# Efficient Charity: Do Unto Others...

This was origi­nally posted as part of the effi­cient char­ity con­test back in Novem­ber. Thanks to Roko, mul­ti­fo­li­aterose, Louie, jmmcd, jsal­vatier, and oth­ers I for­get for help, cor­rec­tions, en­courage­ment, and both­er­ing me un­til I fi­nally re­mem­bered to post this here.

Imag­ine you are set­ting out on a dan­ger­ous ex­pe­di­tion through the Arc­tic on a limited bud­get. The griz­zled old prospec­tor at the gen­eral store shakes his head sadly: you can’t af­ford ev­ery­thing you need; you’ll just have to pur­chase the bare es­sen­tials and hope you get lucky. But what is es­sen­tial? Should you buy the warmest parka, if it means you can’t af­ford a sleep­ing bag? Should you bring an ex­tra week’s food, just in case, even if it means go­ing with­out a rifle? Or can you buy the rifle, leave the food, and hunt for your din­ner?

And how about the field guide to Arc­tic flow­ers? You like flow­ers, and you’d hate to feel like you’re failing to ap­pre­ci­ate the harsh yet del­i­cate en­vi­ron­ment around you. And a digi­tal cam­era, of course—if you make it back al­ive, you’ll have to put the Arc­tic ex­pe­di­tion pics up on Face­book. And a hand-crafted scarf with au­then­tic Inuit tribal pat­terns wo­ven from or­ganic fibres! Wicked!

...but of course buy­ing any of those items would be in­sane. The prob­lem is what economists call op­por­tu­nity costs: buy­ing one thing costs money that could be used to buy oth­ers. A hand-crafted de­signer scarf might have some value in the Arc­tic, but it would cost so much it would pre­vent you from buy­ing much more im­por­tant things. And when your life is on the line, things like im­press­ing your friends and buy­ing or­ganic pale in com­par­i­son. You have one goal—stay­ing al­ive—and your only prob­lem is how to dis­tribute your re­sources to keep your chances as high as pos­si­ble. Th­ese sorts of eco­nomics con­cepts are nat­u­ral enough when faced with a jour­ney through the freez­ing tun­dra.

But they are de­cid­edly not nat­u­ral when fac­ing a de­ci­sion about char­i­ta­ble giv­ing. Most donors say they want to “help peo­ple”. If that’s true, they should try to dis­tribute their re­sources to help peo­ple as much as pos­si­ble. Most peo­ple don’t. In the “Buy A Brush­stroke” cam­paign, eleven thou­sand Bri­tish donors gave a to­tal of £550,000 to keep the fa­mous paint­ing “Blue Rigi” in a UK mu­seum. If they had given that £550,000 to buy bet­ter san­i­ta­tion sys­tems in Afri­can villages in­stead, the lat­est statis­tics sug­gest it would have saved the lives of about one thou­sand two hun­dred peo­ple from dis­ease. Each in­di­vi­d­ual $50 dona­tion could have given a year of nor­mal life back to a Third Wor­lder af­flicted with a dis­abling con­di­tion like blind­ness or limb de­for­mity.. Most of those 11,000 donors gen­uinely wanted to help peo­ple by pre­serv­ing ac­cess to the origi­nal can­vas of a beau­tiful paint­ing. And most of those 11,000 donors, if you asked, would say that a thou­sand peo­ple’s lives are more im­por­tant than a beau­tiful paint­ing, origi­nal or no. But these peo­ple didn’t have the proper men­tal habits to re­al­ize that was the choice be­fore them, and so a beau­tiful paint­ing re­mains in a Bri­tish mu­seum and some­where in the Third World a thou­sand peo­ple are dead. If you are to “love your neigh­bor as your­self”, then you should be as care­ful in max­i­miz­ing the benefit to oth­ers when donat­ing to char­ity as you would be in max­i­miz­ing the benefit to your­self when choos­ing pur­chases for a po­lar trek. And if you wouldn’t buy a pretty pic­ture to hang on your sled in prefer­ence to a parka, you should con­sider not helping save a fa­mous paint­ing in prefer­ence to helping save a thou­sand lives. Not all char­i­ta­ble choices are as sim­ple as that one, but many char­i­ta­ble choices do have right an­swers. GiveWell.org, a site which col­lects and in­ter­prets data on the effec­tive­ness of char­i­ties, pre­dicts that an­ti­malar­ial drugs save one child from malaria per$5,000 worth of medicine, but in­sec­ti­cide-treated bed nets save one child from malaria per $500 worth of net­ting. If you want to save chil­dren, donat­ing bed nets in­stead of an­ti­malar­ial drugs is the ob­jec­tively right an­swer, the same way buy­ing a$500 TV in­stead of an iden­ti­cal TV that costs $5,000 is the right an­swer. And since sav­ing a child from di­ar­rheal dis­ease costs$5,000, donat­ing to an or­ga­ni­za­tion fight­ing malaria in­stead of an or­ga­ni­za­tion fight­ing di­ar­rhea is the right an­swer, un­less you are donat­ing based on some crite­ria other than whether you’re helping chil­dren or not.

Say all of the best Arc­tic ex­plor­ers agree that the three most im­por­tant things for sur­viv­ing in the Arc­tic are good boots, a good coat, and good food. Per­haps they have run highly un­eth­i­cal stud­ies in which they re­lease thou­sands of peo­ple into the Arc­tic with differ­ent com­bi­na­tion of gear, and con­sis­tently find that only the ones with good boots, coats, and food sur­vive. Then there is only one best an­swer to the ques­tion “What gear do I buy if I want to sur­vive”—good boots, good food, and a good coat. Your prefer­ences are ir­rele­vant; you may choose to go with al­ter­nate gear, but only if you don’t mind dy­ing.

And like­wise, there is only one best char­ity: the one that helps the most peo­ple the great­est amount per dol­lar. This is vague, and it is up to you to de­cide whether a char­ity that raises forty chil­dren’s marks by one let­ter grade for $100 helps peo­ple more or less than one that pre­vents one fatal case of tu­ber­cu­lo­sis per$100 or one that saves twenty acres of rain­for­est per $100. But you can­not ab­di­cate the de­ci­sion, or you risk end­ing up like the 11,000 peo­ple who ac­ci­den­tally de­cided that a pretty pic­ture was worth more than a thou­sand peo­ple’s lives. De­cid­ing which char­ity is the best is hard. It may be straight­for­ward to say that one form of an­ti­malar­ial ther­apy is more effec­tive than an­other. But how do both com­pare to fi­nanc­ing med­i­cal re­search that might or might not de­velop a “magic bul­let” cure for malaria? Or fi­nanc­ing de­vel­op­ment of a new kind of su­per­com­puter that might speed up all med­i­cal re­search? There is no easy an­swer, but the ques­tion has to be asked. What about just com­par­ing char­i­ties on over­head costs, the one easy-to-find statis­tic that’s uni­ver­sally ap­pli­ca­ble across all or­ga­ni­za­tions? This solu­tion is sim­ple, el­e­gant, and wrong. High over­head costs are only one pos­si­ble failure mode for a char­ity. Con­sider again the Arc­tic ex­plorer, try­ing to de­cide be­tween a$200 parka and a $200 digi­tal cam­era. Per­haps a parka only cost$100 to make and the man­u­fac­turer takes $100 profit, but the cam­era cost$200 to make and the man­u­fac­turer is sel­l­ing it at cost. This speaks in fa­vor of the moral qual­ities of the cam­era man­u­fac­turer, but given the choice the ex­plorer should still buy the parka. The cam­era does some­thing use­less very effi­ciently, the parka does some­thing vi­tal in­effi­ciently. A parka sold at cost would be best, but in its ab­sence the ex­plorer shouldn’t hes­i­tate to choose the the parka over the cam­era. The same ap­plies to char­ity. An an­ti­malar­ial net char­ity that saves one life per $500 with 50% over­head is bet­ter than an an­tidi­ar­rheal drug char­ity that saves one life per$5000 with 0% over­head: $10,000 donated to the high-over­head char­ity will save ten lives;$10,000 to the lower-over­head will only save two. Here the right an­swer is to donate to the an­ti­malar­ial char­ity while en­courag­ing it to find ways to lower its over­head. In any case, ex­am­in­ing the fi­nan­cial prac­tices of a char­ity is helpful but not enough to an­swer the “which is the best char­ity?” ques­tion.

Just as there is only one best char­ity, there is only one best way to donate to that char­ity. Whether you vol­un­teer ver­sus donate money ver­sus raise aware­ness is your own choice, but that choice has con­se­quences. If a high-pow­ered lawyer who makes $1,000 an hour chooses to take an hour off to help clean up lit­ter on the beach, he’s wasted the op­por­tu­nity to work over­time that day, make$1,000, donate to a char­ity that will hire a hun­dred poor peo­ple for $10/​hour to clean up lit­ter, and end up with a hun­dred times more lit­ter re­moved. If he went to the beach be­cause he wanted the sun­light and the fresh air and the warm feel­ing of per­son­ally con­tribut­ing to some­thing, that’s fine. If he ac­tu­ally wanted to help peo­ple by beau­tify­ing the beach, he’s cho­sen an ob­jec­tively wrong way to go about it. And if he wanted to help peo­ple, pe­riod, he’s cho­sen a very wrong way to go about it, since that$1,000 could save two peo­ple from malaria. Un­less the lit­ter he re­moved is re­ally worth more than two peo­ple’s lives to him, he’s erring even ac­cord­ing to his own value sys­tem.

...and the same is true if his philan­thropy leads him to work full-time at a non­profit in­stead of go­ing to law school to be­come a lawyer who makes $1,000 /​ hour in the first place. Un­less it’s one HELL of a non­profit. The Ro­man his­to­rian Sal­lust said of Cato “He preferred to be good, rather than to seem so”. The lawyer who quits a high-pow­ered law firm to work at a non­profit or­ga­ni­za­tion cer­tainly seems like a good per­son. But if we define “good” as helping peo­ple, then the lawyer who stays at his law firm but donates the profit to char­ity is tak­ing Cato’s path of max­i­miz­ing how much good he does, rather than how good he looks. And this di­chotomy be­tween be­ing and seem­ing good ap­plies not only to look­ing good to oth­ers, but to our­selves. When we donate to char­ity, one in­cen­tive is the warm glow of a job well done. A lawyer who spends his day pick­ing up lit­ter will feel a sense of per­sonal con­nec­tion to his sac­ri­fice and re­live the mem­ory of how nice he is ev­ery time he and his friends re­turn to that beach. A lawyer who works over­time and donates the money on­line to starv­ing or­phans in Ro­ma­nia may never get that same warm glow. But con­cern with a warm glow is, at root, con­cern about seem­ing good rather than be­ing good—albeit seem­ing good to your­self rather than to oth­ers. There’s noth­ing wrong with donat­ing to char­ity as a form of en­ter­tain­ment if it’s what you want—giv­ing money to the Art Fund may well be a quicker way to give your­self a warm feel­ing than see­ing a ro­man­tic com­edy at the cin­ema—but char­ity given by peo­ple who gen­uinely want to be good and not just to feel that way re­quires more forethought. It is im­por­tant to be ra­tio­nal about char­ity for the same rea­son it is im­por­tant to be ra­tio­nal about Arc­tic ex­plo­ra­tion: it re­quires the same aware­ness of op­por­tu­nity costs and the same hard-headed com­mit­ment to in­ves­ti­gat­ing effi­cient use of re­sources, and it may well be a mat­ter of life and death. Con­sider go­ing to www.GiveWell.org and mak­ing use of the ex­cel­lent re­sources on effec­tive char­ity they have available. • Read­ing this and your ar­ti­cle on us­ing dead chil­dren as cur­ren­cies re­minds me of an event a few years ago which might have helped stop me from be­com­ing an­other re­li­gious nut­case. I did not know any­thing about ra­tio­nal­ity or util­i­tar­ian ethics at the time, and I was in­volved in a youth group at church that was go­ing to be mak­ing aid kits for Ethiopia. One of the items that was re­quested was some kind of cloth­ing, so I picked it up from a sec­ond hand store and put the kit to­gether. Later when we were talk­ing about the kits, I was told that we were only sup­posed to bring new items. when I asked why, the per­son in charge said some­thing about re­spect­ing the feel­ings of the peo­ple who were re­ceiv­ing the gifts, and want­ing them to feel like they had been given some­thing spe­cial, in­stead of a dis­carded item. Every­one else in the group seemed to ac­cept this eas­ily, but I asked how many more peo­ple we could have helped with bar­gain items. This time, they pretty much ig­nored what I had just said. I think this was the point when it fi­nally hit me that good in­ten­tions and ap­pear­ing kind are hor­rible in­di­ca­tors that you are re­ally mak­ing the world bet­ter. So any­way, I prob­a­bly would never have tried to find out about web­sites like this with­out my ex­pe­riences deal­ing with re­li­gion. Too bad we can­not all just be taught util­i­tar­ian ethics and ra­tio­nal­ity by our par­ents and school in­stead of dis­cov­er­ing them the hard way. • Ex­cuse my noob ques­tion, but isn’t your sub­tle anti-re­li­gion gen­er­al­iz­ing im­pli­ca­tion some­how ex­actly against the pro-ra­tio­nal at­ti­tude this web­site is spread­ing? Also, when it comes to util­i­tar­ian ethics and ra­tio­nal­ity or any­thing, isn’t “dis­cov­er­ing the hard way” more fruit-bear­ing than hav­ing to learn in schools? • Dis­cov­er­ing the hard way gen­er­ally leads to deeper knowl­edge, but it’s still ex­tremely im­por­tant to learn about, eg, the germ the­ory of dis­ease in school. You may not end up know­ing as much as its origi­nal dis­cov­er­ers about bac­te­ria and their be­hav­ior, but you can still spread a lot fewer dis­ease. • Also, when it comes to util­i­tar­ian ethics and ra­tio­nal­ity or any­thing, isn’t “dis­cov­er­ing the hard way” more fruit-bear­ing than hav­ing to learn in schools? In ad­di­tion to what drethe­lin said, there’s also the prob­lem that dis­cov­er­ing it the hard way is hard. Most peo­ple fail. That way bears no fruit at all. • Don’t be freaked out by my han­dle/​user­name: I like to blend the best of ra­tio­nal­ity with the best of feel­ing/​emo­tion, which I un­der­stand is what ev­ery avatar has done and does to get what I un­der­stand to be wis­dom, al­though their ad­her­ents no so much so. This post re­minds me of the (wrongly!) hated Repub­li­can, T.J. O’Rourke, from his book, Par­li­a­ment of Whores, where he de­scribed hav­ing spo­ken with peo­ple who lived in what he called, “A Sewer In the Sky,” and asked one of them who was say­ing things like, “the land­lords, they treat us like a dog when we ask for ser­vices,” to just imag­ine, what if she and her fel­low ten­ants could own the prop­erty for no more rent than they were presently pay­ing and then THE TENANTS could de­cide/​con­trol what hap­pened to them there and would no longer be sub­ject to any other au­thor­i­ties ex­cept the law, i.e., when and how re­pairs were made etc., and she looked at him, looked at the build­ing and said some­thing to the effect of “I ain’t goin’ for none of that!” This is most peo­ple in our world, I have found re­peat­edly over years of try­ing to help the “less for­tu­nate!” Another for in­stance is that TV shows are blar­ing how there are no jobs, and peo­ple are liv­ing pay­check to pay­check, etc. and when I ap­proach them and tell them that I have a trade­marked sys­tem that can help peo­ple over­come such “hand­i­caps,” and that I have plenty of refer­ences to help peo­ple get jobs, even ex-con­victs that won’t hold their records against them, they block me from be­ing able to post on their blogs, like the re­doubtable, Bill Moy­ers and Com­pany of PBS that has such a rep­u­ta­tion for trust, so they say! From what I can tell from most of the work­ers that I have to deal with daily and they are many in all spheres from Wal­green’s to my health­care providers, the vast ma­jor­ity of work­ers are in­com­pe­tent and have rot­ten at­ti­tudes! If they were do­ing such a great job, why weren’t they re­tained when the rest of the peo­ple were? Em­ploy­ers, though not always fair sup­pos­edly are not stupid ei­ther! Every­body wants the best, but most don’t want to give it to get it! Another ex­am­ple is a friend of al­most 20 years whom I’ve re­cently had to sep­a­rate my­self from be­cause I was be­com­ing a nasty, snarling, ex­traor­di­nar­ily an­gry per­son when faced with what I un­der­stand to be his delu­sional think­ing which has only got­ten worse re­cently. One of our typ­i­cal con­ver­sa­tions was that he told me that ev­ery­one he talks to, most of whom I call “Soup kitchen denizens” agree with him that above and be­yond the nu­tri­tional la­bel­ing on our food, they have a “right” to know what’s in their food and in a democ­racy, the ma­jor­ity rules, which lat­ter is true; how­ever, he ne­glected to un­der­stand that Amer­ica is NOT a democ­racy but a re­pub­lic, the lat­ter of which I un­der­stand is a so­ciety in which the minor­ity opinion is also sup­posed to be con­sid­ered in de­ci­sion mak­ing. Fur­ther, many ad­vance guard doc­tors say that one can eas­ily and sim­ply nav­i­gate our pre­sent food la­bel­ing for the most part by not eat­ing any­thing that has more than five in­gre­di­ents listed on its la­bel or any in­gre­di­ents that one doesn’t know ex­actly what they are; this in mod­er­a­tion is rel­a­tively health­ful. Sounds sen­si­ble and bal­anced to me! I told my friend in an­swer to this that Amer­i­cans are pretty much the only ones who think they have a right to know what’s in their food and most peo­ple in other parts of the world KNOW they have no such right un­less they grow the food them­selves. I also sug­gested that he could do so since he is go­ing to be so damned per­snick­ety, which of course he doesn’t have a clue how to do, e.g., what seed would be best, how to ro­tate same, crops, top soil, soil nu­tri­ents, etc. and he has no in­come what­so­ever and there­fore can’t buy ANY food (!) eats at soup kitchens three times a day or when I fed him, lav­ishly and re­peat­edly out of my in­come, and can’t tell me why he doesn’t have any in­come nor how he pays his rent! It’s some­thing hav­ing to do with the “Will the God!” I even gave this dude a work­ing com­puter FREE of charge (he has great ideas which could serve a lot of peo­ple and even make him a name which would last be­yond his death on al­ter­na­tive tech­nolo­gies which he has amassed on said com­puter, but he ab­solutely will not listen to rea­son and be prac­ti­cal about how to go about im­ple­ment­ing such ideas, which I tried to tell him in de­tail even writ­ing them down!) and all he ever did was com­plain that I wouldn’t listen to most of his bel­ly­ach­ing! Fool, me! I feel re­spon­si­ble that I am at­tract­ing this type of per­son of­ten, but for the most part, there are quite a lot of such over­grown ado­les­cents and I too was one for a very long time. I tell these peo­ple how I over­came that hand­i­cap but they’re in­ter­ested in be­ing pro­fes­sional vic­tims. That’s how they get their at­ten­tion with the, “Oh, woe is me!” syn­drome. Fi­nally, is there any­one here who would like to see my sell sheets for my health­ful, deli­cious food line, in­clud­ing the two most pow­er­ful su­per-foods I’ve heard of, which I fed to the above-men­tioned char­ac­ter for most of last year, again out of loy­alty and mis­guided pity, and which seemed to be the only thing he was satis­fied with, ask­ing me for more and more even when we were both livid with each other, as do many oth­ers? If so, please con­tact me. I’ve also started a ra­tio­nally based church like none I’ve ever heard of be­fore in recorded his­tory (Google: “Free Church of the Div­ine Mar­riage” [it’s not what you may think from its name, I’m pretty sure!] to learn more and to con­tact me from there) as well as many other ac­com­plish­ments un­der ex­treme duress from non-ra­tio­nal peo­ple! I hope this finds you hav­ing a pleas­ant day! • I’m go­ing to as­sign this to my in­tro­duc­tory microe­co­nomics stu­dents to help them un­der­stand op­por­tu­nity costs. • That sort of ter­rifies me, but in a good way. At the risk of toot­ing my own horn, this es­say only in­ci­den­tally ad­dresses op­por­tu­nity costs, but I wrote an­other es­say a few years ago in a differ­ent style that ad­dresses them more di­rectly: A Modest Proposal • I dis­cussed the ideas in this es­say with my stu­dents. I first ask my stu­dents how much an iPad costs. They give me some dol­lar amount, but then I say some­thing like “I don’t want the an­swer in dol­lars but rather in dead Afri­can chil­dren.” Since we have just been dis­cussing op­por­tu­nity costs they catch on quickly to what I’m get­ting at. • Are we count­ing re­sale value, and are we buy­ing new or used? That makes rather a lot of differ­ence. • Have you con­sid­ered sub­mit­ting your es­say to LW? It might not fit the gen­eral ob­jec­tive perfectly well, but I be­lieve it should be pro­moted and that many peo­ple would en­joy read­ing it. That said, I have to thank you for all your great posts. It is a plea­sure to read them. Be­ing clear and con­cise you provide valuable in­sights while dis­solvig im­por­tant top­ics. • I’d cer­tainly up­vote any such sub­mis­sion. I mean: “Not like I am any saint my­self. The past two years, I’ve spent about two dead pup­pies on books from Ama­zon.com alone. I am prob­a­bly go­ing to spend very close to a whole dead child to fly home for my two week win­ter break, and I spent ten dead chil­dren on my trip around the world this sum­mer. I spent four in­fected wounds on fan­tasy map-mak­ing soft­ware. But at least in the back of my mind I re­al­ize I’m do­ing it. Can the peo­ple who spend a dead kid plus a dead puppy on the world’s most ex­pen­sive sun­dae say the same? What about the Ja­panese guy spend­ing 1050 dead kids on a mo­bile phone strap?” Come on! • I can’t fol­low, are you be­ing sar­cas­tic about my sug­ges­tion? I guess it’s a mat­ter of taste. I thought the es­say shows how our util­ity calcu­la­tions are eas­ily in­fluenced by high­light­ing the po­ten­tial of the fuel that is money. Most peo­ple just use their money to feed a fire for its warmth and the beau­tiful sparks. They do not re­al­ize that ev­ery ban­knote is worth more than the printed pa­per it is made up of. Peo­ple do not see that a ban­knote can be used much more effec­tively. Re­nam­ing money is sim­ple yet changes its per­ceived po­ten­tial dra­mat­i­cally. As such the es­say is a metaphor to cau­tion against the burn­ing of books that is fuel­ing the fire of ig­no­rance. Do not burn books if not ab­solutely nec­es­sary, use the po­ten­tial effec­tively, read them! • I wasn’t be­ing sar­cas­tic. The im­plied ex­pan­sion of my last com­ment is ‘Come on [, how can you not like or ap­pre­ci­ate that para­graph among oth­ers?]!’ • You know… that ac­tu­ally seems like po­ten­tially a good idea. Not just a tongue in cheek style good idea, but I’m think­ing that this could be an ac­tu­ally for real good idea, and not just as a way to make “those other peo­ple” see what they’re do­ing. I’d want this im­ple­mented as a way to make it eas­ier for me to keep such things in mind! (The “in­fected wounds” link is bro­ken, though, so mind ex­plain­ing the con­cept re that?) The only real difficulty that I see is that as things change (tech, eco­nomic con­di­tions, etc), the ac­tual cost of sav­ing a child and the rel­a­tive costs of sav­ing a child vs sav­ing pup­pies, etc might shift around. So you’d need some way to dy­nam­i­cally re­name chunks of the cur­rency. For in­stance, if im­prov­ing tech and such leads to the equiv­a­lent of 400$ be­ing suffi­cient to save a child, then what was called a DC would have to be re­named 2 DC.

This would be con­fus­ing.

• The “in­fected wound” origi­nally linked to some or­ga­ni­za­tion that donated first aid kits to those who couldn’t af­ford them. I’ll try to fix that next time I up­date the site.

• Ah, okie, thanks.

• Some seast­ead some­ht­ing should try this I think.

