong we think the nega­tive spillover effect of in­creas­ing wealth is. To see this, we must first recog­nise these other char­i­ties could have nega­tive spillovers: they all even­tu­ally in­crease the wealth of their benefi­cia­ries. While those or­gani­sa­tions provide de­worm­ing in­ter­ven­tions, the vast ma­jor­ity of the benefit of de­worm­ing, ac­cord­ing to GiveWell, comes not from re­duc­ing the phys­i­cal dis­com­fort the worms cause, but from the fact de­wormed chil­dren earn more in later life.[42] If it’s gen­er­ally true that mak­ing some wealthier (and so more satis­fied), re­duces the satis­fac­tion by oth­ers to at least ex­tent, then this con­cern will ap­ply there too. Se­cond, note that, ac­cord­ing to GiveWell, all these other char­i­ties are at least four times more cost-effec­tive than GiveDirectly: De­worm the World is rated as 18.3 times bet­ter (the high­est), and END 5.5 (the low­est). If StrongMinds is only 4 times more cost-effec­tive on GiveDirectly then, as­sum­ing there are no nega­tive spillovers, all these other char­i­ties would be bet­ter (as­sum­ing GiveWell’s anal­y­sis is oth­er­wise cor­rect). If, on the other hand, these char­i­ties have nega­tive spillovers large enough to en­tirely can­cel out their pos­i­tive im­pacts, then StrongMinds will be bet­ter. Thus, the size of nega­tive spillovers is highly im­por­tant, which is why it is frus­trat­ing the ev­i­dence is in­con­sis­tent (at least, it is once we ac­count for the lat­est, un­pub­lished draft of GiveDirectly seen by GiveWell).

I am not able to offer a re­s­olu­tion to this is­sue here. What I think fol­lows from this anal­y­sis is (a) more work is needed on un­der­stand­ing the size of nega­tive spillover effects from wealth in­creases, (b) the pos­si­bil­ity men­tal health in­ter­ven­tion could be more cost-effec­tive than the best poverty-alle­vi­a­tion in­ter­ven­tions should be taken se­ri­ously and (c) given how lit­tle scrutiny men­tal health has re­ceived with effec­tive al­tru­ism, efforts to find even more effec­tive men­tal health in­ter­ven­tions are valuable.

StrongMinds vs GiveWell’s recom­mended life-sav­ing charities

The other im­por­tant com­par­i­son to make is be­tween the life-im­prov­ing men­tal health in­ter­ven­tion, StrongMinds, and GiveWell’s recom­mended life-sav­ing in­ter­ven­tions, such as the Against Malaria Foun­da­tion (AMF). This is anal­y­sis is not very straight­for­ward. First, there are differ­ent philo­soph­i­cal views about the bad­ness of an in­di­vi­d­ual’s death, as I note el­se­where.[43] Se­cond, a thor­ough anal­y­sis would also need to take into ac­count the ‘so­cial value’ of lives, the value a given life has on ev­ery­one else. To as­sess this we’d need to ac­count for the grief to friends and fam­ily sav­ing a life pre­vents, as well as some com­pli­cated and po­ten­tially dis­turb­ing fac­tors that are not usu­ally con­sid­ered, such as the meat eater prob­lem[44] and whether or not the Earth is un­der- or over­pop­u­lated.[45]

I will put the com­plex­ities of the so­cial value of lives to one side, and use the (sim­ple) life com­par­a­tive ac­count of the bad­ness of death, on which the value of sav­ing a life is the to­tal well-be­ing the per­son would have had if they’d lived.

How cost-effec­tive is AMF? Ac­cord­ing to GiveWell’s es­ti­mates, AMF saves a life (i.e. pre­vents a pre­ma­ture death) for around $3,500 [46]. Sup­pose that grants 60 coun­ter­fac­tual years of life. Let’s again use life satis­fac­tion scores as our met­ric of well-be­ing. Aver­age life satis­fac­tion in Kenya, where AMF op­er­ates, is 4.4 out of 10.[47] Now we run into a prob­lem. Life satis­fac­tion sur­veys don’t ask peo­ple to spec­ify what point on the 0 to 10 scale they would con­sider equiv­a­lent to not be­ing al­ive. 0 is la­bel­led ‘ex­tremely dis­satis­fied’ and 10 ‘ex­tremely satis­fied’. In­tu­itively, the mid-point in the scale, 5, would be the neu­tral point. Yet, if that’s true, then sav­ing lives through AMF would, in fact, be bad: 4.4 (out of 10) is be­low the neu­tral point, so AMF are pro­long­ing lives not worth liv­ing.

Let’s sup­pose in­stead the neu­tral point is 4. If this is so, sav­ing the child is worth 0.4 life satis­fac­tion points a year for 60 years, thus 24 LSPs (0.4 x 60). Given the $3,500 cost, we can calcu­late cost-effec­tive­ness as 6.9 LSPs/​$1,000. Ear­lier, I es­ti­mated StrongMinds’ cost-effec­tive­ness was 8 LSPs/​$1000.[48] If these es­ti­mates are cor­rect, then StrongMinds is still more cost-effec­tive, albeit only slightly, than AMF. Of course, we should be cau­tious about tak­ing these es­ti­mates too liter­ally.

