How can you compare helping two different people in different ways?

When peo­ple ask what as­piring effec­tive al­tru­ists work on, I of­ten start by say­ing that we do re­search into how you can help oth­ers the most. For ex­am­ple, GiveWell has found that dis­tribut­ing some 600 bed nets, at a cost of $3,000, can pre­vent one in­fant dy­ing of malaria. For the same price, they have also found you could de­liver 6,000 de­worm­ing treat­ments that work for around a year.

A com­mon ques­tion at this point is ‘how can you com­pare the value of helping these differ­ent peo­ple in these differ­ent ways?’ Even if the num­bers are ac­cu­rate, how could any­one de­ter­mine which of these two pos­si­ble dona­tions helps oth­ers the most?

I can’t offer a philo­soph­i­cally rigor­ous an­swer here, but I can tell you how I per­son­ally ap­proach this puz­zle. I ask my­self the ques­tion:

  • Which would I pre­fer, if, af­ter mak­ing the de­ci­sion, I were equally likely to be­come any one of the peo­ple af­fected, and ex­pe­rience their lives as they would? [1]

Let’s work through this ex­am­ple. First, we’ll make the num­ber of peo­ple we are con­sid­er­ing a man­age­able num­ber: for $5, I could offer 10 chil­dren de­worm­ing treat­ments, or al­ter­na­tively offer 1 child a bed-net, which has a 1 in 600 chance of sav­ing their life. To make this de­ci­sion, I should com­pare three op­tions:

  1. I don’t donate, and so none of the 11 chil­dren re­ceive any help

  2. Ten of the chil­dren re­ceive de­worm­ing treat­ment, but the other one goes with­out a bed-net

  3. The one child re­ceives a bed-net, but the other ten go with­out deworming

If I didn’t know which of these 11 chil­dren I was about to be­come, which choice would be more ap­peal­ing?
Ob­vi­ously 2 and 3 are bet­ter than 1 (no help), but de­cid­ing be­tween 2 and 3 is not so sim­ple. I am con­fi­dent that a malaria net is more helpful than a de­worm­ing tablet, but it is ten times more use­ful?
This ques­tion has the virtue of:
  • Be­ing ‘fair’, be­cause in the­ory ev­ery­one’s in­ter­ests are given ‘equal con­sid­er­a­tion

  • Put­ting the fo­cus on how much the re­cip­i­ents’ value the help, rather than how you feel about it as a donor

  • Mo­ti­vat­ing you to ac­tu­ally try to figure out the an­swer, by putting you in the shoes of the peo­ple you are try­ing to help.

You’ll no­tice that this ap­proach looks a lot like the veil of ig­no­rance, a pop­u­lar method among moral philoso­phers for de­ter­mine whether a pro­cess or out­come is ‘just’. It should also be very ap­peal­ing to any con­se­quen­tial­ist who cares about ‘wellbe­ing’, and thinks ev­ery­one’s in­ter­ests ought to be weighed equally. [2] It also looks very much like the an­cient in­struc­tion to “love your neigh­bor as your­self”.
In my ex­pe­rience, this thought ex­per­i­ment pushes you to­wards ask­ing good con­crete ques­tions like:
  • How much would de­worm­ing im­prove my qual­ity of life im­me­di­ately, and then in the long term?

  • How harm­ful is it for an in­fant to die? How painful is it to suffer from a case of malaria?

  • What risk of death might I be will­ing to tol­er­ate to get the long-term health and in­comes gains offered by de­worm­ing?

  • And so on.

    I find the main weak­ness of ap­ply­ing this ap­proach is that thou­sands of peo­ple might be af­fected in some way by a de­ci­sion. For in­stance, we should not only con­sider the harm to young chil­dren who die of pre­ventable dis­eases, but also the grief and hard­ship ex­pe­rienced by their fam­i­lies as a re­sult. But that’s just the start: health treat­ments de­liv­ered to­day will change the rate of eco­nomic de­vel­op­ment in a coun­try and there­fore the qual­ity of life of all fu­ture gen­er­a­tions. A big part of the case for de­worm­ing is that it im­proves nu­tri­tion, and thereby raises ed­u­ca­tion lev­els and in­comes for peo­ple when they are adults—benefits that are then passed on to their chil­dren and their chil­dren’s chil­dren.
    This doesn’t make this ques­tion the wrong one to ask, but rather that track­ing and weigh­ing the im­pact on the hun­dreds of peo­ple who might be af­fected by an ac­tion is be­yond what most of us can do in a ca­sual way. How­ever, I find you can still make use­ful progress by think­ing through and adding up the im­pacts on pa­per, or in a spread­sheet. [3] When you ap­ply this ap­proach, it is usu­ally pos­si­ble to nar­row down your choices to just a few op­tions, though in my ex­pe­rience you may then not have enough in­for­ma­tion to con­fi­dently de­cide among that re­main­ing hand­ful.
    [1] A very similar, prob­a­bly equiv­a­lent, ques­tion is: Which would I pre­fer if, af­ter mak­ing the de­ci­sion, I then had to se­quen­tially ex­pe­rience the re­main­ing lives of ev­ery­one af­fected by both op­tions?
    [2] One weak­ness is that this ques­tion is am­bigu­ous about how to deal with in­ter­ven­tions that change who ex­ists (for in­stance, poli­cies that raise or lower birth rates). If you as­sume that you must be­come some­one—non-ex­is­tence is not an op­tion—you would end up adopt­ing the ‘av­er­age view’, which ac­tu­ally has no sup­port­ers in moral philos­o­phy. If you sim­ply ig­nored any­one whose ex­is­tence was con­di­tional on your de­ci­sion, you would be adopt­ing the ‘per­son af­fect­ing view’, which it­self has se­ri­ous prob­lems. If you do in­clude those peo­ple in the pop­u­la­tion of peo­ple you could be­come, and add ‘non-ex­is­tence’ as the al­ter­na­tive for the choices which cause those peo­ple not to ex­ist, you would be adopt­ing the ‘to­tal view’.
    [3] Alter­na­tively, if you were con­vinced that these long-term pros­per­ity effects were the most im­por­tant im­pact, and were similarly valuable across coun­tries, you could try to es­ti­mate the in­crease in the rate of eco­nomic growth per $100 in­vested in differ­ent pro­jects, and just seek to max­imise that.