• Float “dead chil­dren” as a cur­rency and reg­u­late that all prices must be ex­pressed in US dol­lars and time-of-pric­ing equiv­a­lent value of dead chil­dren. Deter­mine the ex­change rate not through any nor­mal cur­rency con­cerns but strictly through the change in how many lives US dol­lars save.

Pass­ing a law that does some­thing like this seems al­most fea­si­ble.

• Hrm… That might work

For all shops that don’t yet use elec­tronic price tag type things, there’d ob­vi­ously have to be a grace pe­riod along the lines of hav­ing a week/​month/​what­ever to up­date the equiv­alences (due to changes in the ex­change rate)

Of course, a rather uglier prob­lem would be: “How do we man­age to pro­tect the equiv­alences calcu­la­tion from ex­treme poli­ti­ciza­tion and such?”

Even worse: How do we avoid busi­nesses get­ting to­gether to try to sab­o­tage the efforts of effi­cient char­i­ties, that way lead­ing to a higher dol­lar per child amount. ie, the less dead chil­dren per item, the more will­ing some­one would be to pur­chase it, so there’s a bit of a per­verse in­cen­tive there.

Fi­nally, if we solved all this: how do we push to make it a re­al­ity?

• there’d ob­vi­ously have to be a grace pe­riod along the lines of hav­ing a week/​month/​what­ever to up­date the equiv­alences (due to changes in the ex­change rate)

It doesn’t even need that, just when­ever a price is printed out, it needs the equiv­a­lent in dead chil­dren, at the time of print­ing. This is an in­cen­tive to change or reprint prices when the value of dead chil­dren rises and leave old la­bels alone when the value of dead chil­dren has dipped, but as long as the value of dead chil­dren doesn’t fluc­tu­ate wildly (ie it doesn’t re­spond to spec­u­la­tion about a new dirt cheap cure, only to ex­ten­sive statis­tics on the cur­rent cost to save a child) then it should be mostly right.

The per­verse in­cen­tives, poli­ti­cal in­fluence, and po­ten­tial for Good­hart’s Law and lost pur­poses to come into play are all se­ri­ous con­cerns—all the more ter­rify­ing be­cause these surely play a part in cur­rent aid schemes.

...

You would need some kind of X-Ra­tion­al­ist Re­serve Bank of Dead Chil­dren who re­cite the Li­tany of Tarski (“If this change is to the truth­ful value of dead chil­dren, I de­sire to make this change. If this change is not to the truth­ful value of dead chil­dren, I de­sire not to make this change”) ev­ery morn­ing, and have an in­ves­tiga­tive group em­pow­ered to seek out and pun­ish in­terfer­ence in char­i­ta­ble work, prefer­ably in the form of huge fines payable to the af­fected char­i­ties (the Per­verse In­cen­tive Dis­in­cen­tives Task Force).

Fi­nally, if we solved all this: how do we push to make it a re­al­ity?

Yvain for Pres­i­dent?

• I wrote a post with your Modest Pro­posal as a jump­ing-off point.

• Thanks for the link, and I agree with pretty much all of what you said.

• First off, I don’t like your sug­ges­tion for smaller cur­ren­cies. If you want to deal with frac­tions of a life, just use life-years, life-days, etc.

Se­cond, you spent ten dead chil­dren on a va­ca­tion? Keep­ing it in the back of your head doesn’t mat­ter. What mat­ters is the peo­ple who are still al­ive. A group of peo­ple that’s ten smaller, thanks to your de­ci­sion.

• There are much more worth­while things to do than ac­cuse some­one who is mak­ing a rea­son­able and ra­tio­nal effort to bet­ter the world and help oth­ers do so as well of be­ing in­suffi­ciently self-sac­ri­fi­cial.

• A group of peo­ple that’s also likely hun­dreds of peo­ple larger by his di­rect de­ci­sions, and prob­a­bly thou­sands or tens of thou­sands of peo­ple larger by the de­ci­sions he in­fluenced.

• This week on Face­book, Derek Sivers (founder of CD Baby) wrote that this ar­ti­cle had more im­pact on him than any­thing else he read all year. He said: “Of all the ar­ti­cles I’ve read in the past 6 months, this one had the biggest im­pact on me.”

• I hadn’t seen that! Thanks for bring­ing it up.

• Pretty much a corol­lary of this is Steve Lands­burg’s (for some rea­son con­tro­ver­sial) point that you should only ever be donat­ing money to one char­ity at a time (un­less you’re ridicu­lously rich). The char­ity which makes the best use out of your first $1 dona­tion is al­most cer­tainly also the char­ity which makes the best use out of your 1000th dol­lar as well. Once you’ve done the calcu­la­tion, spread­ing your money be­tween differ­ent char­i­ties isn’t hedg­ing your bets, it’s giv­ing money to the wrong char­ity. See his Slate ar­ti­cle for a slightly more fleshed out ver­sion of the rea­son­ing. • There is one ex­cep­tion to this, which is poli­ti­cal char­i­ties (ACLU, for in­stance). Giv­ing to poli­ti­cal char­i­ties, has a sig­nal­ling effect: a poli­ti­cal char­ity can say “we have twelve mil­lion donors,” and this tells poli­ti­ci­ans that they had bet­ter listen to that char­ity or those twelve mil­lion peo­ple might be vot­ing for some­one else. That said, a$10 dona­tion is enough to get this effect.

• The ad­vice I hear is “limit your­self to three char­i­ties”- use­ful be­cause it al­lows you to broaden your fuzzies (like sup­port­ing eco­nomic liberty and cute an­i­mals and 3rd world de­vel­op­ment) while sig­nifi­cantly de­creas­ing the over­head costs to the char­i­ties. They would much rather have a $1,000 donor than 10$100 donors, es­pe­cially if that donor has made an an­nual com­mit­ment.

• Is that com­pat­i­ble with points five and six here, or is it a stand­ing dis­agree­ment among ac­tivists?

• I sus­pect that SIAI is in a differ­ent po­si­tion from most char­i­ties.

I don’t know what per­centage of char­i­ties are low on pub­lic sup­port, but I sus­pect that is not a se­ri­ous is­sue for most donors, as most donors couldn’t provide more than 2% of a char­ity’s to­tal in­come, even with a third of their to­tal char­ity bud­get.

Most char­i­ties have a prac­tice of send­ing end­less streams of junk mail, and so for most char­i­ties a gift of a few dol­lars is ac­tu­ally a los­ing propo­si­tion in the long term, since you sent the sig­nal you would be re­cep­tive to fu­ture dona­tion re­quests but don’t ac­tu­ally send more money. The SIAI’s strat­egy (and costs for emailing) are differ­ent from most char­i­ties, sug­gest­ing that differ­ent ad­vice makes sense for them.

• I ac­tu­ally tend to ar­gue this point first, and the more gen­eral point about effi­cient char­ity sec­ond. I’m not sure if that’s the most effec­tive way to ar­gue it though.

• I sus­pect con­vinc­ing peo­ple op­ti­mal philan­thropy is a good idea is prob­a­bly one of the most im­por­tant things one could do. Maybe you should find out?

• What about the idea of min­i­miz­ing the pos­si­bil­ity of po­ten­tial failure?

• From the ar­i­cle: “CARE is a no­ble or­ga­ni­za­tion that fights star­va­tion. It would like your sup­port. The Amer­i­can Cancer So­ciety is a no­ble or­ga­ni­za­tion that fights dis­ease. It would like your sup­port, too. Here’s my ad­vice: If you’re feel­ing very char­i­ta­ble, give gen­er­ously—but don’t give to both of them. … Giv­ing to ei­ther agency is a choice at­tached to a clear moral judg­ment. When you give $100 to CARE, you as­sert that CARE is wor­thier than the can­cer so­ciety. Hav­ing made that judg­ment, you are morally bound to ap­ply it to your next$100 dona­tion.”

Lands­burg is wrong, and here’s why. Be­cause the world is shades of gray, not black and white. It’s not clear what the best char­ity is, even by one’s own stan­dards (partly be­cause those stan­dards are not clear, and they some­times con­flict with each other). We know our­selves well enough to know we’re not smart enough to make those judg­ments perfectly, so we don’t bother with try­ing for perfec­tion, but rather with mak­ing sure to do at least some good. It’s hedg­ing our bets know­ing that some of the money is go­ing to the “wrong” char­ity (we’re just not sure which one is “wrong”).

• This de­pends on why you’re donat­ing. If you’re try­ing to get the men­tal state of feel­ing ‘I did some good’ or you’re try­ing to im­press your friends/​fam­ily (cer­tainly le­gi­t­i­mate goals), then this logic might work. If you’re try­ing to help peo­ple as much as pos­si­ble then this logic does not work be­cause the amount you ex­pect to have helped peo­ple rises lin­early with the prob­a­bil­ity that the char­ity helps.

One com­mon strat­egy is to spend a small­ish amount of money giv­ing to var­i­ous causes that make us feel good and/​or im­press oth­ers and a larger amount of money on a sin­gle char­ity op­ti­mized for helping peo­ple as much as pos­si­ble.

• Also, to quote text sim­ply use the mod­ifier >

• I can’t see any flaws in the ar­gu­ment, but the con­clu­sion is far more rad­i­cal than most of us would be will­ing to ad­mit.

Am I the sort of per­son who would value my com­puter over an­other hu­man be­ing’s life? I hope not, that makes me sound like the most hor­rible sort of psy­chopath—it is ba­si­cally the moral­ity of Stalin. But at the same time, did I sell my com­puter to feed kids in Africa? I did not. Nor did any of you, un­less you are read­ing this at a library com­puter (in which case I’m sure I can find some­thing you could have given up that would have al­lowed you to give just a lit­tle bit more to some wor­thy char­i­ta­ble cause.)

It gets worse: Is my col­lege ed­u­ca­tion worth the lives of fifty starv­ing chil­dren? Be­cause I surely paid more than that. Is this house I’m liv­ing in worth eight hun­dred life-sav­ing mosquito nets? Be­cause that’s how much it cost.

Our en­tire eco­nomic sys­tem is based on pur­chases that would be “un­jus­tified”—even im­moral—on the view that ev­ery sin­gle pur­chase must be made on this kind of met­ric. And so if we all stopped do­ing that, our econ­omy would col­lapse and we would be starv­ing in­stead.

I think it comes down to this: Con­se­quen­tial­ism is a lot harder than it looks. It’s not enough to use the sim­ple heuris­tic, “Is this pur­chase worth a child’s life?”; no, you’ve got to carry out the full sys­tem of con­se­quences—in prin­ci­ple, prop­a­gated to our whole fu­ture light cone. (In fact, there’s a very good rea­son not to ask that ques­tion: Be­cause of our so­cial­iza­tion, we have a taboo in our brains about never say­ing that some­thing is worth more than a child—even when it ob­vi­ously is.) You’ve got to note that once the kid sur­vives malaria, he’ll prob­a­bly die of some­thing else, like malnu­tri­tion, or HIV, or a par­a­site in­fec­tion. You’ve got to note that if peo­ple didn’t go to col­lege and be­come sci­en­tific re­searchers, we wouldn’t even know about HIV or malaria or any­thing else. You’ve got to keep in mind the whole sys­tem of mod­ified cap­i­tal­ism and the so­cial demo­cratic welfare state that makes your mas­sive wealth pos­si­ble—and re­ally, I think you should be try­ing to figure out how to ex­port it to places that don’t have it, not skim­ming off the in­come that drives it to save one child’s life at a time.

And if you think, “Ah ha! We’ll just work for the Sin­gu­lar­ity then!” well, that’s a start—and you should, in fact, de­vote some of your time, en­ergy, and money to the Sin­gu­lar­ity—but it’s not a solu­tion by it­self. How much time should you spend try­ing to make your­self happy? How much effort should you de­vote to your fam­ily, your friends? How im­por­tant is love com­pared to what you might be do­ing—and how much will your effec­tive­ness de­pend on you be­ing loved? We might even ask: Would we even want to make a Sin­gu­lar­ity if it meant that no one ever fell in love?

This is why I’m not quite a gung-ho con­se­quen­tial­ist. Ul­ti­mately con­se­quen­tial­ism is right, there can be no doubt about that; but in prac­ti­cal terms, I don’t think most peo­ple are smart enough for it. (I’m not sure I’m smart enough for it.) It might be bet­ter, ac­tu­ally, to make peo­ple fol­low sim­ple rules like “Don’t cheat, don’t lie, don’t kill, don’t steal”; if ev­ery­one fol­lowed those rules, we’d be do­ing all right. (Most of the re­ally hor­rible things in this world are de­on­tic vi­o­la­tions, like tyranny and geno­cide.) At the very least, the stan­dard de­on­tic rules are bet­ter heuris­tics than ask­ing, “Is it worth the life of a child?”

• This is a su­per-duper nice com­ment.

Most of the re­ally hor­rible things in this world are de­on­tic vi­o­la­tions, like tyranny and geno­cide.

Disagree. Most of the re­ally hor­rible things in this world are just ac­ci­dents that not enough peo­ple are pay­ing at­ten­tion to. If an­i­mals can suffer then mil­lions of Holo­causts are hap­pen­ing ev­ery day. If in­sects can suffer then tens of billions are. In any case hu­mans can cer­tainly suffer, and they’re do­ing plenty of that from pure ac­ci­dent. Prob­a­bly less than a twen­tieth of hu­man suffer­ing is in­ten­tion­ally caused by other hu­mans. (Though I will say that the ab­solute mag­ni­tude of hu­man-in­tent-caused hu­man suffer­ing is un­be­liev­ably huge.)

• Upvoted. I re­ally like this com­ment be­cause it shows some of my own con­cerns about con­se­quen­tial­ism. For ex­am­ple I have de­cided that for most cases the de­on­tic an­swers fit the con­se­quen­tial­ist ones so well that we should start out fol­low­ing them and only if they ap­pear to be non­satis­fac­tory we should dive into con­se­quen­tial­ist rea­son­ing. This quite leads to some peace of mind, but it ob­vi­ously is the easy an­swer, not the cor­rect one… Is there a post on less­wrong for de­on­tol­ogy as a sub­set of con­se­quen­tial­ism? (Ac­cord­ing to wikipe­dia there seem to be some sci­en­tists that state a similar opinion.)

• The util­i­tar­ian philoso­pher RM Hare has pro­posed a solu­tion along the lines you sug­gest, it’s called two-level util­i­tar­i­anism. From Wikipe­dia:

As a de­scrip­tive model of the two lev­els, Hare posited two ex­treme cases of peo­ple, one of whom would only use crit­i­cal moral think­ing and the other of whom would only use in­tu­itive moral think­ing. The former he called the ‘ar­changel’ and the lat­ter the ‘prole’.

I think the con­cept has merit, but if you’re smart and will­ing enough to do it, you’d have to act ac­cord­ing to the “crit­i­cal level” (con­ven­tional con­se­quen­tial­ism) any­way.

• we should start out fol­low­ing them and only if they ap­pear to be non­satis­fac­tory we should dive into con­se­quen­tial­ist rea­son­ing.

Your ac­tual val­ues are the ones that de­ter­mine “what ap­pears satis­fac­tory”.

• Of course, that’s why I would call my­self a con­se­quen­tial­ist even though I mainly/​very of­ten ar­gue by us­ing de­on­tic prin­ci­ples. I wasn’t talk­ing about the­ory (or foun­da­tion), but about the prac­ti­cal­ity/​prac­ti­cal use of de­on­tic rea­son­ing ver­sus con­se­quen­tial­ism.

• I have yet to fa­mil­iarize my­self more with effec­tive al­tru­ism to know the de­tails of their met­rics, but it seems like the re­li­ance on ‘num­ber of lives saved per unit money’ doesn’t nec­es­sar­ily al­ign with the goal of helping peo­ple, which i think this post demon­strates well. And then there’s the ar­guably rele­vant is­sue of over-pop­u­la­tion. If ev­ery­one con­tributed some of their ed­u­ca­tion fund­ing on sav­ing lives, wouldn’t the Earth get over-pop­u­lated be­fore suffi­cient tech­nolog­i­cal progress was made to e.g in­habit an­other planet?

• I find I run into a co­nun­drum on this ques­tion, be­cause there is a bias I fear over­com­pen­sat­ing for. I know as a hu­man that I am bi­ased to care more about the one per­son stand­ing in front of me than those ten thou­sand peo­ple starv­ing in In­dia that I’ll never meet, but I find it difficult to ap­ply that in­for­ma­tion. I know that donat­ing money to, say, those malaria nets, will prob­a­bly save more lives than donat­ing to, say, my lo­cal food pantry. By these ar­gu­ments, it seems that that fact should trump all, and I should donate to those malaria nets.

How­ever, I know that my lo­cal food pantry is an or­ga­ni­za­tion that feeds peo­ple who re­ally need food, that it has vir­tu­ally no over­head, and that there are chil­dren who would be mal­nour­ished with­out it. I also know that there are peo­ple all over the world who will con­tribute to malaria nets, but it is highly un­likely that any­one out­side my com­mu­nity will con­tribute to my lo­cal food pantry.

I agree that it is vi­tally im­por­tant to think care­fully about how we spend our char­ity money, and I un­der­stand that the difficulty I am hav­ing with this topic is an in­di­ca­tion that I need to think more deeply on it, but I keep com­ing up against two ba­sic is­sues.

• There is no sim­ple met­ric for “most good done.” What if one dis­ease costs lit­tle to pre­vent death, but leaves sur­vivors crip­pled, while an­other costs much more to pre­vent death but leaves peo­ple healthy? Should I donate to the first, and bur­den the com­mu­ni­ties with many crip­ples, or to the sec­ond, and let peo­ple die? With food and med­i­cal care cost­ing more in the de­vel­oped world, should I only donate to help those in the un­de­vel­oped world, where my dol­lar will go farther?

• Should I feel guilty for donat­ing money to pub­lic ra­dio be­cause it doesn’t save chil­dren? No. My pur­pose in donat­ing money to pub­lic ra­dio is to keep my fa­vorite shows on the air, and my dona­tions do that very effi­ciently. Yes, the money could go to save chil­dren, but so could the money I use to pay my ca­ble bill. I should per­haps not con­sider it as char­ity the way I do a dona­tion that saves chil­dren, but I should not feel guilty. If I have $500 al­lo­cated for en­ter­tain­ment and$500 al­lo­cated for char­ity, per­haps it should come out of the former. How­ever, it would be dis­in­gen­u­ous to say that dona­tions for more frivolous causes, such as sav­ing art­work, could be donated to bet­ter causes, such as malaria nets, un­less we also point out that what we spent on our fancy din­ner or our new dress or go­ing to the movies could also be thus al­lo­cated.

• The sec­ond point is some­thing that re­ally gets me. It seems to me that rather than feel­ing bad about donat­ing to one char­ity rather than a more effi­cient or more “im­por­tant” other char­ity, we should feel bad about spend­ing money on frivoli­ties rather than donat­ing to char­ity. Non­profit or­ga­ni­za­tions are forced to com­pete against each other for slen­der re­sources in many ways, in­clud­ing donor dol­lars—why can’t they com­pete against things that have less moral value in­stead? It would be awe­some if there were more so­cial pres­sure to donate to char­ity rather than go­ing to the movies or buy­ing pretty clothes.

In­ter­est­ingly, how­ever, there is some so­cial stigma against donat­ing “too much”. A few years ago, there was a New York gen­tle­man who donated a much larger than “nor­mal” per­centage of his money to char­ity, as well as his kid­ney, plus some other stuff. (I’m sorry, I re­ally wish I could re­mem­ber his name, but I am very sure I have these de­tails cor­rect, be­cause I read a lot about it at the time.) Peo­ple spec­u­lated in the press about his men­tal sta­tus and other chil­dren mocked his kids at school, al­though his fam­ily was hardly left poor by the ex­pe­rience, and his health was not en­dan­gered.

In terms of the point in the OP about the lawyer who should be work­ing over­time rather than vol­un­teer­ing … I strug­gle with this so much. I spend most of my time do­ing ac­tivism, and I have friends who spend more time than I do (who do things like take very low-pay­ing part-time jobs in or­der to fi­nance spend­ing most of their time do­ing ac­tivism), but most of us are sex-pos­i­tive ac­tivists, and sex-pos­i­tive ac­tivism is ar­guably an ex­tremely “low pri­or­ity” type of ac­tivism. If we are con­cerned about sav­ing more lives, for ex­am­ple, then we should be ded­i­cat­ing our time to other types of ac­tivism, or we should be us­ing our in­tel­li­gence to get awe­some jobs and then spend­ing the money on char­ity. How­ever, I (for one) have tried ded­i­cat­ing all my time to do­ing ac­tivism that seemed “more im­por­tant” (HIV in Africa) rather than the ac­tivism that is most in­ter­est­ing to me (var­i­ous types of sex­u­al­ity stuff in Amer­ica), and I was both less happy and less effec­tive. I am also very sure that I would be un­happy if I ded­i­cated my con­sid­er­able IQ to be­com­ing a cor­po­rate bitch and then donat­ing lots of money, rather than work­ing di­rectly on the is­sues I care about.

Ad­di­tion­ally, it is un­de­ni­able that some­one has to work on the is­sues I care about, or else who would I donate money to even if I had a lot of it?

• I (for one) have tried ded­i­cat­ing all my time to do­ing ac­tivism that seemed “more im­por­tant” (HIV in Africa) rather than the ac­tivism that is most in­ter­est­ing to me (var­i­ous types of sex­u­al­ity stuff in Amer­ica), and I was both less happy and less effective

There’s a story I like to tell when I hear this. Louise and Claire are both con­cerned about global warm­ing. Louise is full of pas­sion for the sub­ject and does what moves her most; through her hard work per­suades a thou­sand peo­ple to un­plug their phone charg­ers at night. Claire can’t get worked up about it even though she un­der­stands it’s im­por­tant; in a drunken con­ver­sa­tion one night she per­suades one friend to turn down their cen­tral heat­ing one de­gree.

Claire’s choice of an effi­cient way to re­duce CO2 emis­sions ab­solutely swamps the differ­ence in en­thu­si­asm; she does con­sid­er­ably more good than Louise.

• This makes me won­der if giv­ing out free cloth­ing vouch­ers in win­ter might be an effec­tive global warm­ing hack.

• Things like this are why I love Less­wrong.

• we should feel bad about spend­ing money on frivoli­ties rather than donat­ing to char­ity.

This is stan­dard re­li­gious dogma. Sec­u­lar ac­tivists rarely have the gump­tion to make it part of their pitches.

In­ter­est­ingly, how­ever, there is some so­cial stigma against donat­ing “too much”.

When you take se­ri­ously some­thing other peo­ple are hyp­o­crit­i­cal about, it makes them edgy.

most of us are sex-pos­i­tive ac­tivists, and sex-pos­i­tive ac­tivism is ar­guably an ex­tremely “low pri­or­ity” type of ac­tivism.

Not for me. Keep up the good work :D

Ad­di­tion­ally, it is un­de­ni­able that some­one has to work on the is­sues I care about, or else who would I donate money to even if I had a lot of it?

Com­par­a­tive ad­van­tage. Com­pare you be­ing an ac­tivist and your donors work­ing (which in­cludes you work­ing a low-value job to donate to your­self) and you work­ing and donat­ing to the marginal ac­tivist. Which sce­nario is su­pe­rior?

The stan­dard lawyer/​sec­re­tary ex­am­ple comes to mind- even if the lawyer types much faster, they’re bet­ter off hav­ing their sec­re­tary type for them. As an ac­tivist, are you a lawyer or a sec­re­tary? If gain­fully em­ployed, would you be a lawyer or a sec­re­tary?

• we should feel bad about spend­ing money on frivoli­ties rather than donat­ing to char­ity.

This is stan­dard re­li­gious dogma. Sec­u­lar ac­tivists rarely have the gump­tion to make it part of their pitches.

That isn’t a counter-ar­gu­ment. The idea is not wrong be­cause re­li­gious peo­ple say it, and re­quiring gump­tion also does not make an idea wrong.

A com­pletely sec­u­lar pre­sen­ta­tion of the idea can be found in The Life You Can Save by Peter Singer.

• That isn’t a counter-ar­gu­ment.