The prob­lem here is that these cost-effec­tive­ness num­bers are highly de­pen­dent on a (so far) ar­bi­trary de­ci­sion about where the neu­tral point goes. If some­one in­stead as­sumes the neu­tral point were 3 - which, in­tu­itively, seems too low—then AMF’s cost-effec­tive­ness would leap to 24.4 LSPs/​$1,000 and it would be more cost-effec­tive than StrongMinds, at least on this sim­plified pic­ture. I note this ap­proach to com­par­ing life-im­prov­ing to life-com­par­ing in­ter­ven­tions (i.e. as­sess­ing how much ad­di­tional life satis­fac­tion is gen­er­ated) is differ­ent GiveWell’s ap­proach (i.e. ask­ing their staff to judge how many years of dou­bled con­sump­tion are morally equiv­a­lent to sav­ing a child’s life).

More work is there­fore ur­gently needed to de­ter­mine where the neu­tral point is. Two po­ten­tial meth­ods for do­ing this would be (1) ask­ing peo­ple to state where they think this neu­tral point is; or (2) us­ing mood re­ports and find­ing out at what score on the life satis­fac­tion scale peo­ple re­port net neu­tral mood.

Why might you—and why might you not—pri­ori­tise this area?

Sec­tion sum­mary: there are a large num­ber of open-ended ques­tions which need to be re­solved be­fore we can de­ter­mine how we do the most good. This sec­tion sets out some of the cru­cial con­sid­er­a­tions, but does not at­tempt to an­swer them.

Ul­ti­mately, you should pri­ori­tise the prob­lem that you think will al­low you to do the most good with your spare re­sources (your time and money). Work­ing out what will do the most good is ob­vi­ously rather com­pli­cated. In this sec­tion I’ll set out some of the con­sid­er­a­tions that seem most rele­vant to de­cid­ing whether men­tal health could be your top pri­or­ity. I do not con­sider this an ex­haus­tive list.

Value: is hap­piness all that ul­ti­mately mat­ters?

Ear­lier, I sug­gested a par­tic­u­lar men­tal health char­ity might be more cost-effec­tive at in­creas­ing hap­piness than at least one of the global de­vel­op­ment char­i­ties GiveWell recom­mend. Pre­sum­ably, hap­piness, a pos­i­tive bal­ance of en­joy­ment over suffer­ing, is one of the things that is in­trin­si­cally good, that is, good in it­self. Some (in­clud­ing this au­thor), be­lieve it is the only in­trin­sic good.

If you be­lieved there were other in­trin­sic goods, that would po­ten­tially change the pri­ori­ties. Sup­pose you thought au­ton­omy, equal­ity and hap­piness were all valuable in them­selves. As a re­sult, you might, per­haps, think poverty is more im­por­tant than men­tal health on the grounds poverty alle­vi­a­tion in­creases au­ton­omy and equal­ity whereas im­prov­ing men­tal health does not.

How­ever, such a con­clu­sion might be a bit too quick. Even if you value things other than hap­piness, a men­tal health in­ter­ven­tion might, from the point of view of your moral the­ory, still do the most good. You could be­lieve tar­get­ing men­tal ill­ness is just as au­ton­omy-en­hanc­ing or equal­ity-in­creas­ing as alle­vi­at­ing poverty, given how dis­abling it can be to be de­pressed, anx­ious, etc. Or, even if treat­ing men­tal health does a poor job at pro­mot­ing au­ton­omy and equal­ity, you might think that it does a good enough job of in­creas­ing hap­piness, some­thing that you, pre­sum­ably, value any­way, so as to com­pen­sate for its ap­par­ent lack of effect on other val­ues.

Pop­u­la­tion ethics: do only some peo­ple mat­ter morally, such as those cur­rently al­ive, or does ev­ery­one who could ever live mat­ter?

The (math­e­mat­i­cally) sim­plest view in pop­u­la­tion ethics is to­tal­ism, which holds the best out­come is the one with the great­est sum to­tal of well-be­ing of ev­ery­one—past, pre­sent, and fu­ture. As a re­sult, all pos­si­ble peo­ple mat­ter ac­cord­ing to to­tal­ism.[49]

In con­trast, many peo­ple in­tu­itively hold a ‘per­son-af­fect­ing’ view of pop­u­la­tion ethics: while the well-be­ing of those who do (or will) ex­ist mat­ters, there is no value in cre­at­ing new lives. This is typ­i­cally jus­tified on the grounds that an out­come can be bet­ter or worse be­cause and to the ex­tent that it’s bet­ter or worse for per­sons, and ex­is­tence can never be bet­ter or worse for a per­son than non-ex­is­tence. Hence there’s no value in cre­at­ing new lives, as be­ing cre­ated is not good for any­one. Much of this is cap­tured by Jan Narve­son’s fa­mous slo­gan “We are in fa­vor of mak­ing peo­ple happy, but neu­tral about mak­ing happy peo­ple.” [50]

One par­tic­u­lar per­son-af­fect­ing view is ne­ces­si­tar­i­anism, which holds the only peo­ple who mat­ter, when choos­ing be­tween two out­comes, are those that are go­ing to ex­ist what­ever we choose to do (i.e. ex­ist nec­es­sar­ily). On this view, we still count the fu­ture peo­ple who will ex­ist any­way. How­ever, in prac­tice, it’s un­likely there are many nec­es­sary, fu­ture peo­ple. The iden­tities of who comes into ex­is­tence de­pends, among other things, on whom the ge­netic par­ents of some­one hap­pen to be.[51] Given the na­ture of hu­man re­pro­duc­tion, if your par­ents had had sex a mo­ment ear­lier, or later, then some­one else would al­most cer­tainly have been born in­stead. Plau­si­bly, even small changes in the world are likely to change who gets con­ceived and thus al­ter the iden­tities of fu­ture peo­ple. As a re­sult, ne­ces­si­tar­i­ans will think our moral con­cerns are, in effect, re­stricted to those who presently ex­ist.

As is prob­a­bly ob­vi­ous, whether you think a per­son-af­fect­ing view (e.g. ne­ces­si­tar­i­anism) or what we can call an ‘im­per­sonal’ view, one that counts all pos­si­ble peo­ple (e.g. to­tal­ism), is cor­rect may be a very im­por­tant con­sid­er­a­tion in pri­ori­tis­ing what to do (there are other views in pop­u­la­tion ethics, but it would be an un­nec­es­sary di­ver­sion to dis­cuss those here). Some causes, such as re­duc­ing ex­is­ten­tial threats to hu­man­ity (e.g. from AI) or im­prov­ing an­i­mal welfare (e.g. by re­duc­ing fac­tory farm­ing) pri­mar­ily af­fect fu­ture peo­ple (tech­ni­cally: sen­tient life), rather than pre­sent peo­ple. By con­trast, while alle­vi­at­ing the men­tal ill­ness or re­duc­ing the poverty of those al­ive to­day will have some effect on fu­ture peo­ple, pre­sum­ably their main im­pact is that they benefit the pre­sent gen­er­a­tion. Thus, if you think an im­per­sonal view is cor­rect, you might think fac­tory farm­ing or re­duc­ing ex­is­ten­tial risks are sim­ply a lot more im­por­tant than men­tal health or poverty; if you think a per­son-af­fect­ing view is cor­rect, men­tal health or poverty may be the pri­or­ity in­stead.

Which is cor­rect? Pop­u­la­tion ethics is a no­to­ri­ously prob­le­matic area, where all of the views can be shown to have ap­par­ently im­plau­si­ble im­pli­ca­tions, and there is not space to get into this here.[52] (Psy­cholog­i­cally au­to­bi­o­graph­i­cal aside: I tend to think per­son-af­fect­ing views are the least-bad of the op­tions).

Em­piri­cal and per­sonal complications

Sim­ply set­tling your moral views is not suffi­cient to tell you what to pri­ori­tise; there are a range of em­piri­cal con­sid­er­a­tions too.

For ex­am­ple, you could hold a per­son-af­fect­ing view but still con­clude re­duc­ing ex­is­ten­tial risks is the top pri­or­ity sim­ply based on the risk they pose to the pre­sent gen­er­a­tion.[53]

Or, you could hold an im­per­sonal view—and there­fore ac­cept that it would be good, in the­ory, to make progress on an­i­mal welfare and shap­ing the far fu­ture—but nev­er­the­less think, in prac­tice, you can do more good by fo­cus­ing on some­thing in the near-term, such as men­tal health or global de­vel­op­ment in­stead. You might reach this con­clu­sion if you held some of the fol­low­ing be­liefs: (a) the ex­pected value of the fu­ture will be nega­tive, such that it would be good if sen­tient life dies out and so re­duc­ing ex­is­ten­tial risks would be bad; (b) there are es­pe­cially good op­por­tu­ni­ties to do good in the pre­sent; (c) there is very lit­tle we can do to im­prove the far fu­ture; (d) we have a par­tic­u­lar moral duty to help other ex­ist­ing hu­mans com­pared to fu­ture hu­mans or an­i­mals; or (e) (prob­a­bly the most im­por­tant of all) your skills, abil­ities and per­sonal mo­ti­va­tion mean you have par­tic­u­larly high ap­ti­tude to work on some nearer-cause area rather than an­other.

Fi­nally, some­one could de­cide that mak­ing cur­rently ex­ist­ing hu­mans lives hap­pier (as op­posed to, say, sav­ing hu­man­ity), is how they’d do the most good, but not think men­tal health is the top pri­or­ity. To state the ob­vi­ous, men­tal ill­ness is not the only source of un­hap­piness. If we could wave a magic wand and fully treat all men­tal health con­di­tions, the world would not be at max­i­mum hap­piness. Some pos­si­ble con­tenders not oth­er­wise men­tioned in this doc­u­ment are:

  • Pain. A re­cent Lancet Com­mis­sion on the sub­ject said, “lack of global ac­cess to pain re­lief and pal­li­a­tive care through­out the life cy­cle con­sti­tutes a global crisis, and ac­tion to close this di­vide be­tween rich and poor is a moral, health, and eth­i­cal im­per­a­tive”.[54]

  • Loneli­ness (lack­ing friends) and ‘love­less­ness’ (lack­ing a ro­man­tic part­ner). Given the amount of mis­ery these seem to pro­duce (e.g. see figure 1 ‘No part­ner’), find­ing a way to progress here could be high im­pact.