It was not in­tended as one.

• Good point re: re­li­gious dogma. I think there are stud­ies show­ing that re­li­gious/​con­ser­va­tive folks are much bet­ter at vol­un­teer­ing and donat­ing to char­ity than liberal/​sec­u­lar folks. It’s too bad.

Re: lawyer/​sec­re­tary, well, the longer I fo­cus my time on ac­tivism the more likely it be­comes that if I were more “gain­fully em­ployed” I’d be a sec­re­tary … :P

• I think the guy you’re think­ing of is Zell Krav­in­sky.

• He donated his kid­ney? They sell in Iran for $3,000 to$5,000.

I don’t know when he donated it. It could be be­fore that was le­gal.

Edit: I ac­ci­den­tally wrote “Iraq” in­stead of “Iran”.

• Kid­ney sales are le­gal in Iran, but not Iraq (they are still sold in Iraq, ob­vi­ously, but it’s a more difficult op­tion).

• A vote for the state­ment that : sex-pos­i­tive ac­tivism is (unar­guably) an ex­tremely “low pri­or­ity” type of ac­tivism.

It might be bet­ter if you can find ways to change what you feel happy about.

Just my 2p.

• Should I feel guilty for donat­ing money to pub­lic ra­dio be­cause it doesn’t save chil­dren? No.

I agree and would go even fur­ther. Guilt is a ter­rible mo­ti­va­tor and one that I would does not ap­ply to any­thing in­volv­ing char­i­ta­ble con­tri­bu­tions. Well ex­cept for, say, mug­ging the aid work­ers to steal other’s con­tri­bu­tions. In such cases guilt serves an en­tirely differ­ent and some­what use­ful role.

This is a sim­ple ques­tion of “What do you want?” If you want to re­duce malaria in­fec­tions buy nets (prob­a­bly). If you want to save a ra­dio sta­tion save a ra­dio sta­tion. If you have mul­ti­ple things you want to pri­ori­tize them and do mul­ti­pli­ca­tions or ap­prox­i­ma­tions thereof.

Never let any­one make you feel guilty for do­ing things that achieve your goals. Even your­self.

• Never let any­one make you feel guilty for do­ing things that achieve your goals.

Really? Sup­pose I want to mur­der my old pri­mary-school teacher, in a fi­nal re­venge for all that ar­ith­metic home­work. Should I not feel guilty?

• If there’s any part of that you should feel guilty about, it’s hav­ing the goal in the first place, not what you do to achieve it. Feel­ing guilty about buy­ing poi­son or sharp­en­ing a knife doesn’t make much differ­ence if you keep think­ing that the mur­der it­self is a good idea.

• Well if you get right down to it, feel­ing guilty only makes it worse. You should just not have the goal in the first place.

The point is that listen­ing to a ra­dio sta­tion should be sig­nifi­cantly be­low sav­ing lives on your list of goals.

• My point was that it is not any more wrong to spend money on pub­lic ra­dio than to spend money on ca­ble tv or a new iPod. Yes, in the­ory all my money not spent on food and shelter could go to sav­ing chil­dren, but you are not go­ing to do that, I am not go­ing to do that, and no one ei­ther of us knows is go­ing to do that.

• Well if you get right down to it, feel­ing guilty only makes it worse. You should just not have the goal in the first place.

Hence the ‘if’ at the be­gin­ning of my com­ment, though in prac­tice I do see how guilt can be use­ful at that stage: Most peo­ple don’t have com­plete con­trol over their emo­tions or what they want, and given the choice be­tween some­one want­ing to mur­der some­one, feel­ing guilty about want­ing that, and not do­ing it be­cause they feel guilty about even con­sid­er­ing it, and some­one want­ing to mur­der some­one, de­cid­ing that that’s a perfectly rea­son­able thing to want, and ac­tu­ally go­ing through with it, the former is pretty clearly prefer­able. Not want­ing to mur­der some­one at all is prefer­able to ei­ther of those, but hu­mans are pretty lousy at want­ing what we want to want.

• The first ques­tion is hard but not con­fus­ing (I’d say “yes” to the de­vel­op­ing world ex­am­ple, though); the sec­ond ques­tion con­fuses me too and I don’t have a good an­swer.

I think this whole “effi­cient char­ity” field is work­ing in the tra­di­tion of util­ity the­ory, where peo­ple’s de­sires are treated as givens and the only in­ter­est­ing ques­tion is how to max­i­mize achieve­ment of those de­sires.

In that con­text, if you de­sire get­ting nice clothes with strength X, and de­sire helping other peo­ple with strength Y, then you di­vide your re­sources ac­cord­ingly and try to max­i­mize the nice­ness of the clothes you get with X re­sources and the num­ber of peo­ple you help with Y re­sources. In that model, “try and help as many peo­ple as you can per char­ity dol­lar” is about all you can say.

This is a ter­ribly over­sim­plified model, both be­cause de­sires might be more com­pli­cated (your de­sire might not be to help peo­ple, but to help Amer­i­cans, or to help peo­ple who en­joy pub­lic ra­dio like you do), and be­cause peo­ple are not util­i­tar­ian agents and it is pos­si­ble to change the strength of your de­sires. A model that takes those into ac­count would have to, among other things, fully un­der­stand moral­ity and what it means to “want” some­thing, and I don’t fully un­der­stand ei­ther, though they’re both re­search in­ter­ests.

So this es­say is only about how to avoid one par­tic­u­larly ob­vi­ous mis­take that’s easy to model in util­ity the­ory, and not about how to avoid more im­por­tant moral and psy­cholog­i­cal mis­takes.

On the harder prob­lems, with­out hav­ing much philo­soph­i­cal foun­da­tion for do­ing so, I recom­mend Giv­ing What We Can

• Late re­sponse, but:

(a) The do­mes­tic vs. in­ter­na­tional is­sue is not clear cut—see, e.g. GiveWell re­search mes­sage board posts by Elie Hassen­feld and by Ja­son Fehr. More gen­er­ally, I think that at least at pre­sent it’s quite un­clear which philan­thropic efforts are most cost-effec­tive.

(b) In re­gards to

How­ever, I know that my lo­cal food pantry is an or­ga­ni­za­tion that feeds peo­ple who re­ally need food, that it has vir­tu­ally no over­head, and that there are chil­dren who would be mal­nour­ished with­out it. I also know that there are peo­ple all over the world who will con­tribute to malaria nets, but it is highly un­likely that any­one out­side my com­mu­nity will con­tribute to my lo­cal food pantry.

see Holden Karnofsky’s post Hunger Here vs. Hunger There.

(c) In re­gards to:

My pur­pose in donat­ing money to pub­lic ra­dio is to keep my fa­vorite shows on the air, and my dona­tions do that very effi­ciently. Yes, the money could go to save chil­dren, but so could the money I use to pay my ca­ble bill. I should per­haps not con­sider it as char­ity the way I do a dona­tion that saves chil­dren, but I should not feel guilty.

You might be in­ter­ested by kom­pon­isto’s com­ments to a post that I made which are in similar spirit.

See also Holden Karnofsky’s Noth­ing wrong with self­ish giv­ing—just don’t call it philan­thropy and the com­ments to it.

• There is no sim­ple met­ric for “most good done.”

I sug­gest the QALY, or qual­ity-ad­justed life year.

Should I feel guilty for donat­ing money to pub­lic ra­dio be­cause it doesn’t save chil­dren?

Yes. Sure you want ra­dio, but they don’t want to die. Who says your wants are more im­por­tant?

Could you jus­tify kil­ling peo­ple for en­ter­tain­ment? Is this any differ­ent?

Ba­si­cally, I am leery of let­ting peo­ple choose their own fac­tors when given a range of 1 be­ing perfect life and 0 be­ing death. For in­stance, a char­ity that cures blind­ness in im­pov­er­ished sec­tions of Africa, with a pro-this-char­ity treat­ment might choose 0.1 as blind, 0.9 as cured (blind­ness is hugely dis­ad­van­ta­geous, giv­ing back sight is there­fore a huge im­prove­ment); an anti-this-char­ity treat­ment might choose 0.1 as blind and 0.3 as cured (the rest of their life still sucks). This means a QALY-based look at the char­ity could over- or un­der-es­ti­mate by as much as a fac­tor of 4! Com­par­i­sons of char­i­ties based on QALYs that are gamed could, pos­si­bly, be only vi­able on or­der-of-mag­ni­tudes.

• Is there a stan­dard for DALYs?

I’m told that there’s some kind of differ­ence, but I still think of them as the same unit.

• Yes. Sure you want ra­dio, but they don’t want to die. Who says your wants are more im­por­tant?

The fact that it’s her own money?

• Sure you want ra­dio, but they don’t want to die. Who says your wants are more im­por­tant?

My au­ton­omy.

• It’s a use­ful ex­er­cise for as­piring economists and ra­tio­nal­ists to dis­sect char­ity into sep­a­rate com­po­nents of warm fuzzies vs. effi­ciency. How­ever, maybe it’s best for the gen­eral pop­u­la­tion not to be fully con­scious that these are sep­a­rate com­po­nents, since the spirit of giv­ing is like a frog: you can dis­sect it, but it dies in the pro­cess (adap­ta­tion of an E.B. White quote).

Lemma: we want char­ity to be en­joy­able, so that more peo­ple are mo­ti­vated to do it. (Anal­ogy: cap­i­tal­ist coun­tries let rich peo­ple keep their riches, to cre­ate an in­cen­tive for eco­nomic growth, even though it might cre­ate more util­ity in the short term to tax rich peo­ple very highly.)

Con­sider this quote from the ar­ti­cle:

If he went to the beach be­cause he wanted the sun­light and the fresh air and the warm feel­ing of per­son­ally con­tribut­ing to some­thing, that’s fine. If he ac­tu­ally wanted to help peo­ple by beau­tify­ing the beach, he’s cho­sen an ob­jec­tively wrong way to go about it.

Sure, but mak­ing the lawyer con­scious of this will give him a com­plete buz­zkill. He will re­al­ize that he was un­con­sciously do­ing the act for self­ish (and kind of silly) rea­sons. Your hope in tel­ling him this is that he will in­stead opt to use his $1000 salary to hire peo­ple, but I ques­tion whether he would ac­tu­ally fol­low through with that kind of giv­ing in the long run, since his un­con­scious origi­nal mo­tive was warm fuzzies, not effi­ciency. In effect, you may have pre­vented him from do­ing any­thing char­i­ta­ble at all. Don’t let the perfect be the en­emy of the good. So, this ar­ti­cle is great fod­der for some­one trained in ra­tio­nal­ist/​eco­nomic thought, but keep in mind that this type of think­ing makes many peo­ple un­easy. • This is a gen­uine prob­lem you’re pre­sent­ing, and I think it re­quires a third solu­tion be­sides the pre­sented op­tions of “Let the lawyer do what he wants” and “Give the lawyer a buz­zkill”. What we need to do is find a way of get­ting the lawyer to un­der­stand what the right thing to do is, with­out mak­ing them feel defen­sive or like a jerk. If we make the bul­let tasty enough, it’ll get eas­ier to swal­low. Ra­tion­al­ist mar­ket­ing FTU (For The Utilons). • So, is my goal in ex­plain­ing this stuff to some­one to max­i­mize effi­ciency at achiev­ing their goals (warm fuzzies), or to max­i­mize effi­ciency at achiev­ing my goals (char­ity)? (Or maybe I want warm fuzzies and the lawyer wants char­ity, what­ever.) • The lawyer wants both warm fuzzies and char­itrons, but has con­flated the two, and will prob­a­bly get buz­zkil­led (and lose out on both mea­sures) if the dis­tinc­tion is made clear. The best out­come is one where the lawyer gets to max­i­mize both, and that hap­pens at the end of a long road that be­gins with in­tro­spec­tion about what warm fuzzies ought to mean. • If you learn about how to give right, some of the warm fuzzies will go away, and fewer peo­ple will donate, but the peo­ple who do donate will donate bet­ter. If all you’re go­ing to be do­ing is pick­ing up lit­ter at a beach, it re­ally doesn’t mat­ter if you stop when you find out it’s not helping peo­ple. You can find an­other hobby. • Not quite the same sce­nario, but close: of­ten when I’m con­sid­er­ing donat­ing to some char­ity, there’s a re­minder in the back of my head that if I were to truly sup­port this char­ity I would donate a much larger amount. This isn’t a happy thought, it gen­er­ates con­flict: there’s an­other part of me that doesn’t like spend­ing large amounts of money. Thus, I of­ten donate noth­ing at all. I’m still work­ing on this con­flict. • Most of us al­lo­cate a par­tic­u­lar per­centage to char­ity, de­spite the fact that most peo­ple would say that nearly noth­ing we spend money on is as im­por­tant as sav­ing chil­drens lives. I don’t know whether you think it’s that we over­es­ti­mate how much we value sav­ing chil­drens lives, or un­der­es­ti­mate how im­por­tant xbox games, so­cial events, large tvs and eat­ing tasty food are to us. Or per­haps you think it’s none of that, and that we’re be­ing sim­ply ir­ra­tional. I doubt that any­one could con­sis­tently live as if the differ­ence be­tween choice of rent­ing a nice flat and rent­ing a dive was one life per month, or that halv­ing nor­mal gro­cery con­sump­tion for a month was a childs life that month, etc. If that’s re­ally the aim, we’re go­ing to have to do a sig­nifi­cant amount of emo­tional en­g­ineer­ing. I also want to stick up for the ne­ces­sity of analysing the way that a char­ity works, not just what they do. For ex­am­ple, char­i­ties that em­ploy lo­cal peo­ple and lo­cal equip­ment may save fewer peo­ple per dol­lar in the short term, but may be less likely to cre­ate a cul­ture of de­pen­dence, and may be more sus­tain­able in the long term. Th­ese con­sid­er­a­tions are im­por­tant too. • I doubt that any­one could con­sis­tently live as if the differ­ence be­tween choice of rent­ing a nice flat and rent­ing a dive was one life per month, or that halv­ing nor­mal gro­cery con­sump­tion for a month was a childs life that month, etc. Then don’t. Use two-level util­i­tar­i­anism. Live like this for a brief amount of time, dur­ing which you de­cide how to live the rest of the time. Work out how much you can cut your bud­get be­fore you start earn­ing less or the risk of giv­ing up goes to high, and live like that for a pe­riod be­fore you go back and check to see if it’s work­ing well again. • Although I definitely agree with the thrust of the ar­ti­cle, I don’t feel that lives-saved is nec­es­sar­ily a very good met­ric of util­ity. A child in the Third World might be saved from malaria, but grow up nu­tri­ent defi­cient lead­ing to re­duced men­tal ca­pac­ity, work on a sub­sis­tence farm, con­tract HIV, and die af­ter hav­ing three kids, who sub­se­quently starve. A char­ity that pre­vented fewer deaths in a pre­dictable causal se­quence might still be a bet­ter util­ity max­i­mizer if it had a greater pos­i­tive effect on peo­ple’s qual­ity of life. Of course, a lot of us already agree on the best available util­ity max­i­miz­ing char­ity, but even among the more “mun­dane” op­tions I think that causes such as pro­mot­ing ed­u­ca­tion in the third world may beat out di­rect life-sav­ing max­i­miz­ers. • I agree with Des­r­topa in that “I don’t feel that lives-saved is nec­es­sar­ily a very good met­ric of util­ity.” Death is bi­nary (dead /​ not dead), but hu­man pain and suffer­ing is not. This should im­pact the anal­y­sis. As­sum­ing the same cost to save the life, if forced to de­cide be­tween sav­ing some­one from a fatal gun­shot wound (per­haps in a war or en­camp­ment some­where) ver­sus sav­ing some­one from pan­cre­atic can­cer (ac­cord­ing to Live­strong, one of the most painful ter­mi­nal dis­eases), the out­come (life-sav­ing) may be the same in ei­ther case, but there is more util­ity in sav­ing the lat­ter be­cause over­all pain would be re­duced. Thanks for this ar­ti­cle; it’s a fan­tas­tic read. • Per­haps a bet­ter idea would be to spend money on ed­u­ca­tion of women in poor ar­eas, some­thing that is known to re­duce the fer­til­ity rate. By re­duc­ing the fer­til­ity rate we also re­duce the num­ber of poor, starv­ing, dy­ing in HIV etc chil­dren born into this world. I think that sim­ply mea­sur­ing the num­ber of dead chil­dren may be use­ful as a sim­plifi­ca­tion, but it’s too sim­plis­tic. Really, to me it seems like it’s just some­thing that peo­ple be­liev­ing in ax­io­matic morals are hav­ing prob­lems deal­ing with. “But, think of the chil­dren!” If the an­swer to “is it bet­ter to spend this money on sav­ing a kids life?” is always yes, I’d say you have a prob­lem with your value sys­tem. • I have a ques­tion. This ar­ti­cle sug­gests that for a given util­ity func­tion there is one sin­gle char­ity that is best and that’s the one one should give money to. That looks a bit prob­le­matic to me—for ex­am­ple, if ev­ery­one in­vests in malaria nets be­cause that’s the sin­gle one that saves most lives, then no­body is in­vest­ing in any other kind of char­ity, but shouldn’t those things get done too ? We can get around this by con­sid­er­ing that the effi­ciency func­tion varies with time—for ex­am­ple, once ev­ery­body gives their money to buy nets the marginal cost of each saved life in­creases, un­til some other char­ity be­comes best and all char­i­ta­ble giv­ing switches to that one. But we don’t have a com­plete and up-to-the-sec­ond knowl­edge of how many lives each marginal dol­lar will save in ev­ery char­ity, all we have to work with is ap­prox­i­ma­tions. In that situ­a­tion, wouldn’t it be best to have a bas­ket of char­i­ties one gives to, with more money go­ing to those that save the most lives but not putting all the money on a sin­gle char­ity ? Or is this con­sid­er­a­tion com­pletely and ut­terly pointless in a world where most peo­ple do NOT act like this, and most peo­ple don’t give enough money to change the game, so ra­tio­nal ac­tors who don’t have mil­lions of dol­lars to give to char­ity should always give to the one that saves the most lives per dol­lar any­way ? • What hap­pens in that situ­a­tion is that peo­ple con­tinue to in­vest in malaria nets, so much that the marginal cost of sav­ing an­other life goes from say,$500 to $700, and for$600 dol­lars you can dig a well, sav­ing an­other per­sons life. In essence, you donate to the most effi­cient char­ity un­til that money has caused the char­ity to have to pay more to save lives, and there­fore stops be­ing the most effi­cient char­ity.

The prob­lem is that ev­ery­one is act­ing in­de­pen­dent and with limited knowl­edge. It’s hard to know what other peo­ple are choos­ing. There may also be long de­lays be­tween you and oth­ers pay­ing and the cost chang­ing.

Say that the op­ti­mal out­come is that out of $1000M,$200M is spent on in­sect nets and $800M on wells, and that you can only donate to one char­ity (too both­er­some or high trans­ac­tion costs or some­thing). Now, if ev­ery­one is ra­tio­nal they are go­ing to donate to the wells, and no one to nets. This is a sub­op­ti­mal out­come. It’d also be difficult to co­or­di­nate the mil­lions of peo­ple donat­ing, so that just the right amount choose nets in­stead of wells. A solu­tion to such co­or­di­na­tion is to roll a dice. If ev­ery­one makes a ran­dom se­lec­tion and lets the prob­a­bil­ity of choos­ing nets be 2/​10ths then the ex­pected out­come is just what we want. Now, you can ad­just this to how many (you think) are play­ing like your­selves. E.g. if you know most peo­ple are go­ing to give to wells, per­haps it’d be bet­ter if you put higher prob­a­bil­ity on nets (per­haps 100%). • The thing about that is, is that not ev­ery­one is donat­ing at the same time, so that they can see the ex­pected value change. • Yes, but there can be long de­lays be­tween a dona­tion hap­pen­ing and up­dates. Co­or­di­nat­ing dona­tions can be non-triv­ial, es­pe­cially when flash crowds ap­pear (e.g. sob story on red­dit). Also, such a ran­dom­ized ap­proach is not nec­es­sary if one can just donate small amounts to mul­ti­ple pro­jects in­stead (i.e. if trans­ac­tion fees are not a prob­lem). • I once donated some money to VillageReach a few min­utes be­fore get­ting the GiveWell newslet­ter is­sue an­nounc­ing that VillageReach wasn’t go­ing to be among the top char­i­ties in the next up­date be­cause their found­ing gap had mostly closed and en­courag­ing peo­ple to wait for the next up­date be­fore de­cid­ing whom to donate money to. True story! • “The lawyer who quits a high-pow­ered law firm to work at a non­profit or­ga­ni­za­tion cer­tainly seems like a good per­son. But if we define “good” as helping peo­ple, then the lawyer who stays at his law firm but donates the profit to char­ity is tak­ing Cato’s path of max­i­miz­ing how much good he does, rather than how good he looks.” Wouldn’t that de­pend on how much harm the lawyer might do by re­main­ing at the high-pow­ered law firm? What if the law firm spe­cial­izes in so­cially-harm­ful ac­tivi­ties, like defend­ing cor­po­rate malfea­sance or (pick your ex­am­ple). How does that fit into the equa­tion? In other words, I don’t think it’s that sim­ple, al­though it’s an ex­cel­lent place to start, and I will cer­tainly check out GiveWell for our next char­i­ta­ble tithing ses­sion. • “Wouldn’t that de­pend on how much harm the lawyer might do by re­main­ing at the high-pow­ered law firm?” Yes. This log­i­cal er­ror is pre­sent in all the char­ity re­lated ar­ti­cles. By the time you own the money it is too late. You have helped a cor­rupt sys­tem to be­come even more cor­rupt. No amount of money donated to char­ity—not even a mul­ti­ple of the money you earned—can right that wrong. Not even with max­i­mally effec­tive char­ity. If the lead­ers are wrong and can’t be de­posed it doesn’t help to try and save their vic­tims, be­cause as soon as you help them, what­ever they gain is taken from them and used to strengthen the rule of their op­pres­sors. The vic­tims can save them­selves if they are ready to serve the op­pres­sors, for ex­am­ple by listen­ing to yvain’s and other’s feel good char­ity, and be­come high pow­ered lawyers, but in that way they only spiral the sys­tem into more and more cor­rup­tion. The fact that they try and make the ones not co­op­er­at­ing with cor­rup­tion look like they are not helping doesn’t make it OK to then make mat­ters worse and claim you are do­ing good. The new speak runs sev­eral lev­els deep here. • Yes. This log­i­cal er­ror is pre­sent in all the char­ity re­lated ar­ti­cles. By the time you own the money it is too late. You have helped a cor­rupt sys­tem to be­come even more corrupt I might be miss­ing some­thing, but this (and the rest of your post) reads ba­si­cally like Marx­ist pro­pa­ganda. Are you se­ri­ously sug­gest­ing that any­one who makes a lot of money has done so through “cor­rup­tion”? I would hope LW was one of the places on the in­ter­net that this sort of “tru­ism” could be avoided. Just about the only way to make a lot of money is to do some­thing that other peo­ple want do­ing, and which you do bet­ter than av­er­age. The fact that they try and make the ones not co­op­er­at­ing with cor­rup­tion look like they are not helping doesn’t make it OK to then make mat­ters worse and claim you are do­ing good I’m se­ri­ously strug­gling to parse this sen­tence, but it seems to be es­sen­tially say­ing that you’re go­ing to stick with your gut in­stinct that work­ing for a high-pow­ered law firm can’t pos­si­bly be as good as work­ing for a nice fluffy non-profit, and damn the num­bers. • Just about the only way to make a lot of money is to do some­thing that other peo­ple want do­ing, and which you do bet­ter than av­er­age. This re­ally isn’t true. But if you said “the ways to make money we’re talk­ing about” then we’d be fine. The most morally ob­jec­tion­able job I’ve seen sug­gested on LW is work­ing in fi­nance, and the worst you can do there is be a con man (though on a pretty mas­sive scale). • Caveat Cer­tain fi­nan­cial ac­tions may in­crease the like­li­hood or mag­ni­tude of a fi­nan­cial crisis. Fi­nan­cial crises are bad on the scale of the fi­nance sec­tor and so this is sig­nifi­cant on the mar­gin. • “I might be miss­ing some­thing, but this (and the rest of your post) reads ba­si­cally like Marx­ist pro­pa­ganda.” Thank you. Marx was a very in­tel­li­gent per­son who un­rav­eled much of the in­ner work­ings of cap­i­tal­ism. His er­ror was—I think—that there is some­thing like a col­lec­tive will of the peo­ple (a CEV, maybe) and that there is an effec­tive way of mea­sur­ing and im­ple­ment­ing it. We all know how badly it turned out. But maybe the idea of har­ness­ing col­lec­tive greed is even worse be­cause it seems flawed already from the be­gin­ning. “Just about the only way to make a lot of money is to do some­thing that other peo­ple want do­ing, and which you do bet­ter than av­er­age.” This is quite wrong. There is also a very big effort to pre­vent other peo­ple from ac­quiring things, and I don’t just mean WMD. Maybe you could read up on the con­cept of ar­tifi­cial scarcity. ″… you’re go­ing to stick with your gut in­stinct that work­ing for a high-pow­ered law firm can’t pos­si­bly be as good as work­ing for a nice fluffy non-profit, and damn the num­bers” My gut in­stinct tells me a lawyer who in his day job se­cures a quar­ter of a billion dol­lar set­tle­ment with an evil regime to pre­vent le­gal per­se­cu­tion of an evil poli­ti­cian in­volved with a ma­jor weapons man­u­fac­turer, can­not offset this with buy­ing a few mosquito nets for chil­dren in that same third world coun­try. • My gut in­stinct tells me a lawyer who in his day job se­cures a quar­ter of a billion dol­lar set­tle­ment with an evil regime to pre­vent le­gal per­se­cu­tion of an evil poli­ti­cian in­volved with a ma­jor weapons man­u­fac­turer, can­not offset this with buy­ing a few mosquito nets for chil­dren in that same third world coun­try. What does you gut in­stinct say about a lawyer who is paid an av­er­age of a quar­ter of a mil­lion dol­lars for a year’s work in which e puts an av­er­age of 2 in­no­cent peo­ple in jail for six years each and donates on av­er­age 2% of er in­come (5000 dol­lars) to buy­ing mosquito nets for third world chil­dren, sav­ing on av­er­age 10 lives? You seem to be con­sid­er­ing the ab­solute worst case sce­nario, and adding in ex­tra­ne­ous con­sid­er­a­tions to un­fairly sway the ar­gu­ment to your side. • What does you gut in­stinct say about a lawyer who is paid an av­er­age of a quar­ter of a mil­lion dol­lars for a year’s work in which e puts an av­er­age of 2 in­no­cent peo­ple in jail for six years each and donates on av­er­age 2% of er in­come (5000 dol­lars) to buy­ing mosquito nets for third world chil­dren, sav­ing on av­er­age 10 lives? My gut in­stinct says “2*6 years of OECD cit­i­zen free­dom > 10*45 years of sub-Sa­haran Afri­can lifes­pan”. Not sure where my point of in­differ­ence is. Note that I’m mak­ing a very char­i­ta­ble as­sump­tion as to how much a mosquito net ex­tends lifes­pans; typ­i­cally in­ter­ven­tions like this just keep you al­ive long enough to hit your next emer­gency. • This is the kind of re­sponse I want: one that doesn’t say “damn the num­bers”. • This is the kind of re­sponse I want: I’m sorry, this is am­bigu­ous. Does “this” re­fer to the sec­ond half of your post, my post, or both? Step­ping through some of the math, in case oth­ers are in­ter­ested: As­sume the peo­ple in ques­tion earn the me­dian Amer­i­can salary- 6 years of not work­ing is 2*6*$32k=$384k, and add on the cost of im­pris­on­ing them: 2*6*$22k=$264k. So the lawyer is do­ing dam­age to the tune of$648,000, and in re­turn is putting $5,000 (that’s .77%) to use sav­ing peo­ple. Let’s as­sume they’re earn­ing, say, the Libe­rian per cap­ita GDP (which is gen­er­ally higher than me­dian in­come) of$424, and again make the char­i­ta­ble as­sump­tion that the $5000 con­verts into 450 years of lifes­pan. We’ve added$190,800 by keep­ing them al­ive.