  • ‘Or­di­nary hu­man un­hap­piness’, which refers to mun­dane, ev­ery­day ex­is­tence be­ing less good than it could be. Even pro­jects that at­tempt to cause seem­ingly triv­ial lifestyle changes, such as ex­er­cis­ing more or en­courag­ing peo­ple to take shorter com­mutes (com­mut­ing is shown, in time-use stud­ies, to be one of the worst parts of the day), are po­ten­tially high im­pact if they have a small, but long-term, daily im­pact across many peo­ple.

How can you help?

Sec­tion sum­mary: some ini­tial char­ity and ca­reer recom­men­da­tions are made. More work is needed here.

Let’s sup­pose you’ve de­cided that men­tal health is how you can do the most good with your time or money. What next?

Re­gard­ing char­i­ta­ble dona­tions, StrongMinds seems to be the cur­rent low-risk favourite.

There are, how­ever, other char­i­ties some­one could con­sider, though none have yet been thor­oughly eval­u­ated against StrongMinds. Th­ese are, with brief ex­pla­na­tions in brack­ets:

Re­gard­ing high im­pact ca­reers, no sys­temic work has been done to eval­u­ate what these might be. In­tu­itively, the fol­low­ing stand out as promis­ing, but these are just some ini­tial ideas:

  1. Cause pri­ori­ti­sa­tion re­search—it’s hard to es­cape this con­clu­sion, given how lit­tle work has been done. For those who are in­ter­ested, I’ve high­lighted some im­por­tant ques­tions in the fol­low­ing doc­u­ment: Hu­man Welfare: A Re­search Agenda. I am in the pro­cess of form­ing an or­gani­sa­tion to con­duct such re­search. If you would like to par­ti­ci­pate, please con­tact me.

  2. Eco­nomics (or other so­cial sci­ence) re­search into hap­piness, e.g. de­ter­min­ing what the most effec­tive gov­ern­ment poli­cies are.

  3. Re­search into new treat­ments for men­tal health, e.g. psychedelics, and/​or ad­vo­cacy for these treat­ments. Lee Sharkey and I have writ­ten about drug policy re­form el­se­where.[55]

  4. Work in de­vel­op­ing world men­tal health. Try­ing to im­prove or­gani­sa­tions already in this area, pos­si­bly with a view to try­ing to start new, more effec­tive ones.

  5. En­trepreneur­ship in men­tal health tech­nol­ogy. There are some effec­tive al­tru­ists work­ing on pro­jects in this area already, such as Mind Ease and UpLift.

  6. Poli­tics or policy. Try­ing to change what gov­ern­ments do and their at­ti­tudes to­wards hap­piness.


Whilst men­tal health has so far not been con­sid­ered one of the world’s most ur­gent prob­lems among effec­tive al­tru­ists, it seems, on fur­ther re­flec­tion, there is a strong case it should as high a pri­or­ity as global health and de­vel­op­ment, a cur­rent EA top-cause. Given the va­ri­ety of op­tions for im­prov­ing men­tal health, and the lack of pri­ori­ti­za­tion re­search on the topic, it seems likely we’ll find even bet­ter ways to make progress on this prob­lem in the near fu­ture. Estab­lish­ing which cause al­lows one to do the most good is com­pli­cated. This doc­u­ment set out some of the cru­cial con­sid­er­a­tions rele­vant to de­cid­ing whether this should (or should not) be some­one’s top pri­or­ity; and sug­gested what you might do if you con­clude work­ing on im­prov­ing men­tal health is how you will do the most good.


[0] I am very grate­ful to Max Dal­ton, Alex Lintz, Siebe Rozen­daal, Jus­tus Arnd, An­drew Fisher and Paul Davies for read­ing and offer­ing com­ments on this draft.

[1] Theo Vos et al., “Global, Re­gional, and Na­tional In­ci­dence, Prevalence, and Years Lived with Dis­abil­ity for 301 Acute and Chronic Diseases and In­juries in 188 Coun­tries, 1990–2013: A Sys­tem­atic Anal­y­sis for the Global Bur­den of Disease Study 2013,” The Lancet 386, no. 9995 (2015): 743–800, https://​​​​10.1016/​​S0140-6736(15)60692-4. De­pres­sion and anx­iety are the ‘com­mon men­tal health di­s­or­ders’; I’ve left out oth­ers such as schizophre­nia, anorexia, bipo­lar di­s­or­der, etc.

[2] Vos et al.