Net dol­lar loss: $457k. So this lawyer’s par­ti­ci­pa­tion in the sys­tem is eat­ing half a mil­lion dol­lars per year; is that worth “ex­tended lives in Libe­ria”—“im­pris­oned years in Amer­ica”? I strongly sus­pect not. • Your post! I am very pleased that some­one ex­am­ined their gut re­ac­tion to my sce­nario with num­bers. :) • Then I’m glad I asked, be­cause I read that the other way first and my first draft re­flected that :P • “What does you gut in­stinct say about a lawyer who is paid an av­er­age of a quar­ter of a mil­lion dol­lars for a year’s work in which e puts an av­er­age of 2 in­no­cent peo­ple in jail for six years each and donates on av­er­age 2% of er in­come (5000 dol­lars) to buy­ing mosquito nets for third world chil­dren, sav­ing on av­er­age 10 lives?” You are for­get­ting the lawyer works to up­hold the sta­tus quo. If that sta­tus quo (“the sys­tem”) also makes arms deal­ers reach and in­vades third world coun­tries or does other de­spi­ca­ble things, the net effect of the lawyers ac­tions can still be nega­tive. Think about it as differ­ent math op­er­a­tors. You can name a num­ber as big as you want but if I can change the sign of it, it will always be smaller than my small pos­i­tive num­ber. “You seem to be con­sid­er­ing the ab­solute worst case sce­nario, and adding in ex­tra­ne­ous con­sid­er­a­tions to un­fairly sway the ar­gu­ment to your side.” I was thrown a bit off bal­ance by the pe­jo­ra­tive “marx­ist pro­pa­ganda”. I don’t want to post too much poli­tics in this thread so this will be my last con­tri­bu­tion. • You are for­get­ting the lawyer works to up­hold the sta­tus quo. If that sta­tus quo (“the sys­tem”) also makes arms deal­ers reach and in­vades third world coun­tries or does other de­spi­ca­ble things, the net effect of the lawyers ac­tions can still be nega­tive. The amount that a crim­i­nal trial lawyer con­tributes to the sta­tus quo as it re­lates to the suffer­ing of third world coun­tries is neg­ligible; it is fair to ig­nore such small con­stants. In­ci­den­tally, I down­voted ben­tarm’s com­ment about Marx­ism and up­voted your re­ply (“Thank you. Marx was a very in­tel­li­gent per­son”) be­cause com­par­ing any­thing to Marx­ism doesn’t strike me as en­gag­ing with the idea (rather, match­ing to an already-re­jected idea so they can re­ject this new idea eas­ily), and your re­sponse in­di­cated your con­cern was sep­a­rate to Marx’s. • You are for­get­ting the lawyer works to up­hold the sta­tus quo. If that sta­tus quo (“the sys­tem”) also makes arms deal­ers reach and in­vades third world coun­tries or does other de­spi­ca­ble things, the net effect of the lawyers ac­tions can still be nega­tive. Think about it as differ­ent math op­er­a­tors. You can name a num­ber as big as you want but if I can change the sign of it, it will always be smaller than my small pos­i­tive num­ber. A cu­ri­ous note: If ‘The Sys­tem’ or any­one with power within it ac­tu­ally took you se­ri­ously they would nec­es­sar­ily im­prison or oth­er­wise crip­ple your abil­ity to take ac­tion. You have es­sen­tially de­clared an in­tent to fight against ev­ery­thing the sys­tem does. That is, if you had any sig­nifi­cance you would be a clear and pre­sent threat to na­tional (and even in­ter­na­tional) se­cu­rity. • My gut in­stinct tells me a lawyer who in his day job se­cures a quar­ter of a billion dol­lar set­tle­ment with an evil regime to pre­vent le­gal per­se­cu­tion of an evil poli­ti­cian in­volved with a ma­jor weapons man­u­fac­turer, can­not offset this with buy­ing a few mosquito nets for chil­dren in that same third world coun­try. Your gut in­stinct needs to learn more eco­nomics. ;) Also note: This board uses mark­down syn­tax. De­tails are in the ‘help’ link just be­low the com­ment box. To quote a para­graph be­gin the line with an >. • I think this might be cor­rect but that hu­mans are prone to pri­ori­tis­ing the welfare of kin and close friends, and so some­one work­ing di­rectly with peo­ple and form­ing some kind of re­la­tion­ship with them may be more likely to donate fi­nan­cial re­sources to that group in fu­ture. The lawyer may be more will­ing to spend money to keep a beach safe and free of lit­ter if he or she has some per­sonal ex­pe­rience which in­creases the im­por­tance of that beach in his mind. Most of us don’t give much weight to mosquito nets be­cause our own ex­pe­rience doesn’t even put that on the radar. I make a point of buy­ing only FairTrade choco­late. My men­tal hack for times when I feel tempted to buy the or­di­nary kind is to think about peo­ple I love and ad­mire, and imag­ine that my spend­ing de­ci­sion ex­tends as far as hav­ing an im­me­di­ate im­pact on whether they are paid fairly for their work. This is not, di­rectly, how the mar­ket works, but as a re-fram­ing ex­er­cise it does help me in stick­ing to a re­s­olu­tion when my own de­sires seem more com­pel­ling than those of the peo­ple who pro­duce the choco­late. I also ques­tion whether money is always more di­rectly effec­tive than time. I think the hu­man re­la­tion­ships which might draw peo­ple to fur­ther fi­nan­cial sup­port are of­ten in and of them­selves benefi­cial. That has cer­tainly been my ex­pe­rience in for­mal and in­for­mal men­tor­ing situ­a­tions. No amount of money can buy lov­ingkind­ness, and while a kind word will not fill an empty stom­ach, some­one with their im­me­di­ate food, shelter and med­i­cal needs met may still be very much in need of that kind word. En­courage­ment and gen­uine care should not be over­looked as fac­tors in in­creas­ing some­one’s qual­ity and du­ra­tion of life. Th­ese things are hard to quan­tify but, for me, they tip the bal­ance to­ward con­tribut­ing time and en­ergy di­rectly to lo­cal causes, es­pe­cially as I earn very lit­tle money any­way. • Wel­come to LessWrong! I’d like to men­tion that here on LessWrong we will try to quan­tify the value of lov­ing kind­ness and en­courage­ment, and af­ter quan­tify­ing we’re go­ing to find that it would fall well be­low the value of im­me­di­ate food, shelter, and med­i­cal needs. es­pe­cially as I earn very lit­tle money any­way. I sus­pect this is a stronger rea­son than the pre­ced­ing para­graph ;) • here on LessWrong we will try to quan­tify the value of lov­ing kind­ness and en­courage­ment, and af­ter quan­tify­ing we’re go­ing to find that it would fall well be­low the value of im­me­di­ate food, shelter, and med­i­cal needs. It helps in this re­gard to be re­ally sure of the se­cu­rity of one’s own im­me­di­ate food, shelter and med­i­cal needs. (Can this be claimed of all LessWrong par­ti­ci­pants? If so, then LW’s par­ti­ci­pant base is not wide enough.) • Yes. This is my ma­jor dis­agree­ment with the “give un­til it hurts” slo­gans you some­times see. Also, I guess yes to your par­en­thet­i­cal. This is a se­lec­tion effect caused by LessWrong’s medium (gen­er­ally, shelter is a nec­es­sary con­di­tion for in­ter­net ac­cess, and food and med­i­cal needs are prob­a­bly—hope­fully? - pri­ori­tized over in­ter­net ac­cess). • gen­er­ally, shelter is a nec­es­sary con­di­tion for in­ter­net ac­cess, and food and med­i­cal needs are prob­a­bly—hope­fully? - pri­ori­tized over in­ter­net access Ac­tu­ally, no, it turns out your view of the world is in­cor­rect and in need of up­dat­ing. I spent a chunk of 2002 couch-sur­fing, liv­ing on the kind­ness of friends, look­ing for work in Lon­don. I se­ri­ously put rather a high value on In­ter­net, be­cause it was the ra­tio­nal choice in se­cur­ing a job. “Well, yes, it’s a house … but there’s no net there.” It’s that im­por­tant. • Wow. I definitely do not treat the in­ter­net as that im­por­tant. Clearly I gen­er­al­ised from my own ex­am­ple in­stead of seek­ing out any data. I can even see how it makes ra­tio­nal sense to pre­fer in­ter­net over shelter, food, and med­i­cal needs; it’s an in­stru­ment to achieve all three ter­mi­nal goals. I just didn’t think that way. Man, that one-mind fal­lacy is in­sidious. • In the situ­a­tion, it would have been ir­ra­tional—blither­ingly stupid—not to make damn sure I had in­ter­net ac­cess in the prospec­tive new place. Med­i­cal needs are fine in the UK (here’s to the NHS!), cheap food ex­ists in small quan­tities, shelter is the crip­pling ex­pense in Lon­don. For­tu­nately my friends are sysad­mins. I would char­ac­ter­ise my situ­a­tion at the time as closer to “dis­tressed gen­tle­man” than “bum”. (1) In any case, I owe the world (and said in­di­vi­d­u­als) lots of kind­ness points, and am quite proud to pay a siz­able chunk of my in­come in tax, be­cause I know per­son­ally what it pays for … More broadly: yes, you ac­tu­ally need In­ter­net to par­ti­ci­pate in Western civil so­ciety these days. Restrict­ing it from the home­less is a way to keep them there. They have phones too these days, and not just as some sort of frip­pery—why do they need them? And also, lov­ing kind­ness and en­courage­ment are how to treat hu­mans; posit­ing that as some­how di­choto­mous with food, shelter and med­i­cal care is a twist of thought I find con­fus­ing. 1. And hadn’t been the former long enough for it to smell like the lat­ter. • What was that t-shirt (from slightly ear­lier than 2002) ‘bout drugs, sex and ’net ac­cess? • One of the sev­eral alt.gothic T-shirts, dat­ing to the mid-1990s. (I had sev­eral but ap­pear to have only the origi­nal 1994 one left.) • The se­cu­rity of one’s own ac­cess to phys­i­cal ne­ces­si­ties is an in­ter­est­ing fac­tor in this. Are those whose se­cu­rity has been un­sta­ble more or less likely to donate time or money to char­ity? For me per­son­ally, un­cer­tainty about my own cir­cum­stances is a dou­ble-edged sword. If I am feel­ing a bit skint I’m un­likely to give money to some­one beg­ging on the street, and if I know my bud­get will be limited I am stingier than usual about char­ity boxes in shops. At the same time, an aware­ness that it is only be­cause of the kind­ness of oth­ers that I am not home­less my­self makes me ea­ger to pass that kind­ness on in un­struc­tured ways (be­ing kind to oth­ers where I can in the course of my work and leisure) and more for­mally (this win­ter, vol­un­teer­ing at a lo­cal night shelter). • Pos­si­bly the peo­ple who give the most, albeit to rel­a­tives, are im­mi­grants from less de­vel­oped to more de­vel­oped coun­tries. Even though for many it means low­er­ing their stan­dards of liv­ing in the US (or wher­ever), they know the re­mit­tance they send is send­ing their younger sister to school, buy­ing a new roof for the fam­ily house in Bo­livia, etc. In the US, the low­est in­come bracket gives a larger per­cent of their in­come than any other bracket. I haven’t seen num­bers on whether this in­cludes peo­ple on the brink of not hav­ing their ba­sic needs met, but I bet a lot of them have been there at some point. • In the US, the low­est in­come bracket gives a larger per­cent of their in­come than any other bracket. I haven’t seen num­bers on whether this in­cludes peo­ple on the brink of not hav­ing their ba­sic needs met, but I bet a lot of them have been there at some point. Note that it’s pos­si­ble that a sub­stan­tial frac­tion of these dona­tions are made to com­mu­nity or­ga­ni­za­tions (churches, etc.) and so may effec­tively serve as mem­ber­ship dues. De­spite this I think that this statis­tic makes a good re­join­der to mid­dle/​up­per class peo­ple who claim that they can’t af­ford to give. • On the other hand, per­haps the poor give too much! They should be re­ceiv­ing the aid, not giv­ing it out! Con­sider all the eco­nomic op­por­tu­ni­ties that poor im­mi­grants are giv­ing up by re­mit­ting so much of their in­come to rel­a­tives where they came from. Per­haps it would be bet­ter if they saved and in­vested in­stead, and then af­ter se­cur­ing them­selves fi­nan­cially, then start giv­ing back? • Per­haps it would be bet­ter if they saved and in­vested instead If you con­sider your­self as, say, a Mex­i­can 30-year-old who comes to the US and works as a car­pen­ter, would you pre­fer to save your earn­ings and in­vest them (de­spite hav­ing lit­tle for­mal ed­u­ca­tion, and thus be­ing un­likely to in­vest well) while your wife, son, and par­ents con­tinue liv­ing in a shack in Chi­a­pas? Know­ing that they would de­spise you for hoard­ing your earn­ings while they scraped by? I bet you would send them part of your pay­check. The op­por­tu­nity cost of sav­ing that money is too high. • Thanks for the wel­come. I won­der how it’s pos­si­ble to quan­tify en­courage­ment and the value of re­la­tion­ships. I have been on the re­ceiv­ing end of a good deal of care and en­courage­ment at a time when my phys­i­cal health was poor and noth­ing could im­me­di­ately be done to im­prove it. This gave me great hope and is ex­pe­rience I still draw courage from when I find life challeng­ing. I don’t have a spare me to ex­per­i­ment on so can only imag­ine how I might have fared with­out that sup­port, but I know it has seemed more in­fluen­tial than the prac­ti­cal sup­port I had, and in some cases I would not have sought prac­ti­cal sup­port had I not had steady emo­tional en­courage­ment. I am for­tu­nate in that I have never been with­out suffi­cient food or ad­e­quate shelter, but that would not have been the case had I been left to my own de­vices. I can only ex­pe­rience the world as my­self, but for me, lov­ing kind­ness and un­con­di­tional pos­i­tive re­gard have been ex­tremely im­por­tant, and are prob­a­bly the de­cid­ing fac­tor in my sub­se­quent at­tempts to help oth­ers. On a wider scale, I’ve of­ten won­dered why we don’t sim­ply set up a tax sys­tem such that ev­ery­one can have a de­cent phys­i­cal stan­dard of liv­ing. Pop­u­la­tion con­cerns aside (given the lower birth rate that ap­pears to re­sult from in­creases in stan­dard of liv­ing this should sort it­srlf out) I think some of this comes back to our ten­dency to pri­ori­tise kin­ship or clan groups over the com­mon good. I would ar­gue that not hav­ing a di­rect re­la­tion­ship with the peo­ple we are try­ing to help makes us more likely to with­draw aid at the first hint of dan­ger. Cer­tainly those with­draw­ing benefits or fi­nan­cial aid from the most dis­ad­van­taged in Bri­tain right now are not those who work with the dis­abled and the home­less on an on­go­ing ba­sis. Yes, good peo­ple ought to donate to char­ity, and funds should be used effi­ciently, but the idea that pay­ing taxes, vot­ing, donat­ing a bit to char­ity and per­haps writ­ing to an MP or go­ing on a protest is enough seems flawed. I think that for the changes to oc­cur which would guaran­tee ev­ery­one a de­cent stan­dard of liv­ing, peo­ple need se­ri­ous mo­ti­va­tion. I see that mo­ti­va­tion com­ing from per­sonal in­volve­ment and re­la­tion­ships more than from a cost/​benefit anal­y­sis of how to spend the “char­ity” por­tion of a house­hold bud­get. The lat­ter is im­por­tant and I am glad there are or­gani­sa­tions like GiveWell which at­tempt some of the ar­ith­metic, but I ques­tion whether money-only donors will, in gen­eral, eval­u­ate the rest of their spend­ing and ac­tivity with a view to in­creas­ing the com­mon good, and I sus­pect that the ab­stract con­nec­tions formed by fi­nan­cial dona­tions are frail, mak­ing such aid more likely to be with­drawn if it is in­con­ve­nient. I don’t sug­gest that peo­ple who donate money to char­ity should dis­con­tinue that sup­port but I do think it helpful if they also spend some time, per­haps as lit­tle as an hour per week or month, do­ing some kind of aid work that offers the op­por­tu­nity for a gen­uine re­la­tion­ship not based on who has more money. As most peo­ple do not spend all their wak­ing hours work­ing, this need not de­tract from their fi­nan­cial con­tri­bu­tions. I would be in­ter­ested in see­ing any data that sup­port or re­fute this; I am ex­trap­o­lat­ing from my own ob­ser­va­tions. • I don’t sug­gest that peo­ple who donate money to char­ity should dis­con­tinue that sup­port but I do think it helpful if they also spend some time, per­haps as lit­tle as an hour per week or month, do­ing some kind of aid work that offers the op­por­tu­nity for a gen­uine re­la­tion­ship not based on who has more money. As most peo­ple do not spend all their wak­ing hours work­ing, this need not de­tract from their fi­nan­cial con­tri­bu­tions. This is an im­por­tant point: perfectly spher­i­cal ra­tio­nal­ists of uniform den­sity in a vac­uum at ab­solute zero might make a more pro­duc­tive con­tri­bu­tion to char­ity by work­ing and donat­ing rather than per­sonal con­tri­bu­tion of time, but perfectly spher­i­cal ra­tio­nal­ists of uniform den­sity in a vac­uum at ab­solute zero are in some­what short sup­ply. In the world of hu­mans, a bit of hands-on par­ti­ci­pa­tion makes it far more likely that they will bother to con­tinue to con­tribute to that char­ity at all. • In the world of hu­mans, a bit of hands-on par­ti­ci­pa­tion makes it far more likely that they will bother to con­tinue to con­tribute to that char­ity at all. Ex­actly what I was try­ing to say, but much shorter! Thanks. • If you are to “love your neigh­bor as your­self” Why use that par­tic­u­lar phrase? I think I don’t need to love my neigh­bor as much as me to be in­ter­ested in char­ity. And while I sup­pose the phrase sits well aes­thet­i­cally in the text, I think it might un­for­tu­nately evoke with it a few Chris­tian cached toughs. Pure self­less­ness re­warded in af­ter­life don’t re­ally seem ap­pli­ca­ble to what peo­ple here want to do. • Any­one know what the prob­a­bil­ity of a whole blood or platelet dona­tion sav­ing a life is? That isn’t rated by GiveWell, and I failed at find­ing the data in a Google search. • The Red Cross claims that 1 pint saves “up to 3 lives”. I’m not sure what to make of that, given that it’s an up­per bound and pre­sented by a non-par­tial source. If any­one can do bet­ter, I would be very in­ter­ested in know­ing the an­swer. I always try to give blood as of­ten as pos­si­ble un­der the as­sump­tion that I save at least one life each time, but a more ro­bust figure would be nice. • I always try to give blood as of­ten as pos­si­ble un­der the as­sump­tion that I save at least one life each time That can’t pos­si­bly be right, not on the mar­gins. • Given you and wedrifid’s re­sponses, I am now up­dat­ing my es­ti­mate of num­ber of lives saved sig­nifi­cantly down­wards. How­ever, I am cu­ri­ous as to why it’s ob­vi­ous to you that 3 lives is too high of a num­ber on the mar­gins. • Maybe I’m just be­ing naive here, but in a case that straight­for­ward and that pos­si­ble for the av­er­age per­son to un­der­stand, where there’s noth­ing odd or un­pres­ti­gious about the ac­tion and lots of peo­ple are do­ing it already, where, on the mar­gins, an ad­di­tional Amer­i­can life is saved each time an­other per­son donates blood, I have trou­ble be­liev­ing that even a world this in­sane wouldn’t push blood dona­tions a lit­tle harder. • I take it that you’re sug­gest­ing marginal anal­y­sis based on the stan­dard cor­rect clas­si­cal causal de­ci­sion the­ory (in which no one is re­spon­si­ble for sav­ing a life by donat­ing blood un­less some­one would have ac­tu­ally died had that dona­tion not been made) out of ei­ther be­lated hu­mil­ity about the prob­a­bil­ity of an SIAI-origi­nat­ing de­ci­sion the­ory be­ing cor­rect, or be­cause you’re plan­ning to ac­tu­ally con­vince some­one and you don’t want to in­voke Hofs­tad­te­rian su­per­ra­tional­ity in place of the stan­dard cor­rect de­ci­sion the­ory? :) My guess would be that at the mar­gin, a blood dona­tion saves less than 0.00001 lives. (Other­wise, com­pen­sa­tion would be in­creased for the paid donors). But, if you want to use a TDT/​UDT style anal­y­sis, here are some rele­vant statis­tics from the Amer­i­can Red Cross: • The num­ber of blood dona­tions col­lected in the U.S. in a year: 16 mil­lion (2006). • The num­ber of pa­tients who re­ceive blood in the U.S. in a year: 5 mil­lion (2006). Given these num­bers, I would es­ti­mate that roughly 0.5 mil­lion (US) lives are saved (more ac­cu­rately, ex­tended) by blood prod­ucts an­nu­ally. If you adopt the as­sump­tion that all blood comes from vol­un­tary, un­com­pen­sated dona­tions, and di­vide those 0.5 mil­lion lives among the 16 mil­lion an­nual dona­tions, you get one life saved for ev­ery 32 pints donated—not as much as jstein­hardt hoped, but still sig­nifi­cant enough to earn a ma­jor warm-and-fuzzy. • I hap­pen to ad­minister a lot of blood to my pa­tients, so let me an­swer some of the fac­tual ques­tions. 1. The way they calcu­late “up to 3 lives” is in the most triv­ial way: blood you donate is frac­tionated into red cells, plasma, and platelets. Each of those may go to a differ­ent re­cip­i­ent. 2. All blood ad­ministered to pa­tients comes from vol­un­tary, un­com­pen­sated dona­tions. Plasma used in re­search stud­ies may be com­pen­sated, but may not be trans­fused. This is the most im­por­tant fac­tor keep­ing our blood sup­ply safe, and is far more effec­tive than lab­o­ra­tory test­ing alone. 3. Given that blood banks need to keep a suffi­cient store of blood available of each type, rarer blood types are gen­er­ally in greater need than, say, A After all, a larger pro­por­tion of blood of those types must be dis­carded. O blood is ob­vi­ously highly use­ful in trauma situ­a­tions, and is there­fore in high de­mand as well. 4. The dis­tri­bu­tion of donors’ and re­cip­i­ents’ blood types should not be as­sumed to be equal: peo­ple with blood type A are sig­nifi­cantly more likely to donate than peo­ple with blood type B. This ex­ac­er­bates the dis­crep­an­cies due to point 3. 5. The num­ber of lives saved can be calcu­lated in two ways: a. the feel-good way. Every time a physi­cian gives a unit of blood to a pa­tient e does so be­liev­ing it is a life-sav­ing pro­ce­dure. So if 3 units are given the pa­tient’s life was saved 3 times in rapid suc­ces­sion. (You have to be will­ing to save a life mul­ti­ple times, be­cause that’s the anal­y­sis we’re us­ing for the rest of this dis­cus­sion: mul­ti­ple mosquito nets saved the same kid’s life mul­ti­ple times over his life­time; that same kid was then saved by anti-di­ar­rheal treat­ments; etc. The same anal­y­sis be­longs here). Now, we sub­tract the num­ber of pa­tients who die, but that’s a small num­ber. So 26 mil­lion trans­fu­sions/​16 mil­lion dona­tions = 1.6 lives saved per dona­tion. b. the marginal way. Dona­tions are cur­rently suffi­cient for us­age; we benefit in three ways from more dona­tions. First, we can be slightly more profli­gate with trauma pa­tients who have a low sur­vival chance; this saves a min­i­mal num­ber of lives. Se­cond, fresher blood is as­so­ci­ated with bet­ter out­comes than older blood; the ex­tent of this effect is un­known but is an area of cur­rent re­search in­ter­est. The calcu­la­tion would have to look at the like­li­hood that your dona­tion re­duced the av­er­age shelf age of the blood be­ing ad­ministered times the sur­vival im­prove­ment from the fresher blood. Third, blood from mul­ti­parous women is as­so­ci­ated with ARDS; an in­crease in dona­tion would al­low us to stop us­ing it. • peo­ple with blood type A are sig­nifi­cantly more likely to donate than peo­ple with blood type B I’ve donated blood a few times and I’m type A+. Why is it that B’s are less likely to donate, or is that un­known? Are my dona­tions likely to be marginally use­less? I have mostly donated blood in the past for sig­nal­ing rea­sons, con­ver­sa­tional high ground, and a vague de­sire to match the 15-gal­lon mark that my grand­father got his name in the pa­per for. There’s a plaque of the news­pa­per men­tion in my grandma’s house and I’ve been look­ing at it my whole life. Also I figure the Red Cross will let me know if I come down with one of the dis­eases they screen for, and it’s a free way to get my iron lev­els checked (at­tempt­ing to donate blood was how I found out I was ane­mic in the first place). Th­ese rea­sons aren’t likely to evap­o­rate if I find that I have been sav­ing only tiny frac­tions of ex­pected lives, but I would prob­a­bly en­dure less in­con­ve­nience in or­der to donate for only these rea­sons as op­posed to these rea­sons on top of life­sav­ing. • Your dona­tions are not marginally use­less! (un­less you’ve been preg­nant a cou­ple times—in that case, con­sider stop­ping). The rea­son for the dis­crep­an­cies in dona­tion rates be­tween types A and B is both sim­ple and com­plex: eth­nic­ity. In the in­ter­ests of safety (avoidance of Hepatitis C, HIV, etc) we’ve set up a sys­tem that sub­tly en­courages cer­tain types of donors and dis­cour­ages oth­ers. The sys­tem is not racist per se, but it is most effec­tive in ob­tain­ing dona­tions from white, mid­dle-aged, mid­dle-class males. Re­gard­ing sig­nal­ing rea­sons: we are ob­vi­ously very afraid of blood donated for sig­nal­ing pur­poses. Ac­cord­ingly, we do not al­low peo­ple to donate to their rel­a­tives ex­cept un­der very un­usual cir­cum­stances. Ad­di­tion­ally, we give peo­ple an “out” by check­ing a box which tells the cen­ter to draw and dis­card their blood. That way peo­ple who fear they may be high-risk donors can get the so­cial ap­proval of donat­ing with­out harm­ing any pa­tients. • un­less you’ve been preg­nant a cou­ple times—in that case, con­sider stopping I’ve never been preg­nant, but what is it about mul­ti­ple preg­nan­cies that ren­ders the blood non-preferred? • Ob­vi­ous guess: Your blood then con­tains an­ti­bod­ies to the blood type of your ba­bies. • Essen­tially this. The A/​B/​O blood groups rep­re­sent the most rele­vant anti­gens in hu­man blood. There are a host of oth­ers (Rh, Duffy, Kell, etc.) which typ­i­cally cre­ate only minor prob­lems in a trans­fu­sion and which can be ig­nored in an emer­gency. But a per­son who has been ex­posed to al­lo­geneic blood via mul­ti­ple trans­fu­sions or preg­nan­cies be­comes more likely to de­velop an­ti­bod­ies to some of these anti­gens. The donor’s an­ti­bod­ies or white cells can re­act to the per­son be­ing trans­fused, caus­ing lung dam­age. • There are a host of oth­ers (Rh, Duffy, Kell, etc.) which typ­i­cally cre­ate only minor prob­lems in a trans­fu­sion and which can be ig­nored in an emer­gency. In the case of the Rh­e­sus fac­tor it should be noted that it is minor once and then only minor for males. Be­ing there­after un­able to safely give birth to healthy Rh+ chil­dren is definitely not a minor con­se­quence even if it is bet­ter than ‘prob­a­bly go­ing to die to­day’. (Un­less, I sup­pose, you hap­pen to some Rh+ an­ti­serum ly­ing around but no Rh- blood, which will usu­ally avoid the fu­ture difficul­ties.) • I use “minor” differ­ently than you do, to mean “un­likely to cause death”. Ob­vi­ously cross-matched blood is always prefer­able for a va­ri­ety of rea­sons (in­clud­ing pos­si­ble in­fer­til­ity, in the case of young fe­males). I would avoid RhoGam in the case of a pa­tient who needs a RBC trans­fu­sion, in­ci­den­tally. It would be un­likely to be safe or effec­tive. • I use “minor” differ­ently than you do, to mean “un­likely to cause death”. I would like to ex­press that my ap­proval of this phrase ex­tends be­yond the ca­pac­ity of up­vot­ing and into the ca­pac­ity of a com­ment ex­press­ing ap­proval. • I use “minor” differ­ently than you do, to mean “un­likely to cause death”. And hav­ing both your arms re­moved is “Just a flesh wound!”. • Oops, I see that this has already been asked. we’ve set up a sys­tem that sub­tly en­courages cer­tain types of donors and dis­cour­ages others While we’ve got you here, can you ex­plain why gay men can­not donate? This up­sets a lot of gay peo­ple that I know. I un­der­stand that it’s eas­ier to catch STDs (not just HIV/​AIDS) from a man than from a woman. But the cur­rent U.S. rule (A man can­not donate if he’s had sex with a man; a woman can­not donate if she’s had sex with a man who’s had sex with man.) is lop­sided. The even-handed rule that you can­not donate if you’ve had sex with a man would keep the sup­ply safe with­out hav­ing to rely on peo­ple’s be­ing able to trust their part­ners. But it would keep most women from donat­ing, so maybe it’s not worth it. The even-handed rule that you can­not donate if you’ve had sex with man who’s had sex with a man would still keep out most gay men, but it would prob­a­bly help to heal the rift. • The even-handed rule that you can­not donate if you’ve had sex with man who’s had sex with a man would still keep out most gay men If a man is gay and sex­u­ally ac­tive, he’s al­most cer­tainly had sex with a man who’s had sex with a man, even if the men he has had sex with has only had sex with him. I don’t see how this phras­ing of the rule would be an im­prove­ment. • My phras­ing was un­clear; make it “if you’ve had sex with man who’s pre­vi­ously had sex with a man (other than you)”. There wouldn’t be any point in for­bid­ding me from donat­ing (if I’m male) be­cause the man that I’ve had sex with has had sex with me! This change would in­clude more peo­ple; it in­cludes monog­a­mous gay male cou­ples who be­gan their re­la­tion­ship as vir­gins (as well as some other peo­ple). Not many more, but it makes it clear that the blood col­lec­tor is only will­ing to trust you and your part­ners, no fur­ther. Frankly, the first even-handed rule (no sex with a man, pe­riod) makes more sense to me. Why should the blood col­lec­tor trust that I know (if I’m a woman) whether all of the men that I’ve had sex with have had sex only with women? (No doubt many women are donat­ing con­trary to guidelines be­cause they don’t know this about their part­ners.) But be­cause this would cut the po­ten­tial donor pool in half, the blood col­lec­tor is ba­si­cally forced to trust me about my part­ners too. In fact, the blood col­lec­tors trust women to know the sex­ual his­tory of their part­ners, but not men. They are not ask­ing ev­ery­body the same ques­tions. • Another pos­si­ble solu­tion, not even-handed, but more hon­est: Just don’t ask women any­thing about the sub­ject. The idea that a per­son can be trusted to know about their part­ners’ part­ners is pre­pos­ter­ous; no other ques­tion (in the U.S.) asks the donor about other peo­ple’s be­havi­our, and for good rea­son. In­stead of half-ass­edly try­ing to be even-handed about it, just ad­mit what they’re do­ing: rul­ing out men who’ve had sex with men, be­cause many of their part­ners will have had sex with other men, and so on back (in many cases) a long way; but ac­cept­ing women who’ve had sex with men, be­cause most of their part­ners won’t have had sex with men, stop­ping the trans­mis­sion-from-men se­quence. I’m con­fi­dent that they already ac­cept blood from most women who’ve had sex with men who’ve had sex with men (be­cause the women don’t know this about their part­ners), and they are surely aware of this (if I am cor­rect) fact. So why are they ask­ing ques­tions of peo­ple who don’t ac­tu­ally know the an­swers? Gay peo­ple will still be up­set that they can’t donate, but I at least would be more will­ing to trust that the blood col­lec­tors are ac­tu­ally mak­ing an hon­est de­ci­sion. • Mostly, my faith in the qual­ity of the blood sup­ply de­rives from what test­ing they’re do­ing to the blood, not from what un­en­force­able poli­cies they’re sug­gest­ing to the donors. I’d ac­tu­ally be sur­prised if the lat­ter sig­nifi­cantly af­fected the qual­ity of the blood. Mostly, I think the prob­lem they are a solu­tion for is main­tain­ing pub­lic con­fi­dence in the blood sup­ply. Which I ac­knowl­edge is an im­por­tant prob­lem. And it may well be that be­ing per­ceived as ex­clud­ing gay men and their part­ners is a bet­ter solu­tion to that prob­lem than any­thing else they might do; I don’t know. That said, if I’m wrong and these poli­cies re­ally do solve a prob­lem re­lated to the blood sup­ply, yet an­other pos­si­ble solu­tion is: don’t al­low peo­ple who have had un­pro­tected sex to donate. Or, if that’s too big a chunk of your po­ten­tial donor base, make it peo­ple who have had un­pro­tected sex out­side of a monog­a­mous re­la­tion­ship. • The origi­nal rule bars ‘a man who has had sex with a man’ - X—and then any women who’ve had sex with X. It’s a log­i­cal phras­ing but un­for­tu­nately X maps ex­actly onto “gay man”, so it feels like gay men are be­ing speci­fi­cally tar­geted. The rephras­ing mol­lifies that sense of tar­get­ing with­out, as far as I can tell, chang­ing the in­cluded or ex­cluded peo­ple. The origi­nal phrase is even-handed, how­ever. If you over­speci­fied an even-handed rule and said “1) You can­not donate if you’re a man who has had sex with a man who has had sex with a man, and 2) you can­not donate if you’re a woman who has had sex with a man who has had sex with a man”—ie, pre­vent “man who has had sex with a man” from com­ing into sex­ual con­tact with any donor—you could re­duce 1) down to “man who has had sex with a man” (it log­i­cally im­plies three, four, and so on iter­a­tions). This, there­fore, re­duces down to the ac­tual rule they have in place. • Every time a physi­cian gives a unit of blood to a pa­tient e does so be­liev­ing it is a life-sav­ing pro­ce­dure. So if 3 units are given the pa­tient’s life was saved 3 times in rapid suc­ces­sion. (You have to be will­ing to save a life mul­ti­ple times, be­cause that’s the anal­y­sis we’re us­ing for the rest of this dis­cus­sion: mul­ti­ple mosquito nets saved the same kid’s life mul­ti­ple times over his life­time; that same kid was then saved by anti-di­ar­rheal treat­ments; etc. The same anal­y­sis be­longs here) There are not many times I see a line of rea­son­ing and have to re­ject it at ev­ery sin­gle step. Apart from be­ing con­cep­tu­ally ab­surd the very thought is morally ob­jec­tion­able. It to­tally de­val­ues the value of ‘sav­ing a life’ to the point of ut­ter mean­ingless. How could that ever make some­one ‘feel-good’? • It to­tally de­val­ues the value of ‘sav­ing a life’ to the point of ut­ter mean­ingless. Which part? I thought that started silly (it’s ex­plain­ing the logic be­hind a non-profit’s puffery, did you ex­pect it to be rigor­ous?) but then got bet­ter. The idea of “sav­ing a life” is pretty mean­ingless when you poke at it- it’s all just lifes­pan ex­ten­sion. And so the idea that each emer­gency treat­ment ex­tends lifes­pans by the ‘nat­u­ral span of a life’ is silly. If some­one would die if they don’t re­ceive a unit of blood at 50 sep­a­rate oc­ca­sions on their life, should each trans­fu­sion get the full moral weight of sav­ing a life? If so, we just gave this per­son 50 lives. If not, then we need to aban­don the lan­guage of “sav­ing a life” and talk about “ex­tend­ing a lifes­pan” (be­cause we can say those units of blood each added a year to the per­son’s life, for ex­am­ple). • Thanks, this is ex­cep­tion­ally in­for­ma­tive. I didn’t re­al­ize that dona­tions were suffi­cient for us­age. Is this barely main­tained by call­ing peo­ple when blood sup­plies are low, or does blood reg­u­larly get thrown out, or is there some other rea­son that sup­plies closely match need? • A com­bi­na­tion of the above. We have a core group of donors who can be called in emer­gency situ­a­tions, we in­crease the in­ten­sity of blood drives when sup­plies are low, we re­duce marginally-benefi­cial uses of blood when sup­plies be­come low, and we are bet­ter able to dis­card the old­est least-effec­tive blood when­ever sup­plies in­crease. We are likely to face challenges in meet­ing fu­ture need. The co­hort that most reg­u­larly donates blood is ag­ing... • We are likely to face challenges in meet­ing fu­ture need. The co­hort that most reg­u­larly donates blood is ag­ing... I’m as­sum­ing you are from the US, do you think the same is true for other coun­tries? Also which de­mo­graphic are you refer­ring too? • I’m as­sum­ing you are from the US, do you think the same is true for other coun­tries? Also which de­mo­graphic are you refer­ring too? I’m more in­ter­ested in the co­hort that isn’t ag­ing. What is their se­cret? A new and im­proved Calorie Restricted diet? Per­haps that ex­plains their in­abil­ity to gen­er­ate suffi­cient ex­cess blood for dona­tion. • They could also be brain up­loads, which would also ex­plain the in­abil­ity. • Good point. I won­der if they would con­sider donat­ing CPU time in­stead! • Thanks for data! Only vaguely re­lat­edly: if you have poin­t­ers to (or are will­ing to syn­the­size) a re­li­able calcu­la­tion of ex­pected lives-saved/​deaths-caused by main­tain­ing or dis­card­ing the ex­ist­ing Red Cross poli­cies about who is “al­lowed” to donate blood, es­pe­cially the rel­a­tively con­tro­ver­sial ban on male donors with ho­mo­sex­ual acts in their sex­ual his­tory, I would be in­ter­ested. Full dis­clo­sure: I do have a per­sonal/​emo­tional stake in this ques­tion, but I re­ally re­ally don’t want to set off a poli­ti­cal/​eth­i­cal con­ver­sa­tion about it. I’m ask­ing it here be­cause, as with a lot of poli­ti­cally charged top­ics, the ar­gu­ments I’ve found on both sides are mostly a case of fram­ing the ques­tion so as to give the an­swer one wants to give, rather than so as to an­swer the ques­tion that was asked, and I’m look­ing for a more ob­jec­tive anal­y­sis. • I also wanted to ask this ques­tion. Giv­ing blood is im­por­tant to me. It is so im­por­tant that I have cho­sen not to pur­sue re­la­tion­ships with other men in or­der than I can con­tinue to give blood with­out ly­ing to do so. I ex­pect that sooner or later, I will choose oth­er­wise, and a sex­ual re­la­tion­ship will be im­por­tant enough to me to sac­ri­fice my abil­ity to ever give blood again, and this dis­tresses me. I can ac­cept that the risks of HIV may be high enough to make this a rea­son­able choice on the part of United Blood Ser­vices /​ Red Cross. How­ever, I would like to be quite sure that this is the case, or to be told that my blood isn’t as im­por­tant as I pre­vi­ously though it was. I was pre­vi­ously giv­ing blood on the im­pres­sion that each dona­tion saves around a twen­tieth of a life; this thread doesn’t change that es­ti­mate enough for me to feel like I can stop donat­ing in good con­science. • Giv­ing blood is im­por­tant to me. It is so im­por­tant that I have cho­sen not to pur­sue re­la­tion­ships with other men in or­der than I can con­tinue to give blood with­out ly­ing to do so. On the mar­gins, I ex­pect that each marginal pint of blood saves only a very small frac­tion of a life. As sev­eral read­ers pointed out, this doesn’t mean that we should or­di­nar­ily be calcu­lat­ing on the mar­gins, since it’s not like you can use a pint of blood for some­thing else in­stead; in terms of moral credit, you should think of your­self as part of a refer­ence class of peo­ple who all choose to donate blood for around the same rea­sons, and who all get an equal share of the lives saved. How­ever, the Red Cross has already de­cided that they’re will­ing to X out the en­tire ho­mo­sex­ual com­mu­nity, and I would ex­pect the refer­ence class of those who re­frain from sex­ual ac­tivity in or­der to con­tinue donat­ing blood to be small, and I would guess that if this en­tire refer­ence class re­frained from donat­ing blood, not a sin­gle ad­di­tional life might be lost. Modern-day hos­pi­tals are not, so far as I know, blood-limited. They need a rou­tine flow of blood in or­der to rou­tinely save lives. They do not need more blood to save more lives. That’s the im­pres­sion I got, any­way; some quick Googling even said that they usu­ally have enough blood to just use O-nega­tive in­stead of match­ing types. I hate to say this, but I think you’re mak­ing the wrong sac­ri­fices here. I es­ti­mate a very high in­for­ma­tion value for fur­ther in­ves­ti­ga­tion on your part; I would ex­pect it to show that you were safe to stop donat­ing blood and re­sume sex­ual ac­tivity with­out cost­ing any­one one-twen­tieth of a life. If you’re re­ally feel­ing guilty or wor­ried, re­sume sex­ual ac­tivity and send a dona­tion to the Sin­gu­lar­ity In­sti­tute as a car­bon offset. If you can speed up a pos­i­tive Sin­gu­lar­ity by one minute that works out to around 100 lives, never mind in­creas­ing the prob­a­bil­ity. • I think I was ac­ci­den­tally mis­lead­ing by failing to add that I’m bi­sex­ual. Not giv­ing blood re­duces my pool of po­ten­tial ro­man­tic part­ners by roughly 10%, and doesn’t pre­vent me from hav­ing fulfilling re­la­tion­ships. I don’t think I would ab­stain from sex in or­der to give blood even if I knew I could save a life with each dona­tion. Even if that’s an in­cred­ibly self­ish de­ci­sion, I’m just not that good a per­son. Re­gard­less, the sup­port of ev­ery­one who replied is very much ap­pre­ci­ated. • ...tech­ni­cally, doesn’t speed­ing up a nega­tive sin­gu­lar­ity also save lives—the lives of those who would oth­er­wise have been born and then kil­led but were in­stead never born and there­fore couldn’t be kil­led? In fact, I think speed­ing up a nega­tive sin­gu­lar­ity ac­tu­ally “saves” more lives than speed­ing up a pos­i­tive one us­ing this calcu­la­tion—a quick Google search in­di­cates ~250 peo­ple are born ev­ery minute and ~100 peo­ple die ev­ery minute. • Re­place “save lives” with “ex­tend lifes­pans.” All the math will sud­denly start work­ing out bet­ter. • In a fairly mean­ingful sense, no life has ever been saved be­fore. No­body has ac­tu­ally been pre­vented from dy­ing yet. A pos­i­tive sin­gu­lar­ity could change that. • I be­lieve you can make an eas­ier calcu­la­tion: change the de­nom­i­na­tor from lives to units of blood. How much effort/​money/​so­cial cap­i­tal would it take you to con­vince one more per­son to donate one more unit? [ig­nore the cost to that per­son, as it’s likely zero or slightly benefi­cial]. Calcu­late the effort it there­fore would take you to re­place your­self as a donor while keep­ing the blood sup­ply con­stant; this should serve as an up­per bound for the self-sac­ri­fice you should make in terms of sex­ual re­straint. • You make an ex­cel­lent point. I clar­ified that the sex­ual re­straint re­quired is not as great as it may seem, but con­vinc­ing other peo­ple to donate reg­u­larly (I have done so at least twice in my life) is still much less of a sac­ri­fice. • (nods) For me, it’s not a prag­matic ques­tion of whether I donate or not: af­ter ~20 years in a mu­tu­ally monog­a­mous re­la­tion­ship, I am con­fi­dent that my donat­ing blood re­duces the per­centage of in­fected blood in the sup­ply, re­gard­less of my gen­der, and that’s the met­ric that mat­ters. But I spent some time try­ing to make sense of the ar­gu­ments pro and con, a few years back, and mostly came to the con­clu­sion that I didn’t trust any­one’s ar­gu­ments. It is cer­tainly true that if you di­vide the com­mu­nity of po­ten­tial donors into two groups, and the fre­quency of blood-born pathogens is higher in group A than group B, and your fil­ter­ing mechanisms aren’t 100% re­li­able, then the blood sup­ply is N% safer if you re­move group A from po­ten­tial donors. It is equally cer­tainly true that you can do that di­vi­sion in thou­sands of differ­ent ways, and each way of do­ing that di­vi­sion gets you a differ­ent N. I was hop­ing to find a com­par­i­son of es­ti­mated Ns for differ­ent plau­si­ble poli­cies, and per­haps a recom­men­da­tion for the best policy. What I found in­stead was that defen­ders of the ex­ist­ing policy were mak­ing the first ar­gu­ment and say­ing “See? The policy makes the blood sup­ply N% safer! We have to keep do­ing it, to do oth­er­wise would be un­safe!” while at the same time dis­re­gard­ing ques­tions about how large N ac­tu­ally was (i.e.., how many lives were ac­tu­ally at stake? 