[3] Ibid

[4] Bran­don H. Hi­daka, “De­pres­sion as a Disease of Moder­nity: Ex­pla­na­tions for In­creas­ing Prevalence,” Jour­nal of Affec­tive Di­sor­ders 140, no. 3 (Novem­ber 2012): 205–14, https://​​​​10.1016/​​j.jad.2011.12.036.

[5] NHS, “Adult Psy­chi­a­tric Mor­bidity Sur­vey: Sur­vey of Men­tal Health and Wel­lbe­ing, England, 2014 - NHS Digi­tal,” 2016, https://​​digi­​​data-and-in­for­ma­tion/​​pub­li­ca­tions/​​statis­ti­cal/​​adult-psy­chi­a­tric-mor­bidity-sur­vey/​​adult-psy­chi­a­tric-mor­bidity-sur­vey-sur­vey-of-men­tal-health-and-wellbe­ing-england-2014#key-facts.

[6] Jean M. Twenge et al., “Birth Co­hort In­creases in Psy­chopathol­ogy among Young Amer­i­cans, 1938–2007: A Cross-Tem­po­ral Meta-Anal­y­sis of the MMPI,” Clini­cal Psy­chol­ogy Re­view 30, no. 2 (March 2010): 145–54, https://​​​​10.1016/​​j.cpr.2009.10.005.

[7] John F. Hel­liwell, Richard La­yard, and Jeffrey Sachs, World Hap­piness Re­port 2017, chap­ter 5 (Sus­tain­able Devel­op­ment Solu­tions Net­work, 2017), http://​​wor­ld­hap­­port/​​ed/​​2017/​​.

[8] Ibid. Those in­ter­ested in how life satis­fac­tion (one of the com­po­nents of what is some­times called ‘sub­jec­tive well-be­ing’) is mea­sured, and how re­li­able those mea­sures are should see Ed Diener, Ron­ald In­gle­hart, and Louis Tay, “The­ory and Val­idity of Life Satis­fac­tion Scales,” So­cial Indi­ca­tors Re­search 112, no. 3 (July 13, 2013): 497–527, OECD, Guidelines on Mea­sur­ing Sub­jec­tive Well-Be­ing (OECD Pub­lish­ing, 2013) and Paul Dolan and Mathew P. White, “How Can Mea­sures of Sub­jec­tive Well-Be­ing Be Used to In­form Public Policy?,” Per­spec­tives on Psy­cholog­i­cal Science 2, no. 1 (March 21, 2007): 71–85

[9] Dolan and White, “How Can Mea­sures of Sub­jec­tive Well-Be­ing Be Used to In­form Public Policy?” Ed Diener, Ron­ald In­gle­hart, and Louis Tay, “The­ory and Val­idity of Life Satis­fac­tion Scales,” So­cial Indi­ca­tors Re­search 112, no. 3 (July 13, 2013): 497–527, https://​​​​10.1007/​​s11205-012-0076-y. Ruut Veen­hoven, “Cross-Na­tional Differ­ence in Hap­piness: Cul­tural Mea­sure­ment Bias or Effect of Cul­ture?,” In­ter­na­tional Jour­nal of Wel­lbe­ing 2, no. 4 (De­cem­ber 13, 2012), https://​​­ter­na­tion­aljour­nalofwellbe­​​in­dex.php/​​ijow/​​ar­ti­cle/​​view/​​98.

[10] Ibid. See also Sarah Fleche and Richard La­yard, “Do More of Those in Misery Suffer from Poverty, Unem­ploy­ment or Men­tal Ill­ness?,” Kyk­los 70, no. 1 (Fe­bru­ary 1, 2017): 27–41, https://​​​​10.1111/​​kykl.12129.

[11] Another ap­proach would be to use health met­rics. This is less use­ful that us­ing hap­piness scores for two rea­sons. First, that only al­lows us to com­pare health states, and we want to able to com­pare health states to poverty, which we can do with hap­piness scores. Se­cond, health met­rics (DALYs and QALYs) re­flect how peo­ple who have mostly never ex­pe­rienced these ill­nesses imag­ine they would feel if they did so. A bet­ter al­ter­na­tive is to mea­sure di­rectly how peo­ple ac­tu­ally feel when they ac­tu­ally do ex­pe­rience the ill­ness. When QALYs have been com­pared to hap­piness scores, it was found the pub­lic hugely un­der­es­ti­mated by how much men­tal pain (com­pared with phys­i­cal pain) would re­duce their satis­fac­tion with life, as dis­cussed by Paul Dolan and Robert Met­calfe, “Valu­ing Health,” Med­i­cal De­ci­sion Mak­ing 32, no. 4 (July 2, 2012): 578–82, https://​​​​10.1177/​​0272989X11435173.

[12] Shekhar Sax­ena et al., “WHO’s Assess­ment In­stru­ment for Men­tal Health Sys­tems: Col­lect­ing Essen­tial In­for­ma­tion for Policy and Ser­vice De­liv­ery,” Psy­chi­a­tric Ser­vices 58, no. 6 (June 2007): 816–21, https://​​​​10.1176/​​ps.2007.58.6.816.

[13] WHO, Men­tal Health At­las 2011 (World Health Or­ga­ni­za­tion, 2011).