1000? .001? Some­where in be­tween?) and whether a differ­ent policy might get you a much larger N, while op­po­nents of the policy were dis­re­gard­ing the first ar­gu­ment al­to­gether. • But I spent some time try­ing to make sense of the ar­gu­ments pro and con, a few years back, and mostly came to the con­clu­sion that I didn’t trust any­one’s ar­gu­ments. My con­clu­sion is some­what re­lated. I have no par­tic­u­larly good rea­son to be­lieve that I am bet­ter able to es­tab­lish blood dona­tion and us­age policy than the Red Cross or the med­i­cal prac­ti­tion­ers. I just give them my blood and they can use it or not as they see fit. I’d do it just for the health benefits any­way. • For my own part, I ap­pre­ci­ate that the Red Cross (and etc.) is try­ing to satisfy mul­ti­ple con­straints, only one of which is the ac­tual safety of their blood sup­ply, and I don’t ob­ject to that. But the con­straints that ap­ply to them in ar­tic­u­lat­ing a policy don’t nec­es­sar­ily ap­ply to me in donat­ing blood. • On the other hand you have con­straints that they do not have, not least of which is the lack of scal­ing benefits for your re­search and de­ci­sion mak­ing efforts. We are left with an op­ti­mal ap­proach of con­sid­er­ing what we know of our own blood that the col­lec­tion agency does not (or is for­bid­den from dis­crim­i­nat­ing on). We can ap­prox­i­mate whether this knowl­edge would make the blood more suit­able or less. Only if ‘less’ do we need worry about how sig­nifi­cant that ex­tra knowl­edge is. • We also need to worry if the an­swer is ‘more’ and be­cause of that we de­cide to lie on the an­swer form so that we can donate. I kind of get the im­pres­sion that TheOtherDave is do­ing that, or at least would con­done it un­der cir­cum­stances very much like his. • I don’t do it, mostly be­cause I’m so ir­ri­tated by the policy that I’ve worked my way into a com­pletely coun­ter­pro­duc­tive “F—k it, then, donate your own f—king blood, see if I care” kind of sulk about it. I’m not proud of this, but there it is. Yes, I con­done it… in­deed, I en­dorse it… in situ­a­tions very much like mine. • They aren’t as­sess­ing that risk in a log­i­cal fash­ion. If they were, they would have similar re­stric­tions on dona­tion by eth­nic group. (It is pos­si­ble that the Red Cross would like to do that also but knows that it is poli­ti­cal un­fea­si­ble.) • Thx. All blood ad­ministered to pa­tients comes from vol­un­tary, un­com­pen­sated dona­tions. Plasma used in re­search stud­ies may be com­pen­sated, but may not be trans­fused. This is the most im­por­tant fac­tor keep­ing our blood sup­ply safe, and is far more effec­tive than lab­o­ra­tory test­ing alone. This ar­ti­cle on the ethics and prag­mat­ics of blood source—com­pen­sated vs un­com­pen­sated—was fas­ci­nat­ing, IMO. Though it may be some­what out-of-date. • This is Holden Karnofsky, the co-Ex­ec­u­tive Direc­tor of GiveWell, which is refer­enced in the top-level ar­ti­cle and el­se­where on this thread. I think there is an im­por­tant differ­ence be­tween dis­cussing the marginal im­pact of a blood dona­tion and the marginal im­pact of a vote. When it comes to blood dona­tions, it is pos­si­ble for ev­ery­one to si­mul­ta­neously fol­low the rule: “Give blood only when the sup­ply of dona­tions is low enough that an ad­di­tional dona­tion would have high ex­pected im­pact”, with a rea­son­able out­come. It is not pos­si­ble for ev­ery­one to be­have this way in elec­tions: no voter is able to con­sider the ex­ist­ing dis­tri­bu­tion of votes be­fore cast­ing their own. I am only ca­su­ally fa­mil­iar with TDT/​UDT, but it seems to me that that “Give blood only when the sup­ply of dona­tions is low enough that an ad­di­tional dona­tion would have high ex­pected im­pact” should get about the same amount of credit un­der TDT/​UDT as giv­ing blood, and thus the ex­tra im­pact of ac­tu­ally giv­ing blood (as op­posed to fol­low­ing that rule) is small re­gard­less of what de­ci­sion the­ory one is us­ing. I bring this up be­cause the dis­cus­sion of marginal blood dona­tions is par­allel to anal­y­sis GiveWell of­ten does of the marginal im­pact of dona­tions. We do ev­ery­thing we can to un­der­stand the marginal (not av­er­age) im­pact of a dona­tion and recom­mend or­ga­ni­za­tions on this ba­sis, and we be­lieve this is a very im­por­tant and unique el­e­ment of what we offer (more on this is­sue). We try to push donors to un­der­funded char­i­ties and away from over­funded ones, and I do not think the val­idity of this de­pends on any con­tro­ver­sial (even con­tro­ver­sial-within-Less-Wrong) view on de­ci­sion the­ory, though I am open to ar­gu­ments that it does. • Com­pletely agree with your gen­eral point on marginal anal­y­sis (al­though I’m a TDT skep­tic), and am a fan of GiveWell, but this is triv­ially wrong: It is not pos­si­ble for ev­ery­one to be­have this way in elec­tions: no voter is able to con­sider the ex­ist­ing dis­tri­bu­tion of votes be­fore cast­ing their own. This seems to as­sume away in­for­ma­tion about the size of the elec­torate as well as any pre­dic­tive power about the out­come. Surely the marginal benefit of a Pres­i­den­tial vote in a small swing state is mas­sively higher than in a large solidly Demo­cratic state, for ex­am­ple. And in ad­di­tion to his­tor­i­cal re­sults, there is pol­ling data in ad­vance of the elec­tion to im­prove pre­dic­tions. Be­sides this be­ing the­o­ret­i­cally true, we can see it em­piri­cally from the spend­ing pat­terns of both Pres­i­den­tial cam­paigns and poli­ti­cal par­ties on Con­gres­sional races. They al­lo­cate money to the states /​ races where they be­lieve it will do the most marginal good, which is of­ten a very in­equal dis­tri­bu­tion. Thus they do, in fact “con­sider the ex­ist­ing dis­tri­bu­tion of votes be­fore cast­ing” their ad­ver­tis­ing dol­lars. • Pa­tris­simo, fair enough. I was think­ing that vot­ers can’t vote with the same de­gree of knowl­edge of the ex­ist­ing situ­a­tion that they can have with blood dona­tions. Ar­gu­ments over TDT cer­tainly seem more rele­vant to vot­ing than to blood dona­tions. But you are right that vot­ers have lots of rele­vant in­for­ma­tion about the likely dis­tri­bu­tion of votes that can be pro­duc­tively fac­tored into their de­ci­sions re­gard­less of the TDT de­bate. Glad to hear you’re a fan of GiveWell. • 0.00001 sounds low to me. Given that hos­pi­tals aren’t nor­mally blood-limited, there are always fluc­tu­a­tions around the av­er­age, and I’d be sur­prised if be­com­ing blood-limited hap­pens less than 1 day in 10^5. Two trauma cases can be enough to cre­ate a lo­cal crisis • My ac­tions and your ac­tions aren’t perfectly cor­re­lated, be­cause we’re some­what differ­ent. No mat­ter how you han­dle this, it seems to sug­gest that my dona­tion would acausally af­fect some frac­tion of other peo­ple’s dona­tions. So it might count as, e.g., +/​- 2 mil­lion, which is still more-or-less marginal, since the costs are mul­ti­plied as well. Maybe more than that? It’s stil a far cry from just di­vid­ing 5 mil­lion/​16 mil­lion. Edit: Isn’t util­ity max­i­mized if the ab­stract com­pu­ta­tions “What hu­mans do” and “The thing with great­est marginal benefit” equal­ized, though? If util­ity is con­vex, yes. So there should be some other rule, like, if you’re at a bad equil­ibrium, act so as to break it. I am un­sure how this works. • If you adopt the fic­tion that all blood comes from vol­un­tary, un­com­pen­sated donations I know that plasma and such are com­pen­sated, but where is blood paid for? Places where it’s cheaper than trans­port­ing it from ar­eas that have sur­plus vol­un­teer efforts? Or they don’t pub­li­cize that com­pen­sa­tion is available be­cause that would shrink the vol­un­teer base? • How­ever, I am cu­ri­ous as to why it’s ob­vi­ous to you that 3 lives is too high of a num­ber on the mar­gins. Around 15 mil­lion pints of whole blood are donated per year in the US. At 3 lives per pint that comes out to 45 mil­lion. We can also as­sume that if lives per pint is 3 at the mar­gin then the more effi­cient cases it will be even more than that. The pop­u­la­tion of the US just isn’t high enough to ac­count for that. Oh, then there there is the fact that a lot of cases use a whole heap more than one pint of blood. (For ex­am­ple.) • Dead ba­bies or chil­dren are a bad met­ric pre­cisely be­cause of this rea­son. Years in good phys­i­cal and men­tal health seem a bet­ter way to mea­sure what peo­ple are go­ing for. A dona­tion of blood saves less than one life in my es­ti­mates, but it im­proves qual­ity of life and adds in my opinion a few years of healthy happy life. • I won­der if get­ting too fo­cused on the best (or worst) case sce­nario is a named log­i­cal er­ror. I’m also not sure whether giv­ing a rare blood type is likely to save more lives than giv­ing a more com­mon blood type. • It seems odd the more peo­ple of a cer­tain blood type would sus­tain in­juries re­quiring blood, or that peo­ple of a cer­tain blood type would care more about blood dona­tion So if the peo­ple who heard about blood banks and were in­ter­ested, and the peo­ple who needed them are nearly ran­domly dis­tributed in the pop­u­la­tion, I would ex­pect the de­mand vs sup­ply of each type to av­er­age out to the same figure. How­ever, I have heard blood banks ask speci­fi­cally for peo­ple with rare blood types to donate, so it would ap­pear that this the­ory is wrong. Alter­na­tively, there is an equal short­age of all types, and some­one in mar­ket­ing thought that the speci­fi­ca­tion would at­tract more peo­ple. (Even though if I was go­ing to use the dark arts to make a group more likely to come, I would tar­get the largest one) • It seems odd the more peo­ple of a cer­tain blood type would sus­tain in­juries re­quiring blood, or that peo­ple of a cer­tain blood type would care more about blood dona­tion. Blood types vary by eth­nic­ity, SES varies by eth­nic­ity, in­juries and dona­tions vary by SES. • For the same rea­son donat­ing or­gans if one is mixed race is worth more (than or­gan dona­tions by in­di­vi­d­u­als of pre­dom­i­nantly mono­ra­cial an­ces­try) be­cause of com­pli­ca­tions in com­pat­i­bil­ity for the de­mo­graphic. • The rea­son for this is the com­pat­i­bil­ity of the blood types, for ex­am­ple O-nega­tive-blood can be donated to ev­ery­one and is there­fore used in emer­gen­cies where the blood type of the re­cip­i­ent is not known. • I know pretty much noth­ing about the me­chan­ics of sav­ing and donat­ing blood, but I’d ex­pect lo­gis­ti­cal effects to push de­mand for rare types up even if the ac­tual need for them is pro­por­tional to their rar­ity. • The Red Cross claims that 1 pint saves “up to 3 lives”. I’m not sure what to make of that, given that it’s an up­per bound and pre­sented by a non-par­tial source. I pre­sume it means that the first pint of blood donated, if al­lo­cated with effi­cient triage, could save three lives. At the mar­gin I as­sume the figure is a small frac­tion of a life per pint donate. • I won­der how many peo­ple would just give up on char­ity al­to­gether if they ac­cepted this ar­gu­ment. I know I did, I pretty much see char­ity as su­pereroga­tory now. • Don’t peo­ple nor­mally see it as su­pereroga­tory? • And like­wise, there is only one best char­ity: the one that helps the most peo­ple the great­est amount per dol­lar. I dis­agree. Giv­ing money to char­ity is not differ­ent from spend­ing money on a latte at Star­bucks. I spend money ac­cord­ing to my val­ues. And I still buy lat­tes. I am not Zachary Baumk­let­terer. Even Je­sus said, “The poor you will have with you always”, to jus­tify spend­ing an INCREDIBLE amount of money (enough to buy ten peo­ple’s en­tire lives, in an era with no in­fla­tion, mak­ing it com­pa­rable to ten mil­lion US dol­lars to­day) on pour­ing per­fume once on Je­sus’ feet. The guy was tired and de­pressed and about to be cru­ci­fied and wanted his damn per­fume, like I want my damn latte. Similarly, peo­ple who gave money to keep a paint­ing in a mu­seum, might also spend con­sid­er­ably more money to buy paint­ings to hang in their houses, than it would take to save a life in an­other coun­try. Th­ese peo­ple value art, and they value benefit­ting oth­ers. Draw a 2D plot, and la­bel the axes “self­ish … un­selfish” and “spiritual … phys­i­cal” (“spiritual” stand­ing for art and other “im­prac­ti­cal” val­ues). One per­son might • buy a paint­ing to hang in their bed­room (spiritual, self­ish) • buy a paint­ing to hang in their guest room (spiritual, sorta self­ish) • spend to pre­serve a paint­ing in a mu­seum (spiritual, un­selfish) • buy fuzzy slip­pers (phys­i­cal, self­ish) • spend money for vac­cines in Africa (phys­i­cal, un­selfish) And each of those things could have similar util­ity for them. I don’t think this is ir­ra­tional. Ir­ra­tional is spend­ing any money at all on “char­ity” in­stead of spend­ing it ac­cord­ing to your util­ity func­tion. This post con­tains the hid­den pre­sup­po­si­tion that char­ity, us­ing a col­lec­tive util­ity func­tion, is more moral than self-ori­ented ac­tions; and there­fore, fol­low­ing our util­ity func­tions is im­moral. This is an as­ser­tion about moral­ity and ra­tio­nal­ity that has huge im­pli­ca­tions! It is res­o­nant with a very com­mon meme that says that “moral” be­hav­ior is be­hav­ior that we don’t want to do, be­cause we are fun­da­men­tally im­moral. I say, in­stead, that morals are part of our util­ity func­tion—that we have these things called morals be­cause part of us re­ally wants to be nice to other peo­ple. They are just an­other part of our util­ity func­tion. En­courag­ing un­selfish be­hav­ior can be done by ma­nipu­lat­ing peo­ples’ self­ish de­sires to pro­duce “un­selfish” be­hav­ior (give to char­ity and get so­cial benefits, or stay out of Hell), as a mechanism to solve PD prob­lems with a given pay­off ma­trix. But it can also be done by treat­ing peo­ple in ways that en­courage what nat­u­ral un­selfish ten­den­cies they have—solv­ing PD prob­lems by chang­ing peo­ple’s pay­off ma­tri­ces. Ap­ply Kant’s im­per­a­tive. This post sug­gests that we have 2 util­ity func­tions, one for ev­ery­day life, and an­other for char­ity; and that the one for char­ity is more moral. But if ev­ery­one used such a char­ity util­ity func­tion for ev­ery­thing they did, it would re­sult in a global race to the bot­tom as economies im­ploded af­ter spend­ing all na­tional wealth on ame­lio­rat­ing suffer­ing while un­der­cut­ting all pri­vate mo­ti­va­tion. There­fore, it is less moral. It is not only not ob­vi­ously moral, it is im­moral, if that means any­thing, for a gov­ern­ment, or a per­son, to spend ev­ery last dol­lar on helping the un­for­tu­nate be­fore spend­ing any money on ed­u­ca­tion, roads, defense, art, or even en­ter­tain­ment. • To some de­gree, this ar­ti­cle is less about mor­al­iz­ing and more of a “how to” guide. If you want to help peo­ple, this is how to do it. If you don’t want to help peo­ple, and you pre­fer to have lat­tes or works of fine art or what­ever, then a how-to guide on how to help peo­ple isn’t rele­vant to your in­ter­ests. To the de­gree that it is more than that, the ar­ti­cle is an at­tempt to ex­pose cer­tain thought pro­cesses into con­scious­ness so that they can be eval­u­ated by con­scious sys­tems. Peo­ple may be donat­ing to these in­effi­cient char­i­ties be­cause they feel like it and they don’t ex­am­ine their feel­ings, even though if they were to con­sciously think the prob­lem through they would give to more effi­cient char­i­ties. If, af­ter re­al­iz­ing that the choice is be­tween one kid’s life or 1/​1000 of a paint­ing, some­one still prefers the paint­ing, I don’t re­ally have any­thing more I can say—but my guess is that’s not a lot of the pop­u­la­tion. You made a re­ally good point in your mys­ti­cism post on Dis­cus­sion, about the differ­ence be­tween cat­e­go­riz­ing things by their causes and cat­e­go­riz­ing things by their effects. When you talk about spiritual and un­selfish choices, you’re cat­e­go­riz­ing things by their causes—a dona­tion to the paint­ing come from the same warm feel­ings that also pro­duce a dona­tion to vac­cines. Effi­cient char­ity is about cat­e­go­riz­ing things by their effects—it doesn’t mat­ter how no­ble the feel­ings that pro­duced a cer­tain ac­tion, only how much that ac­tion did what you wanted it to do. If you want to help peo­ple, it’s about how many peo­ple you helped. Cat­e­go­riz­ing things by their causes is an aca­demic ac­tivity that can only de­clare some peo­ple to be more “un­selfish” than oth­ers and ac­cord them brag­ging rights. In my opinion this doesn’t have as much to do with the ac­tual work of sav­ing the world as cat­e­go­riz­ing things by effects. You say this ar­ti­cle claims things about moral­ity, but that’s re­ally not its pur­pose. Its pur­pose is—if you’ve seen all sorts of hor­rible things in the world, and it’s reached the point where you’re so mad you don’t care what can or can’t be clas­sified as moral, you just want to fix things as quickly as pos­si­ble, what do you do then? I think the idea of some­thing to pro­tect is rele­vant here. • It should be noted that if you want to help peo­ple then donat­ing some­thing helps more peo­ple than be­ing dis­cour­aged to the point of not donat­ing at all due to the pos­si­bil­ity that your con­tri­bu­tion might be used some or­ders of mag­ni­tude less effec­tively than pos­si­ble. Many peo­ple do not (yet) have the abil­ity (or nerves/​time etc.) to read up on and make sense of the ar­gu­ments, or the data, to sub­se­quently com­pute the an­swer of what would be the most effec­tive way to spend their money in case they want to help other peo­ple. So be­fore you give up and do not donate any­thing at all, bet­ter split your money and give some to the SIAI (or even Wikipe­dia etc.). Ad­di­tion­ally use a ser­vice like GiveWell. And also don’t worry helping to ex­hibit some paint­ing. All of those con­tri­bu­tions will help some peo­ple, if only by mak­ing them happy (as in the case of the paint­ing). It will make a differ­ence! And it will make a huge differ­ence com­pared to do­ing noth­ing at all. • In­deed, it’s re­mark­able how lit­tle we would have to spend to end the worst poverty and in­jus­tice in the world to­day, if only peo­ple were will­ing to do it. We liter­ally spend more on cat food than it would take to elimi­nate the UN ab­solute poverty level. • To some de­gree, this ar­ti­cle is less about mor­al­iz­ing and more of a “how to” guide. The spe­cific quote the grand­par­ent was re­ply­ing to is about mor­al­iz­ing. And like­wise, there is only one best char­ity: the one that helps the most peo­ple the great­est amount per dol­lar. One could strip the mor­al­iz­ing el­e­ment from the quote (and the ar­ti­cle) in a fairly straight­for­ward man­ner. The best char­ity some­one can donate to is sub­jec­tively ob­jec­tive: the one that achieves the most benefit per dol­lar ac­cord­ing to that per­sons val­ues, al­tru­is­tic or oth­er­wise. • The spe­cific quote the grand­par­ent was re­ply­ing to is about mor­al­iz­ing. The prob­lem with the word “best” there is the same prob­lem the word “good” always runs into—the differ­ence be­tween “a good car” and “a good per­son”. I’m us­ing “best char­ity” in the same sense I would use “best Arc­tic sur­vival gear”—best at achiev­ing the pur­pose you are as­sumed to have. Although I think there is a case for that also be­ing the morally best for most moral sys­tems in which “morally best” makes sense, that would be way out­side the scope of this dis­cus­sion. • I un­der­stand what you are do­ing in the post and fol­low the sense of ‘best’. What I am ob­serv­ing is that the claim “you are mor­al­iz­ing” is fac­tu­ally cor­rect. The mor­al­iza­tion is not in the form of a di­rect ‘should’ nor is it in the way in which you use best. It can be seen here: best at achiev­ing the pur­pose you are as­sumed to have. That is an ex­tremely pow­er­ful moral gam­bit. • What a pro­vok­ing ar­ti­cle—ex­cel­lent! It’s healthy for us to be ask­ing these ques­tions. But I won­der about the du­al­is­tic na­ture of the ques­tions posed in your ‘how to guide’. Some­times, in fact of­ten, it is not a sim­ple choice be­tween two. Bio­di­ver­sity, like cul­ture, is much more com­plex than a graph can de­pict. The mul­ti­ple lay­ers move at differ­ent rhythms & speed and are in­structed by differ­ing mo­ti­va­tions such as hor­mone, in­stinct, sex, sur­vival, power, em­pa­thy (to name only a few). My point is that sys­temic change is not a mat­ter of choos­ing be­tween the best char­ity—that ap­proach only has one out­come which is how many lives to save in one mon­e­tary act—if we look at the world in a con­nected web than demon­strat­ing em­pa­thy & care by look­ing af­ter one’s place (clean­ing up the lo­cal beach) or pro­tect­ing a rain­for­est for the fu­ture health of the planet—these are all re­spon­si­bil­ities with differ­ent im­pacts that con­tribute to a greater whole. Helping a rain­for­est now may save mil­lions of lives in the fu­ture com­pared to 10 lives treated for malaria now. And this is not just about hu­mans! I don’t think you can mea­sure what you are try­ing to mea­sure—it de­nies the com­plex­ity of life and re­duces it to an eco­nomic plan. Yes you can look at a ‘how to guide’ if you want to find the best char­ity and you do make great ex­am­ples of how to make that de­ci­sion—but sus­tain­ing life and sur­vival is much deeper, chaotic and un­known. • Even Je­sus said, “The poor you will have with you always”, to jus­tify spend­ing an INCREDIBLE amount of money (enough to buy ten peo­ple’s en­tire lives, in an era with no in­fla­tion, mak­ing it com­pa­rable to ten mil­lion US dol­lars to­day) on pour­ing per­fume once on Je­sus’ feet “Even Je­sus”? Does it oc­cur to you that mak­ing this a re­li­gious ex­am­ple is ac­tu­ally even MORE likely to get us to no­tice the moral dis­so­nance, not con­vince us to ex­cuse it? • It is not only not ob­vi­ously moral, it is im­moral, if that means any­thing, for a gov­ern­ment, or a per­son, to spend ev­ery last dol­lar on helping the un­for­tu­nate be­fore spend­ing any money on ed­u­ca­tion, roads, defense, art, or even en­ter­tain­ment. This seems a false di­chotomy; the un­for­tu­nate will also be helped by money spent on ed­u­ca­tion, roads and other mea­sures which in­crease the com­mon good (so long as they do not make the plight of the un­for­tu­nate worse). Whether to spend money on medicine for the sick, ed­u­ca­tion for those who can­not get ac­cess to it with their own re­sources, or art and eter­tain­ment by which a cul­ture might ex­am­ine these prob­lems strikes me as be­ing a bit like med­i­cal triage in an emer­gency room. Per­haps it makes sense to treat per­sonal re­source man­age­ment similarly. • Well, think of it this way: What would an econ­omy look like, if ev­ery­one in it obeyed the max­ims of Peter Singer? It seems to me it would be a com­plete mess, far worse than what we have to­day. Now, if ev­ery­one in the world gave just a small amount of their in­come (5%? 10%?) to a wide va­ri­ety of char­i­ties they care about—e.g. sci­en­tific re­search, medicine, eco­nomic de­vel­op­ment, and yes, arts and cul­ture—we would get all the benefits of our pre­sent sys­tem and elimi­nate a lot of the worst flaws. US GDP is$14 trillion. US de­vel­op­ment aid and pri­vate char­ity are more like $300 billion (about 2% if you’re play­ing at home). Step that up to$600 billion, or $1 trillion, and what we could ac­com­plish! But I don’t think we’re go­ing to get there by mak­ing peo­ple feel guilty about sup­port­ing one thing rather than an­other. Far bet­ter, it seems, to get them to just make a habit of writ­ing a check—think of it like an­other bill to pay—and not wor­ry­ing so much about whether it is go­ing the best pos­si­ble place. • Well, think of it this way: What would an econ­omy look like, if ev­ery­one in it obeyed the max­ims of Peter Singer? They would donate up un­til the point of diminish­ing marginal re­turns as de­ter­mined by ex­perts in the rele­vant fields and then spend on them­selves. Seems like a pretty good world. • It seems like a world in which most re­sources are con­trol­led by “ex­perts in rele­vant fields.” When I con­sider this pos­si­ble world should I imag­ine it with the ex­perts we have now, or with more ideal­ized ex­perts? • (Sorry for bad html, I’ll try to learn to use the in­ter­face when I’m next at a real com­puter.) • When re­ply­ing to a com­ment, click the “help” link to the right of the “can­cel” but­ton (it’s all the way over in the cor­ner). • Thanks. • I just had a con­ver­sa­tion with my father on this sub­ject which sig­nifi­cantly clar­ified my think­ing, and re­solved most of my in­ter­nal dilemma. The ar­gu­ment put for­ward in this post is cor­rect, but there is one sig­nifi­cant prob­lem. I care about more than just sav­ing chil­dren. I also care about how effi­ciently it is done, what periph­eral good a char­ity is do­ing in the com­mu­nity by, say, em­ploy­ing lo­cals, and any num­ber of other things. “Chil­dren saved” is an im­por­tant met­ric and should ab­solutely be con­sid­ered, and it is a de­ci­sion that should be made care­fully, but it is not the only met­ric to con­sider. We should spend our money effi­ciently, but we first need to clar­ify our goals in or­der to do so, and “sav­ing chil­dren” is not nec­es­sar­ily our only goal, even in cases where it is pri­mary. • “sav­ing chil­dren” is not nec­es­sar­ily our only goal Un­less you have a huge “they are in an­other coun­try” dis­count on chil­dren’s lives, or a huge “they are in my com­mu­nity” boost to the other goals, I can’t name any goals off the top of my head that can com­pete with sav­ing chil­dren’s lives. • I didn’t say that other goals could com­pete, but there are other goals that can be con­sid­ered si­mul­ta­neously. If one char­ity saves ten chil­dren for$100 and an­other saves nine and ac­com­plishes a few other things, that is not a choice we should make mind­lessly. we can’t let “sav­ing chil­dren be­come a buz­zword that cuts off thought. What if the sec­ond char­ity saves the chil­dren from death and gives them some skills that will help them make a liv­ing and help their com­mu­ni­ties? In that case, I would prob­a­bly choose the sec­ond char­ity. Think of it as a lin­ear alge­bra prob­lem, with nu­mer­ous pa­ram­e­ters with differ­ent weights. You end up with an op­ti­mal solu­tion for all vari­ables to­gether rather than for a sin­gle vari­able alone. Just be­cause sav­ing chil­dren is the most heav­ily weighted vari­able doesn’t mean that it is the only one.

• At the risk of pro­vok­ing defen­sive­ness I will say that it re­ally sounds like you are try­ing to ra­tio­nal­ize your prefer­ences as be­ing ra­tio­nal when they aren’t.

I say this be­cause the ex­am­ples that you were giv­ing (lo­cal food kitchen, pub­lic ra­dio), when com­pared to truly effi­cient char­i­ties (save lives, im­prove health, foster lo­cal en­trepreneur­ship), are noth­ing like “save 9 kids + some other benefits” vs. “save 10 kids and noth­ing else”. It″s more like “save 0.1 kids that you know are in your neigh­bor­hood” vs. “save 10 kids that you will never meet” (and that’s prob­a­bly an over­es­ti­mate on the lo­cal op­tion). Your choice of a close num­ber is sus­pi­cious be­cause it is so wrong and so ap­peal­ing (by jus­tify­ing the giv­ing that makes you happy).

The amount of hap­piness that you cre­ate through lo­cal first world char­i­ties is or­ders of mag­ni­tude less than third world char­i­ties. There­fore, if you are choos­ing lo­cal first world char­i­ties that help “mal­nour­ished” kids who are fab­u­lously nour­ished by third world stan­dards, we can in­fer that the weight you put on “sav­ing the lives of chil­dren” (and with it, “max­i­miz­ing hu­man qual­ity-ad­justed life years”) is ba­si­cally zero. There­fore, you are al­most cer­tainly buy­ing warm fuzzies. That’s con­sump­tion, not char­ity. I’m all for con­sump­tion, I just don’t like peo­ple pre­tend­ing that it’s char­ity so they can tick their men­tal “give to char­ity” box and move on.

• I agree with you com­pletely about con­sump­tion vs. char­ity, and had even men­tioned the con­cept in my point about NPR dona­tion guilt.

I also agree that the close num­ber is wildly in­ac­cu­rate, but even in con­text it wasn’t ap­plied to lo­cal char­i­ties and it was in­tended to make the point that mul­ti­ple fac­tors could and should be con­sid­ered when pick­ing char­i­ties, even when the im­por­tance mul­ti­pli­ers on some fac­tors are or­ders of mag­ni­tude higher than for other fac­tors.

I hope this clar­ifies my mean­ing with­out defen­sive­ness, be­cause none was meant.

• Ok, great, I’m glad I mi­s­un­der­stood.

• Let’s say you want to start a school, be­cause you like ed­u­ca­tion. You could found a very large school that ed­u­cates lots of chil­dren, but at a so-so qual­ity. Or you could spend the same amount of in­struc­tion to make a tiny, amaz­ing school, a lit­tle gem. Some peo­ple might find it more fulfilling to build the small, won­der­ful school. When you’ve achieved your goal, a tiny cor­ner of the world is just perfect, and it’s a part you have con­trol over.

I think this is part of the rea­son peo­ple some­times are more mo­ti­vated to im­prove con­di­tions in their own coun­try than abroad. On some level, I’d rather make one per­son re­ally happy and suc­cess­ful than make 100 peo­ple just barely bet­ter off than dead.

• Think of it as a lin­ear alge­bra prob­lem, with nu­mer­ous pa­ram­e­ters with differ­ent weights. You end up with an op­ti­mal solu­tion for all vari­ables to­gether rather than for a sin­gle vari­able alone.

This is what I had in mind; I just felt that that the “sav­ing chil­drens’ lives” vari­able would have a mul­ti­plier of a few hun­dred in front of it (be­cause lives are im­por­tant) and the other vari­ables like “im­proves their com­mu­nity” would have mul­ti­pli­ers of two or three at best. I couldn’t think of any other vari­able that would have a similar mul­ti­plier to “child’s life”.

• Some of those other vari­ables will feed back in to the “child’s life” vari­able, a gen­er­a­tion or two down the road.

• Feed­ing back into “child’s life” a few gen­er­a­tions down the road is not a mul­ti­plier of a few hun­dred. That it feeds back gives it an ex­tra 10% or so; even with the feed­ing back, do­ing any­thing that isn’t di­rectly sav­ing as many lives as pos­si­ble right now is an ob­jec­tively worse op­tion.

• I’m both­ered by the in­tertem­po­ral im­pli­ca­tions of this, i.e. if I have $100 that I will spend to help the most hu­mans pos­si­ble, then I could ei­ther spend it to­day or in­vest it and spend$105 next year (as­sumed 5% ROR). Will I then ever spend the money on char­ity? Or will I always in­vest it, and just let this amassed wealth be dis­tributed when I die?

• As­sum­ing that char­i­ties can in­vest and bor­row at pre­vailing in­ter­est rates (and large char­i­ta­ble trusts can in fact bor­row from their en­dow­ment), you should be in­differ­ent to this choice. Robin Han­son has ad­dressed this is­sue here.

• The good you do can com­pound too. If you save a childs life at $500, that child might go on to save other chil­drens lives. I think you might well get a higher rate of in­ter­est on the good you do than 5%. There will be a sav­ings rate at which you should save in­stead of give, but I don’t think we’re near it at the mo­ment. • This, in­ci­den­tally, is also an ar­gu­ment for sup­port­ing less im­me­di­ately-effi­cient char­i­ties. If you spend$500 on mosquito nets, you are sav­ing the life of a child whose ex­pected life­time earn­ing po­ten­tial is low. This is won­der­ful, but the rate of “in­ter­est” may well be small. If you spend $500 on sav­ing the paint­ing Blue Rigi, you have not saved a sin­gle life in the short run. But it con­tributes to the ed­u­ca­tion of thou­sands of Bri­tish chil­dren, many of whom will grow up to cre­ate and donate large amounts of wealth/​knowl­edge. Your in­cre­men­tal im­pact on their ed­u­ca­tion may plau­si­bly pre­vent more malar­ial deaths than your dona­tion of mosquito nets, though I’ve no idea how to calcu­late this. At the very least, I’d sug­gest that anal­ogy of “set­ting out on an Arc­tic jour­ney” sets us up to men­tally dis­count fu­ture benefits in fa­vor of im­me­di­ate re­sults. In­stead we might imag­ine that we’ve set up an Arc­tic village, or are plan­ning a jour­ney a decade from now. Our spend­ing habits would change ac­cord­ingly. • If you spend$500 on sav­ing the paint­ing Blue Rigi, [...] it con­tributes to the ed­u­ca­tion of thou­sands of Bri­tish chil­dren, many of whom will grow up to cre­ate and donate large amounts of wealth/​knowledge

Con­tributes how much? For each child, how much more knowl­edge do you ex­pect they will cre­ate be­cause they saw the origi­nal, rather than a fac­simile, Blue Rigi? My es­ti­mate for this is so close to 0 that I can’t con­science pay­ing even $1 for Blue Rigi, ex­cept for aes­thetic rea­sons. • Is this an­other way of say­ing that schools should fo­cus on math and sci­ence, ig­nor­ing art? Or is this an ar­gu­ment that we need to re­struc­ture the way pub­lic mu­se­ums work, slash­ing the cost by re­plac­ing the paint­ings with copies? • It’s just an ar­gu­ment that art is not in the same bucket as sav­ing lives. I’m not go­ing to tell you how to spend your money, but if your stated ob­jec­tive is to help peo­ple, sav­ing Blue Rigi is not a cost effec­tive way of do­ing that. The way we run schools, math and sci­ence aren’t very use­ful to be­gin with. Slash­ing art bud­gets is prob­a­bly not a use­ful place to start. • Well, I want to make sure I un­der­stand it. Which of the fol­low­ing do you mean: a. If Bri­tish peo­ple be­come more pro­duc­tive that pro­duc­tivity won’t trans­late into more char­ity/​in­ven­tions that will save lives? b. Ed­u­ca­tion does not im­prove pro­duc­tivity? c. Art mu­se­ums are not an im­por­tant part of ed­u­ca­tion (at least not in terms of sci­en­tific/​eco­nomic pro­duc­tivity)? d. Blue Rigi does not im­prove the over­all qual­ity of the Tate? e. Ac­tu­ally none of the above, but Blue Rigi was sim­ply priced too high? To clar­ify/​ad­dress ArisKat­saris’s points: I am not at­tempt­ing to make an ar­gu­ment in this post. I am try­ing to iden­tify the point at which data­dataev­ery­where first has a prob­lem. For in­stance, I don’t need to dis­cuss whether the cul­tural given (fetish?) that our mu­se­ums will seek out origi­nals is eas­ily muta­ble if his ob­jec­tion re­ally starts ear­lier in my list. For in­stance, is it pos­si­ble that the ed­u­ca­tion of Bri­tish chil­dren is a bet­ter way to save Afri­can lives than the im­me­di­ate pur­chase of mosquito nets? If that’s im­plau­si­ble, then the ques­tion of how one ed­u­cates a child is ir­rele­vant to this dis­cus­sion. • Aris’ ex­panded ex­pla­na­tion is ex­cel­lent, and what I would have tried to say at first. I find it pretty im­plau­si­ble that the ed­u­ca­tion of Bri­tish chil­dren in the art­work of an 18th cen­tury Bri­tish land­scape painter is a bet­ter method of sav­ing Afri­can lives than a proven method that cur­rently saves lives and is reck­oned to be one of the cheap­est meth­ods per life saved. Over the long term, how we ed­u­cate chil­dren prob­a­bly de­ter­mines a great deal about what our world looks like in the fu­ture. How­ever, un­less you have an or­a­cle, or are ed­u­cat­ing them in some­thing speci­fi­cally re­lated, such as the con­cept of Effi­cient Char­ity, I would place the up­per and lower guesses of the me­dian in­crease in QALY/​DALY well be­low and above zero, re­spec­tively, in­di­cat­ing that you shouldn’t do it on that ba­sis. • Down­voted for ex­treme amounts of mud­dled think­ing, and a line of ar­gu­men­ta­tion that’s so hole-rid­den it gives me a headache. Also he has an­swered you already: He ar­gued that dis­play­ing the origi­nal Blue Rigi as op­posed to a fac­simile doesn’t con­tribute one iota to the ed­u­ca­tion of any child. You ei­ther didn’t pay at­ten­tion, or are try­ing to wear him out by keep on ask­ing some­thing he already an­swered. • Maybe. But I still don’t know if that’s be­cause art doesn’t con­tribute or be­cause origi­nals are the same as fac­similes. Any­way, can you help me un­der­stand what you con­sider the holes/​mud­dle? • Mud­dled think­ing is when your line of ar­gu­men­ta­tion “paint­ing con­tributes to mu­seum, mu­seum con­tributes to ed­u­ca­tion, ed­u­ca­tion con­tributes to pro­duc­tivity, pro­duc­tivity con­tributes to char­ity” im­plies there’s some sin­gle met­ric each of these in­crease, which can be traced from one to the other sim­ply, step by step. An origi­nal paint­ing may con­tribute to mu­seum’s “qual­ity”, but it needn’t con­tribute to the ed­u­ca­tional qual­ity of the mu­seum, so you can’t trans­fer that sort of con­tri­bu­tion down that next step. An art mu­seum con­tributes to ed­u­ca­tion, but it needn’t con­tribute to ed­u­ca­tion in such a man­ner that it be­comes the sort of “pro­duc­tivity” that saves lives. Art is about aes­thet­ics, which con­tribute to qual­ity of life, but not the preser­va­tion of such. Art con­tributes, but it con­tributes differ­ently—and you were told that already. Ed­u­ca­tion may con­tribute to pro­duc­tivity, but de­pend­ing what you’re ed­u­cated to value, it may in­crease or de­crease the amounts of char­ity pro­vided. For ex­am­ple, if you’re taught to value the pres­ence of origi­nal paint­ings, you’ll prob­a­bly give money to keep origi­nal paint­ings in your na­tion, not to save lives. Want­ing an origi­nal paint­ing, as op­posed to a copy, isn’t about ed­u­cat­ing, it’s about satis­fy­ing a fetish. A na­tional fetish in this case, much the way that Greece was ob­sess­ing with Olympic Games and mu­se­ums to house the un­re­turned Parthenon mar­bles, while in the mean­time its econ­omy was go­ing down the crap­per. In that way I could eas­ily ar­gue that the origi­nal is of less util­ity than a fac­simile, ex­actly be­cause it en­courages such un­pro­duc­tive fetishes, while be­ing aes­thet­i­cally iden­ti­cal. • Upvoted, but dis­agreed with: In that way I could eas­ily ar­gue that the origi­nal is of less util­ity than a fac­simile, ex­actly be­cause it en­courages such un­pro­duc­tive fetishes, while be­ing aes­thet­i­cally iden­ti­cal. It seems to me that scarcity and au­then­tic­ity can both play into aes­thet­ics, but be­sides those two con­tex­tual vari­ables that’s spot on. • I don’t think the prefer­ence for origi­nal paint­ings is just a fetish. Ac­cu­rate color re­pro­duc­tion is hard [1], and in many cases, it’s pos­si­ble to get close enough to the origi­nal to see the brush­strokes and tex­ture. I don’t think we’re at the tech yet for re­ally ex­cel­lent re­pro­duc­tions, but please let me know if I’m miss­ing some­thing. Origi­nals vs. re­pro­duc­tions may not be worth the cost, but that’s a differ­ent ques­tion. [1] The col­ors in a paint­ing may change with time, but re­pro­duc­tions add an­other layer of in­ac­cu­racy. I don’t know how good color re­pro­duc­tion can be if a ma­jor effort is made. I do know that if I go to the mu­seum shop af­ter an ex­hi­bi­tion, I’m always struck by how far off the col­ors are com­pared to the paint­ings. • Tex­ture re­pro­duc­tion is ac­tu­ally an eas­ier prob­lem than color re­pro­duc­tion, and is pretty much solved at less than a$5000 cost. Color is hard par­tially be­cause peo­ple want the paint­ing to look the same un­der all light­ing con­di­tions; un­der just one, we can solve the prob­lem pretty well, but un­der all, we nearly need to use the same ma­te­ri­als as were origi­nally used. Need­less to say, the cost of re­pro­duc­tions scales with the qual­ity, and can be­come quite high.

• I won­der if enough peo­ple would go to a mu­seum of high qual­ity re­pro­duc­tions to make it worth­while.

• ex­cept for aes­thetic reasons

I think that’s the point.

• that child might go on to save other chil­drens lives.

Or, of course, go on to harm them. Or be neu­tral. It seems al­most cer­tain that on av­er­age there is some benefit from the stan­dard trade and com­par­a­tive ad­van­tage rea­sons, but I have no idea how to even ap­proach that calcu­la­tion.

• Will I then ever spend the money on char­ity?

What’s likely to hap­pen is that the RoR and benefit of char­ity will fluc­tu­ate over time and over the size of your pot- so your pot will grow un­til there’s a need, then you’ll spend, and then it’ll go back to grow­ing. The prob­lem is that re­quires ac­tive man­age­ment (which is hard to con­tinue af­ter your death) and typ­i­cally the view is that if you value warm fuzzies, you can find some char­ity that re­turns more than the RoR of prof­itable ven­tures.

There is quite a bit of warm fuzzies in gen­er­at­ing a gi­ant pot of cash and then en­dow­ing it to stand per­pet­u­ally- but be­yond sta­bil­ity effects I’m not sure there is much to recom­mend that model of char­ity.

• In or­der this to be true for­ever, the world would have to never end, which would mean that there’s in­finite util­ity no mat­ter what you do.

If this is false even­tu­ally, there is no para­dox. Whether or not It’s worth while to in­vest for a few cen­turies is an open ques­tion, but if it turns out it is, that’s no rea­son to aban­don the idea of com­par­ing char­i­ties.

• In or­der this to be true for­ever, the world would have to never end, which would mean that there’s in­finite util­ity no mat­ter what you do.

That doesn’t sound right… even if I’m ex­pect­ing an in­finite fu­ture I think I’d still want to live a good ex­is­tence rather than a mediocre one (but with >0 util­ity). So it does mat­ter what I do.

Say I have two op­tions:

• A, which offers on av­er­age 1.. utilon per sec­ond? (Are utilons mea­sures of util­ity of a time pe­riod, or in­stan­ta­neous util­ity?)

• B, which offers on av­er­age 2 utilons /​ s

The limits as t ap­proaches in­finity are U(A) = t, U(B) = 2t. Both are “in­finite” but B is yet larger than A, and there­fore “bet­ter”.

• You can switch be­tween A and B just by re­ar­rang­ing when events hap­pen. For ex­am­ple, imag­ine that there are two planets mov­ing in op­po­site di­rec­tions. One is a Utopia, the other is a Distopia. From the point of refer­ence of the Utopia, time is slowed down in the Distopia, so the world is worth liv­ing in. From the point of refer­ence of the Distopia, it’s re­versed.

This gets even worse when you start deal­ing with ex­pected util­ity. As messed up as the idea is that the or­der of events mat­ter, there at least is an or­der. With ex­pected util­ity, there is no in­her­ent or­der.

The best I can do is as­sign the pri­ors for in­finite util­ity to zero, and make my pri­ors fall off fast enough to make sure ex­pected util­ity always con­verges. I’ve man­aged to prove that my pos­te­ri­ors will also always have a con­verg­ing ex­pected util­ity.

• So we need to for­mal­ize this, ob­vi­ously.

Method 1: Ex­po­nen­tial dis­count­ing.

Prob­lem: You don’t care very much about fu­ture peo­ple.

Method 2: Tak­ing the av­er­age over all time (speci­fi­cally the limit as t goes to in­finity of the in­te­gral of util­ity from 0 to t, di­vided by t)

Con­clu­sion which may be prob­le­matic: If hu­man­ity does not live for­ever, noth­ing we do mat­ters.

Caveat: Depend­ing on our an­throp­ics, we can ar­gue that the uni­verse is in­finite in time or space with prob­a­bil­ity 1, in which case there are an in­finite num­ber of copies of hu­man­ity, and so we can always calcu­late the av­er­age. This seems like the right ap­proach to me. (In gen­eral, us­ing the same math for your ethics and your an­throp­ics has nice con­se­quences, like avoid­ing most ver­sions of Pas­cal’s Mug­ging.)

• Prob­lem: You don’t care very much about fu­ture peo­ple.

Why is this a prob­lem? This seems to match re­al­ity for most peo­ple.

• So does self­ish­ness and ir­ra­tional­ity. We would like to avoid those. It also is in­tu­itive that we would like to care more about fu­ture peo­ple.

• Ex­ces­sive self­ish­ness, sure. Some de­gree of self­ish­ness is re­quired as self-defense, cur­rently, oth­er­wise all your own needs are sub­sumed by sup­ply­ing oth­ers’ wants.. Even in a com­pletely sym­met­ric so­ciety with ev­ery­body act­ing more for oth­ers’ good than their own is worse than one where ev­ery­body takes care of their own needs first—be­cause each in­di­vi­d­ual gen­er­ally knows their own needs and wants bet­ter than any­one else does.

I don’t know the needs and wants of the fu­ture. I can’t know them par­tic­u­larly well. I have worse and worse un­cer­tainty the farther away in time that is. Un­less we’re talk­ing about species-ex­tinc­tion level of events, I damn well should punt to those bet­ter in­formed, those closer to the prob­lems.

It also is in­tu­itive that we would like to care more about fu­ture peo­ple.

Not to me. Heck. I’m not en­tirely sure what it means to care about a per­son who doesn’t ex­ist yet, and where my choices will in­fluence which of many pos­si­ble ver­sions will ex­ist.