[14] Vic­to­ria de Me­nil, “Missed Op­por­tu­ni­ties in Global Health: Iden­ti­fy­ing New Strate­gies to Im­prove Men­tal Health in LMICs,” 2015.

[15] Ibid

[16] A Roth and P Fon­agy, What Works for Whom?: A Crit­i­cal Re­view of Psy­chother­apy Re­search, 2nd ed. (Guilford Publi­ca­tions, 2005).

[17] Richard La­yard and David M. (David Millar) Clark, Thrive : The Power of Ev­i­dence-Based Psy­cholog­i­cal Ther­a­pies, n.d.

[18] David M. Clark, “Real­iz­ing the Mass Public Benefit of Ev­i­dence-Based Psy­cholog­i­cal Ther­a­pies: The IAPT Pro­gram,” An­nual Re­view of Clini­cal Psy­chol­ogy 14, no. 1 (May 7, 2018): 159–83, https://​​​​10.1146/​​an­nurev-clinpsy-050817-084833.

[19] La­yard and Clark, Thrive : The Power of Ev­i­dence-Based Psy­cholog­i­cal Ther­a­pies. P143

[20] B Boeck­ing, “Mechanism of Change in Cog­ni­tive Ther­apy for So­cial Pho­bia” (King’s Col­lege Lon­don, 2010).

[21] Ni­cola J Wiles et al., “Long-Term Effec­tive­ness and Cost-Effec­tive­ness of Cog­ni­tive Be­havi­oural Ther­apy as an Ad­junct to Phar­ma­cother­apy for Treat­ment-Re­sis­tant De­pres­sion in Pri­mary Care: Fol­low-up of the CoBalT Ran­domised Con­trol­led Trial,” The Lancet Psy­chi­a­try 3, no. 2 (Fe­bru­ary 1, 2016): 137–44, https://​​​​10.1016/​​S2215-0366(15)00495-2.

[22] La­yard and Clark, Thrive : The Power of Ev­i­dence-Based Psy­cholog­i­cal Ther­a­pies.

[23] For a longer dis­cus­sion of men­tal health and StrongMinds, see John Halstead and James Snow­den, “Cause Re­port—Men­tal Health,” n.d., https://​​founder­​​re­search/​​Cause Re­port—Men­tal Health.pdf.

[24] KC Cukrow­icz and TE Joiner, “Com­puter-Based In­ter­ven­tion for Anx­ious and De­pres­sive Symp­toms in a Non-Clini­cal Pop­u­la­tion,” Cog­ni­tive Ther­apy and Re­search, 2007, http://​​​​ar­ti­cle/​​10.1007/​​s10608-006-9094-x. E Kal­ten­thaler et al., “Com­put­er­ised Cog­ni­tive–be­havi­oural Ther­apy for De­pres­sion: Sys­tem­atic Re­view,” The Bri­tish Jour­nal Of, 2008, http://​​​​con­tent/​​193/​​3/​​181.short.

[25] An­drea Cipri­ani et al., “Com­par­a­tive Effi­cacy and Ac­cept­abil­ity of 21 An­tide­pres­sant Drugs for the Acute Treat­ment of Adults with Ma­jor De­pres­sive Di­sor­der: A Sys­tem­atic Re­view and Net­work Meta-Anal­y­sis.,” Lancet (Lon­don, England) 391, no. 10128 (April 7, 2018): 1357–66, https://​​​​10.1016/​​S0140-6736(17)32802-7. Irv­ing Kirsch et al., “Ini­tial Sever­ity and An­tide­pres­sant Benefits: A Meta-Anal­y­sis of Data Sub­mit­ted to the Food and Drug Ad­minis­tra­tion,” ed. Phillipa Hay, PLoS Medicine 5, no. 2 (Fe­bru­ary 26, 2008): e45, https://​​​​10.1371/​​jour­nal.pmed.0050045.

[26] Keith S. Dob­son et al., “Ran­dom­ized Trial of Be­hav­ioral Ac­ti­va­tion, Cog­ni­tive Ther­apy, and An­tide­pres­sant Med­i­ca­tion in the Preven­tion of Re­lapse and Re­cur­rence in Ma­jor De­pres­sion.,” Jour­nal of Con­sult­ing and Clini­cal Psy­chol­ogy 76, no. 3 (June 2008): 468–77, https://​​​​10.1037/​​0022-006X.76.3.468.

[27] Olivia F O’Leary, Ti­mothy G Di­nan, and John F Cryan, “Faster, Bet­ter, Stronger: Towards New An­tide­pres­sant Ther­a­peu­tic Strate­gies,” Euro­pean Jour­nal of Phar­ma­col­ogy 753 (April 2015): 32–50, https://​​​​10.1016/​​j.ejphar.2014.07.046. Mar­ije aan het Rot et al., “Safety and Effi­cacy of Re­peated-Dose In­tra­venous Ke­tamine for Treat­ment-Re­sis­tant De­pres­sion,” Biolog­i­cal Psy­chi­a­try 67, no. 2 (Jan­uary 15, 2010): 139–45, https://​​​​10.1016/​​J.BIOPSYCH.2009.08.038. Re­becca Brach­man, Could a Drug Prevent De­pres­sion and PTSD? | TED Talk, 2016, https://​​​​talks/​​re­becca_brach­man_could_a_drug_pre­vent_de­pres­sion_and_ptsd/​​tran­script?lan­guage=en.

[28] MAPS, “FDA Grants Break­through Ther­apy Des­ig­na­tion for MDMA-As­sisted Psy­chother­apy for PTSD, Agrees on Spe­cial Pro­to­col Assess­ment for Phase 3 Tri­als,” 2017, https://​​​​news/​​me­dia/​​6786-press-re­lease-fda-grants-break­through-ther­apy-des­ig­na­tion-for-mdma-as­sisted-psy­chother­apy-for-ptsd,-agrees-on-spe­cial-pro­to­col-as­sess­ment-for-phase-3-tri­als. Michael C Mithoefer et al., “Dura­bil­ity of Im­prove­ment in Post-Trau­matic Stress Di­sor­der Symp­toms and Ab­sence of Harm­ful Effects or Drug Depen­dency af­ter 3,4-Methylene­dioxymetham­phetamine-As­sisted Psy­chother­apy: A Prospec­tive Long-Term Fol­low-up Study.,” Jour­nal of Psy­chophar­ma­col­ogy (Oxford, England) 27, no. 1 (Jan­uary 2013): 28–39, https://​​​​10.1177/​​0269881112456611.

[29] Robin L Carhart-Har­ris et al., “Psilo­cy­bin with Psy­cholog­i­cal Sup­port for Treat­ment-Re­sis­tant De­pres­sion: An Open-La­bel Fea­si­bil­ity Study,” The Lancet Psy­chi­a­try 3, no. 7 (July 2016): 619–27, https://​​​​10.1016/​​S2215-0366(16)30065-7. See also D. E. Ni­chols, M. W. John­son, and C. D. Ni­chols, “Psychedelics as Medicines: An Emerg­ing New Paradigm,” Clini­cal Phar­ma­col­ogy and Ther­a­peu­tics 101, no. 2 (2017), https://​​​​10.1002/​​cpt.557.

[30] Michael Plant and Lee Sharkey, “High Time For Drug Policy Re­form. Part 1/​4: In­tro­duc­tion and Cause Sum­mary,” Effec­tive Altru­ism Fo­rum, 2017, http://​​effec­tive-al­tru­​​ea/​​1d8/​​high_time_for_drug_policy_re­form_part_14/​​.

[31] Sid­ney H. Kennedy et al., “Deep Brain Stim­u­la­tion for Treat­ment-Re­sis­tant De­pres­sion: Fol­low-Up After 3 to 6 Years,” Amer­i­can Jour­nal of Psy­chi­a­try 168, no. 5 (May 1, 2011): 502–10, https://​​​​10.1176/​​appi.ajp.2010.10081187.

[32] J Brunelin et al., “Effi­cacy of Repet­i­tive Tran­scra­nial Mag­netic Stim­u­la­tion (RTMS) in Ma­jor De­pres­sion: A Re­view,” L’En­cephale 33, no. 2 (2007): 126–34, https://​​​​10.1016/​​S0013-7006(07)91542-0.

[33] Ale­jan­dro Adler, “Teach­ing Well-Be­ing In­creases Aca­demic Perfor­mance: Ev­i­dence From Bhutan, Mex­ico, and Peru,” Publi­cly Ac­cessible Penn Disser­ta­tions, Jan­uary 1, 2016, https://​​repos­i­​​edis­ser­ta­tions/​​1572.

[34] I’ve put all the calcu­la­tions for the cost-effec­tive­ness of both GiveDirectly and StrongMinds, in­clud­ing refer­ences, in the fol­low­ing spread­sheet: Michael Plant, “Life Satis­fac­tion Im­pact of Treat­ing Men­tal Health vs Alle­vi­at­ing Poverty,” 2018, n.d., https://​​​​spread­sheets/​​d/​​1FcpfiP6P-nxJ7ilqOH0Vv-6TwYd­sWc2bOtWLvBmt8wM/​​edit#gid=0.

[35] Jo­hannes Haushofer, James Reis­inger, and Jeremy Shapiro, “Your Gain Is My Pain: Nega­tive Psy­cholog­i­cal Ex­ter­nal­ities of Cash Trans­fers, Work­ing Paper,” 2015.

[36] An­drew E. Clark, “Hap­piness, In­come and Poverty,” In­ter­na­tional Re­view of Eco­nomics 64, no. 2 (June 21, 2017): 145–58, https://​​​​10.1007/​​s12232-017-0274-7.

[37] Clark, An­drew E., Natavudh Powdthavee, Sarah Flèche, Richard La­yard, and Ge­orge Ward. The Ori­gins of Hap­piness : The Science of Well-Be­ing over the Life Course, 2018.

[38] Jo­hannes Haushofer and Jeremy Shapiro, “THE LONG-TERM IMPACT OF UNCONDITIONAL CASH TRANSFERS: EXPERIMENTAL EVIDENCE FROM KENYA,” 2018, http://​​jere­myp­​​pa­pers/​​Haushofer_Shapiro_UCT2_2018-01-30_pa­per_only.pdf.

[39] Available at: https://​​​​2018/​​05/​​04/​​new-re­search-on-cash-trans­fers/​​

[40] A list GiveWell top char­i­ties is available here: https://​​​​char­i­ties/​​top-charities

[41] Plant, “Life Satis­fac­tion Im­pact of Treat­ing Men­tal Health vs Alle­vi­at­ing Poverty.” https://​​​​spread­sheets/​​d/​​1FcpfiP6P-nxJ7ilqOH0Vv-6TwYd­sWc2bOtWLvBmt8wM/​​edit#gid=0. Note the im­pact is nearly iden­ti­cal whether an as­sum­ing of a con­stant benefit of four years (i.e. the in­fer­ence for Wiles et al. 2016) or if we as­sume a 75% an­nual re­ten­tion of benefits, which is the method taken by Halstead and Snow­den, “Cause Area Re­port: Men­tal Health”, Founders Pledge, https://​​founder­​​re­search/​​Cause%20Re­port%20-%20Men­tal%20Health.pdf

[42] GiveWell “2018 Cost-effec­tive­ness anal­y­sis -ver­sion 4” states that 2% of the cost-effec­tive­ness of the de­worm­ing char­i­ties (DtW, SCI, Sight­savers, END) comes from ‘short-term health effects’ and 98% from ‘even­tual in­come and con­sump­tion gains’. See Re­sults tab in this spread­sheet: https://​​​​spread­sheets/​​d/​​1moyxmsn4UjhH3CzFJmPwAN7LUAk­m­maoXDb6bdW3WILg/​​edit#gid=1364064522.

[43] Michael Plant, “Are You Sure You Want To Donate To The Against Malaria Foun­da­tion?,” Effec­tive Altru­ism Fo­rum, 2016, http://​​effec­tive-al­tru­​​ea/​​14k/​​are_you_sure_you_want_to_donate_to_the_against/​​.

[44] “The Meat-Eater Prob­lem—Effec­tive Altru­ism Con­cepts,” ac­cessed Septem­ber 21, 2018, https://​​con­cepts.effec­tivealtru­​​con­cepts/​​the-meat-eater-prob­lem/​​.

[45] Hilary Greaves, “Op­ti­mum Pop­u­la­tion Size,” in Oxford Hand­book of Pop­u­la­tion Ethics, ed. Ar­rhe­nius, Bykvist, and Camp­bell (Oxford Univer­sity Press, n.d.).

[46] GiveWell, “2018 GiveWell Cost-Effec­tive­ness Anal­y­sis — Ver­sion 4,” 2018, https://​​​​spread­sheets/​​d/​​1moyxmsn4UjhH3CzFJmPwAN7LUAk­m­maoXDb6bdW3WILg/​​edit#gid=1364064522.

[47] Hel­liwell, La­yard, and Sachs, World Hap­piness Re­port 2017 p28

[48] Plant, “Life Satis­fac­tion Im­pact of Treat­ing Men­tal Health vs Alle­vi­at­ing Poverty.” https://​​​​spread­sheets/​​d/​​1FcpfiP6P-nxJ7ilqOH0Vv-6TwYd­sWc2bOtWLvBmt8wM/​​edit#gid=0.

[49] For a sum­mary of the prob­lems of pop­u­la­tion ethics, see Hilary Greaves, “Pop­u­la­tion Ax­iol­ogy,” Philos­o­phy Com­pass, 2017. Note Greaves is un­sym­pa­thetic to per­son-af­fect­ing views.

[50] Jan Narve­son, “Mo­ral Prob­lems of Pop­u­la­tion,” The Mon­ist, 1973, 62–86.

[51] See Parfit, Rea­sons and Per­sons, part 4, 1984.

[52] Gus­tav Ar­rhe­nius, “An Im­pos­si­bil­ity The­o­rem for Welfarist Ax­iolo­gies,” Eco­nomics & Philos­o­phy 16, no. 2 (2000): 247–66.

[53] Gre­gory Lewis, “The Per­son-Affect­ing Value of Ex­is­ten­tial Risk Re­duc­tion,” Effec­tive Altru­ism Fo­rum, 2018, http://​​effec­tive-al­tru­​​ea/​​1n0/​​the_per­son­af­fect­ing_value_of_ex­is­ten­tial_risk/​​.

[54] Feli­cia Marie Knaul et al., “Alle­vi­at­ing the Ac­cess Abyss in Pal­li­a­tive Care and Pain Relief-an Im­per­a­tive of Univer­sal Health Cover­age: The Lancet Com­mis­sion Re­port.,” Lancet (Lon­don, England) 0, no. 0 (Oc­to­ber 11, 2017), https://​​​​10.1016/​​S0140-6736(17)32513-8.

[55] Plant and Sharkey, “High Time For Drug Policy Re­form. Part 1/​4: In­tro­duc­tion and Cause Sum­mary.”