• each in­di­vi­d­ual gen­er­ally knows their own needs and wants bet­ter than any­one else does.

I don’t know the needs and wants of the fu­ture.

Ex­pected-util­ity calcu­la­tion already takes that into effect. Uncer­tainty about whether an ac­tion will be benefi­cial trans­lates into a lower ex­pected util­ity. Dis­count­ing, on top of that, is dou­ble count­ing.

Knowl­edge is a fact about prob­a­bil­ities, not util­ities.

Not to me.

Let’s hope our differ­ent in­tu­itions are re­solv­able.

I’m not en­tirely sure what it means to care about a per­son who doesn’t ex­ist yet, and where my choices will in­fluence which of many pos­si­ble ver­sions will ex­ist.

Surely it’s not much more difficult than car­ing about a per­son who your choices will dra­mat­i­cally change?

If you have a set E = {X, Y, Z...} of pos­si­ble ac­tions, A (in E) is the util­ity-max­imis­ing ac­tion iff for all other B in E, the limit

$\\lim\_\{t\\rightarrow\\infty\}\\left\(\\int\_0^t Eu\(A, t'\$dt’%20-%20\int_0%5Et%20{Eu(B,%20t’)dt’%20\right))

is greater than zero, or ap­proaches zero from the pos­i­tive side. Caveat: I have no ev­i­dence this doesn’t im­plode in some way, per­haps by the limit be­ing un­defined. This is just a stupid idea to con­sider. A pos­si­bly equiv­a­lent for­mu­la­tion is

$\\exists e\\forall t\.~\(t>e\$%20\im­plies%20\left(\int_0%5Et%20Eu(A,%20t’)dt’%20\geq%20\int_0%5Et%20Eu(B,%20t’)dt’\right))

The in­equal­ity be­ing greater or equal al­lows for two or more ac­tions be­ing equiv­a­lent, which is un­likely but pos­si­ble.

• Side com­ment: that math equa­tion image gen­er­a­tor you used is freakin’ ex­cel­lent. The image it­self is gen­er­ated based from the URL, so you don’t have to worry about host­ing. Edi­tor is here.

• I pre­fer this one, which au­to­mat­i­cally gen­er­ates the link syn­tax to paste into a LW com­ment. There’s a short dis­cus­sion of all this on the wiki.

• Func­tions whose limit is +in­finity and -in­finity can be dis­t­in­guished, so your good there.

I think it’s the same as my sec­ond: As long as the prob­a­bil­ity given both ac­tions of a hu­man­ity last­ing for­ever is nonzero, and the differ­ences of ex­pected util­ities far in the fu­ture is nonzero, noth­ing that hap­pens in the first mil­lion billion years mat­ters.

• The differ­ence in ex­pected util­ity would have to de­crease slow enough (slower than ex­po­nen­tial?) to not con­verge, not just be nonzero. [Which would be why ex­po­nen­tial dis­count­ing “works”...]

How­ever I would be sur­prised to see many de­ci­sions with that kind of last­ing im­pact. The prob­a­bil­ity of an ac­tion hav­ing some effect at time t in the fu­ture “de­cays ex­po­nen­tially” with t (as­sum­ing p(Effect_t | Effect_{t-1}, Ac­tion) is ap­prox­i­mately con­stant), so the differ­ence in ex­pected util­ity will in gen­eral fall off ex­po­nen­tially and there­fore con­verge any­way. Ex­cep­tions would be choices where the util­ities of the likely effects in­crease in mag­ni­tude (ex­po­nen­tially?) as t in­creases.

Any­way I don’t see in­fini­ties as an in­her­ent prob­lem un­der this scheme. In par­tic­u­lar if we don’t live for­ever, ev­ery­thing we do does in­deed mat­ter. If we do live for­ever, what we do does mat­ter, ex­cepts how it af­fects us might not if we an­ti­ci­pate caus­ing “per­manant” gain by do­ing some­thing.

• Can’t think about the un­der­ly­ing idea right now due to headache, but in­stead of talk­ing about any sort of limit, just say that it’s even­tu­ally pos­i­tive, if that’s what you mean.

• Bostrom would dis­agree with your con­clu­sion that in­fini­ties are un­prob­le­matic for util­i­tar­ian ethics: http://​​www.nick­bostrom.com/​​ethics/​​in­finite.pdf

• The points made here are sound. I was par­tic­u­larly awak­ened by call­ing out the rule about over­head as wrong, since that has been a ma­jor fac­tor in my char­i­tiable giv­ing in the past.

How­ever, if we imag­ine ev­ery­one be­hav­ing ac­cord­ing to these rules, we wind up with very few (in­com­pe­tent) peo­ple run­ning a few char­i­ties with piles of cash. If no lawyers take time off and con­tribute their ex­per­tise to a char­ity, then how do char­i­ties pro­tect them­selves from law­suits, for ex­am­ple? The op­ti­mal char­ity solu­tion is not for ev­ery­one to fol­low your guidelines, but for al­most ev­ery­one to fol­low your guidelines, and a few peo­ple to de­vi­ate. Yet, how do we know whether we should be the ones who de­vi­ate?

• How­ever, if we imag­ine ev­ery­one be­hav­ing ac­cord­ing to these rules, we wind up with very few (in­com­pe­tent) peo­ple run­ning a few char­i­ties with piles of cash.

If the choice is be­tween char­i­ties mak­ing an­ti­malar­ial drugs run by com­pe­tent peo­ple, and char­i­ties mak­ing (more use­ful) mosquito nets run by in­com­pe­tent peo­ple, then yes on the short term you might see in­com­pe­tent peo­ple with loads of cash, but then other char­i­ties will prob­a­bly pop up mak­ing malar­ial nets with low over­head, and then they’ll get the most donar­i­ons.

Or if you’re con­cerned about com­pe­tent peo­ple all get­ting a “real” job and donat­ing money: it’s only ra­tio­nal to do so when the marginal util­ity of vol­un­teer­ing is less than the marginal util­ity of work­ing and donat­ing. If that’s the case now (too many vol­un­teers, not enough money), that doesn’t mean that all vol­un­teers should stop and go get a job.

• If no lawyers take time off and con­tribute their ex­per­tise to a char­ity, then how do char­i­ties pro­tect them­selves from law­suits, for ex­am­ple?

The lawyer ex­am­ple wasn’t about lawyers donat­ing lawyer­ing to char­i­ties, it was about lawyers buy­ing fuzzies by do­ing vol­un­teer work like pick­ing up lit­ter or work­ing at a food bank in­stead of do­ing over­time le­gal work and us­ing the ex­tra money to gen­er­ate ten times as much char­i­ta­ble work.

Un­der some cir­cum­stances, the most effi­cient thing might be for a lawyer to provide pro-bono le­gal work to a char­ity, if a good lawyer is will­ing to do that, but in gen­eral, the an­swer to “how do char­i­ties pro­tect them­selves from law­suits?” is “by pay­ing for le­gal rep­re­sen­ta­tion with part of the money peo­ple donate to them”.

• The an­swer is that un­til the world’s cul­ture of giv­ing changes mas­sively, you should not be the one to de­vi­ate. And you’ll no­tice when the world’s cul­ture of giv­ing is chang­ing mas­sively. And then we can solve the new prob­lem of “but the marginal gain of one lawyer from zero re­ally is large!”, but un­til then, it’s noth­ing more than a hy­po­thet­i­cal.

Un­less! I haven’t thought about this be­fore, but what if the great ma­jor­ity of peo­ple of im­por­tant-to-char­ity cat­e­gory X who cur­rently donate their time are also the sort of peo­ple who will switch to these much bet­ter guidelines be­fore it be­comes a wor­ld­wide phe­nomenon? Does such a cat­e­gory ex­ist? It’s the only thing that would make a “just switch to these guidelines and fix the other prob­lems if these guidelines are widely adopted” policy turn out badly if im­ple­mented, I think...

• this was cov­ered here: http://​​less­wrong.com/​​lw/​​65/​​money_the_unit_of_car­ing/​​

“If the soup kitchen needed a lawyer, and the lawyer donated a large con­tigu­ous high-pri­or­ity block of lawyer­ing, then that sort of vol­un­teer­ing makes sense—that’s the same spe­cial­ized ca­pa­bil­ity the lawyer or­di­nar­ily trades for money. But “vol­un­teer­ing” just one hour of le­gal work, con­stantly de­layed, spread across three weeks in ca­sual min­utes be­tween other jobs? This is not the way some­thing gets done when any­one ac­tu­ally cares about it, or to state it near-equiv­a­lently, when money is in­volved.”

• Very well pre­sented.

Just a minor tech­ni­cal note: All the links that linked to other LW pages are bro­ken. It looks like some­how the links ended up hav­ing those ar­ti­cles’s names be­ing ap­pended to the link for this one.

For in­stance, the one that was sup­posed to link to money be­ing the unit of car­ing in­stead tries to link to this: http://​​less­wrong.com/​​lw/​​3gj/​​effi­cient_char­ity_do_unto_oth­ers/​​lw/​​65/​​money_the_unit_of_car­ing/​​

• Hm, never seen that par­tic­u­lar er­ror be­fore. Thanks and fixed.

• No prob. I’m guess­ing it was some­thing along the lines of those links some­how end­ing up be­ing turned into rel­a­tive links.

• With re­spect to the lawyer ex­am­ple, I un­der­stand that the lawyer can max­i­mize the good he does by re­main­ing a lawyer in­stead of work­ing for a non-profit. But if all the most tal­ented/​pro­duc­tive peo­ple (and thus those with the high­est po­ten­tial salaries in the pri­vate sec­tor) took pri­vate sec­tor jobs, then only the least tal­ented/​pro­duc­tive peo­ple would be available to start and run the non-prof­its. Given that we can ex­pect this low tal­ent pool to make many mis­takes, a lot of the high tal­ent pool’s dona­tions will be squan­dered and wasted. So hav­ing all the $1,000 lawyers, doc­tors, CEO etc… stay in the pri­vate sec­tor may not max­i­mize the to­tal good achieved. In my view we need some of the most tal­ented/​pro­duc­tive to run the non-prof­its. Yes they will make less than they will in their pri­vate sec­tor oc­cu­pa­tion and there­fore will not be able to per­son­ally donate as much. But they will make the over­all char­ity sys­tem more effi­cient. As for the self-in­ter­est piece of this puz­zle the tal­ented non-prof­i­teers would need to have a strong prefer­ence for char­ity so the gains in util­ity from their oc­cu­pa­tion offset the de­crease in pur­chas­ing power from the lower salary. • For the “lawyer work for an­other hour and donate the money vs. vol­un­teer”, it also what mat­ters what the side effects of his la­bor are, right? If the lawyer can make$1000 an hour, but only in ways that ac­tu­ally harm so­ciety (work­ing on frivolous law­suits against hos­pi­tals, for ex­am­ple), then work­ing for an­other hour and donat­ing the money isn’t nec­es­sar­ily the best thing he could be do­ing. Now, on the other hand, if his work also gen­uinely helps so­ciety and cre­ates more wealth for peo­ple, then it’s even bet­ter then the null case.

• For the “lawyer work for an­other hour and donate the money vs. vol­un­teer”, it also what mat­ters what the side effects of his la­bor are, right? If the lawyer can make $1000 an hour, but only in ways that ac­tu­ally harm so­ciety (work­ing on frivolous law­suits against hos­pi­tals, for ex­am­ple), then work­ing for an­other hour and donat­ing the money isn’t nec­es­sar­ily the best thing he could be do­ing. Then the equa­tion be­comes ($1,000 - dam­age-per-hour + damgage-per-hour-if-the-next-lawyer-available-did-it > next-best-value-op­por­tu­nity). It is highly un­likely that the lawyer choos­ing to not in­volve him­self in such law­suits will make much differ­ence at all, at the mar­gin. Lawyer availa­bil­ity isn’t a par­tic­u­larly limit­ing fac­tor.

Par­ti­ci­pat­ing in the so­ciety-de­struc­t­ing be­hav­ior, tak­ing huge amounts of money for do­ing so (and us­ing it well) while still vot­ing for laws that pro­hibit or limit such be­hav­ior in gen­eral is prob­a­bly the cor­rect thing to do.

• It’s pos­si­ble that that might be true, in the­ory, but con­sid­er­ing that we are run­ning on cor­rupted hard­ware, I would be very sus­pi­cious of ac­tu­ally try­ing to put any plan into effect that sounded like an ex­cuse for “I’m go­ing to do some­thing un­eth­i­cal to make my­self re­ally rich, but I’ll make sure I use the money for good, and if I don’t do it some­one else will any­way.” Even if that is ac­tu­ally your plan, in re­al­ity you will prob­a­bly end up do­ing more harm then good.

• I would be very sus­pi­cious of ac­tu­ally try­ing to put any plan into effect that sounded like an ex­cuse for “I’m go­ing to do some­thing un­eth­i­cal to make my­self re­ally rich, but I’ll make sure I use the money for good, and if I don’t do it some­one else will any­way.” Even if that is ac­tu­ally your plan, in re­al­ity you will prob­a­bly end up do­ing more harm then good.

I im­ple­ment eth­i­cal in­hi­bi­tions too. The differ­ence is in a differ­ent un­der­stand­ing of the eco­nomic im­pli­ca­tions of the be­hav­ior. Put­ting “be a lawyer who is will­ing to work on ‘frivolous’ law­suits” in the class of things that must be in­hibited for eth­i­cal rea­sons is a mis­take.

• There’s one flaw in the ar­gu­ment about Buy a Brush­stroke vs Afri­can san­i­ta­tion sys­tems, which is the as­sump­tion/​im­pli­ca­tion that if they hadn’t given that money to Buy a Brush­stroke they would have given it to Afri­can san­i­ta­tion sys­tems in­stead. It’s a false di­chotomy. Sure, the money would have been bet­ter spent on Afri­can san­i­ta­tion sys­tems, but you can say that about any­thing. The money they spent on their cars, the money I just spent on my lunch, in fact some­where prob­a­bly over 99.9% of all non-Afri­can-san­i­ta­tion-sys­tem-pur­chases made in the first-world would be bet­ter to have been made on Afri­can san­i­ta­tion sys­tems. It makes the Buy a Brush­stroke cam­paign look ac­tively mal­i­cious, de­spite the fact that all it re­all did was redi­rect money from per­sonal junk lux­ury items to, well, an­other more pub­lic junk lux­ury item. Neu­tral at worst.

To me, it’s silly to only ap­ply the san­i­ta­tion sytems com­par­i­son to peo­ple’s char­i­ta­ble dona­tions. They’re a softer tar­get, be­cause it’s ob­vi­ous that peo­ple could have spared that money, but the end re­sult is peo­ple who’ve given noth­ing to any­one sit­ting there think­ing “Well at least I’m not that stupid to have made such sub­op­ti­mal dona­tions”, and feel­ing su­pe­rior about them­selves com­pared to those who are at least giv­ing some­thing to a cause that’s not them­selves. Not to men­tion peo­ple feel­ing ac­tively guilty about rais­ing money for a good lo­cal cause just be­cause ev­ery dona­tion they gather is money those peo­ple could have given to a bet­ter cause.

I agree with your point on the whole I just think these side-effects of that com­par­i­son are worth rais­ing.

• This speaks in fa­vor of the moral qual­ities of the cam­era manufacturer

Why is this? Is it un­eth­i­cal to profit from trade? This made my lit­tle in­ner Ob­jec­tivist cringe pretty hard… but oth­er­wise, I like this post a lot. It makes an im­por­tant point about effi­ciency that isn’t ob­vi­ous and needs to be re­in­forced.

• It can speak in favour of the moral qual­ities of the cam­era man­u­fac­turer with­out speak­ing against the moral qual­ities of the parka man­u­fac­turer. Of course it’s fine to profit from trade, but gen­eros­ity is still praise­wor­thy.

(Though, in prac­tice, sel­l­ing a $200 cam­era at no profit is prob­a­bly not nearly the best way for a cam­era man­u­fac­turer to be al­tru­is­tic any­way, so that may not be the best ex­am­ple.) • This is an ex­cel­lent ar­ti­cle. A quick com­ment on one of the sen­tences: GiveWell.org, a site which col­lects and in­ter­prets data on the effec­tive­ness of char­i­ties, pre­dicts that an­ti­malar­ial drugs save one child from malaria per$5,000 worth of bed medicine, but in­sec­ti­cide-treated bed nets save one child from malaria per $500 worth of drugs. I find this sen­tence some­what con­fus­ing. Should the “worth of bed medicine” be just “medicine”, and the “worth of drugs” be “worth of net­ting”? • Thank you, I have no idea what hap­pened to my brain there. • The Pol­li­na­tion Pro­ject is run by a guy who gives$1,000 a day, to a differ­ent re­cip­i­ent ev­ery day. Ra­tional jus­tifi­ca­tions for this ap­proach in­clude min­i­miz­ing the model risk—i.e., per­haps the model you used to de­cide which sin­gle char­i­ta­ble cause is the best is wrong. Also, small dona­tions seem likely to pro­duce a high ve­loc­ity of the money donated.

• In­ter­est­ing ar­ti­cle.

But if you re­ally re­ally want to do more good, you also have to change the way you see peo­ple in need. This might in­volve buy­ing a ticket to Su­dan and see­ing chil­dren face to face while they are starv­ing to death.

So in the long run, your $5k trip to Su­dan might do more good than giv­ing a one time$5k dona­tion to an or­ga­ni­za­tion, be­cause by chang­ing the way you think about starv­ing peo­ple, you will have the ur­gency of ac­tu­ally chang­ing your pri­ori­ties. So you will liter­ally work for those who die of star­va­tion in other coun­tries, and de­prive your­self from buy­ing use­less stuff or en­ter­tain­ment in or­der to feed more chil­dren in Su­dan.

… un­less you’re psy­chopath, of course. ;-)

• It is im­por­tant to be ra­tio­nal about char­ity for the same rea­son it is im­por­tant to be ra­tio­nal about Arc­tic ex­plo­ra­tion: it re­quires the same aware­ness of op­por­tu­nity costs and the same hard-headed com­mit­ment to in­ves­ti­gat­ing effi­cient use of resources

In his Mars Direct talks, Robert Zubrin cited the shoestring bud­get Amund­sen ex­pe­di­tion through the North­west Pas­sage in com­par­i­son to around 30 con­tem­po­rary gov­ern­ment funded ex­pe­di­tions with state of the art steam fri­gates and huge lo­gis­tics trains. The Amund­sen ex­pe­di­tion trav­eled in a cheap lit­tle seal­ing boat and fed them­selves largely through rifles and am­mu­ni­tion they brought with them.

• Con­grat­u­la­tions! I Liked the ar­ti­cle very much.

I just have some doubts about two spe­cific points:

In the part that there’s the text:

“And when your life is on the line, things like im­press­ing your friends and buy­ing or­ganic pale in com­par­i­son.”

I got the im­pres­sion that it’s miss­ing the end of the sen­tence. As my en­glish is not good, maybe it’s my fault. Sorry if that is the case.

The other thing is that I found a lit­tle prob­le­matic the math com­par­ing U$10.000 spent on a U$ 500 char­ity with over­head of 50% and the one spent on a U$10.000 char­ity with 0% over­head. It’s said that in the first case we could save 10 peo­ple and in the sec­ond just 2. Isn’t that 20 peo­ple in the first case? • “And when your life is on the line, things like im­press­ing your friends and buy­ing or­ganic pale in com­par­i­son” I got the im­pres­sion that it’s miss­ing the end of the sen­tence. As my en­glish is not good, maybe it’s my fault. Sorry if that is the case. Parse it as fol­lows: And when your life is on the line, cer­tain other things (like “im­press­ing your friends” and “buy­ing or­ganic foods”) pale in com­par­i­son. • I agree with the idea that effi­ciency should be taken into ac­count when con­sid­er­ing char­i­ta­ble ac­tions, but I do not know if I agree with your con­clu­sion of what is most effi­cient. Alle­vi­at­ing a prob­lem does not cure it. While pay­ing for malaria nets, clean­ing up the beach, donat­ing to char­i­ties alle­vi­ates real so­cial is­sues, it does not ad­dress the is­sue of their cau­sa­tion. In my opinion, what is most effi­cient is not con­cen­trat­ing on re­cu­per­a­tion, but at­tack­ing the sick­ness. Without chang­ing the causal con­di­tions the dis­ease will con­tinue to grow end­lessly no mat­ter how much you sup­press it. This is why even af­ter go­ing through suc­cess­ful re­hab, ad­dicts will ex­pe­rience re­lapse if rein­tro­duced into their origi­nal en­vi­ron­ments be­cause noth­ing has changed to pre­vent the same symp­toms from aris­ing again. What then is the cause of so­cial trav­es­ties? I would ar­gue a lack of high-level em­pa­thy. In my opinion the ques­tion then be­comes does fi­nan­cial dona­tion in­crease a per­son’s em­pa­thetic ca­pac­ity? I do not think it does. It definitely in­crease the amount of pure cap­i­tal be­ing pushed at a prob­lem, but i do not think that nec­es­sar­ily cures the prob­lem. I know that some of the poor­est schools in Amer­ica have re­cently got­ten state of the art equip­ment, smart rooms, i-pads, new schools, but their test scores are not chang­ing. That is be­cause the prob­lem is the val­ues be­ing pushed into the kids not the amount of money. What does pro­mote em­pa­thy? Pierre Bour­dieu is a promi­nent so­ciol­o­gist who is best known for the idea of habi­tus. The gen­eral idea of habi­tus is that cog­ni­tive and emo­tional pat­terns are shaped by hu­man phys­i­cal­ity. Aris­tole’s virtue ethics rep­re­sent the idea that moral­ity is de­vel­oped through habi­tus. Men­cius, the sec­ond most fa­mous con­fu­cian moral­ist, also had no­tions of em­pa­thy be­ing like a mus­cle that must be strength­ened within peo­ple. From this the­o­ret­i­cal frame­work the type of char­ity pro­posed in the es­say above would be in­effi­cient. While I can­not for cer­tain say what ac­tion/​ en­vi­ron­ments cul­ti­vate em­pa­thy/​moral­ity; I think it is a safe bet to say that work­ing to make more money and spend­ing money does not. if that was the case, then the most suc­cess­ful busi­ness men and women would also be the most em­pa­thetic peo­ple. No, it seems more likely that em­pa­thy would be de­vel­oped through com­mit­ting time to peo­ple and places that are not read­ily iden­ti­fied with the ac­tor’s self. Mean­ing that the lawyer tak­ing an hour each day to work on the beach would in­crease his virtue/​em­pa­thetic ca­pac­ity, mak­ing that work­ing more valuable than the thou­sands of dol­lars that he could have earned in that time see­ing as there is no way to buy morals. I am not sure if my idea is cor­rect. If the au­thor of this es­say writes from a fun­da­men­tally eco­nomic frame, I write from a fun­da­men­tally so­ciolog­i­cal/​an­thro­polog­i­cal/​con­fu­cian one. The cor­rect an­swer is prob­a­bly a me­di­a­tion of the two de­pend­ing on type of char­ity and cir­cum­stance. Thanks for your thought­ful writ­ing. • Your com­ment tries to an­swer the ques­tion, “How can I make my­self more char­i­ta­ble?” rather than the ques­tion, “Now that I’m very char­i­ta­ble, how can I max­i­mize my im­pact?” If some­one is not a very char­i­ta­ble per­son, yes some learned em­pa­thy might fix that, and hands-on ex­pe­rience might be the best way to do that. But such a per­son would not be ask­ing the first ques­tion above. If some­one is already a very char­i­ta­ble per­son, then they should be con­cerned with how much of an im­pact their ac­tions are ac­tu­ally hav­ing—then, the work on the beach is in­effi­cient as com­pared to the thou­sands of dol­lars. • Well said, but I would tweak your word­ing of my ques­tion to “now that I am a good per­son, how can max­i­mize my im­pact?” What is the es­ti­mate of a good per­son? I would ar­gue that a good per­son is one who pro­duces mean­ingful re­la­tion­ships in the world. The model of effi­ciency above touches only on how to most im­pact the per­son-cap­tial re­la­tion­ship, i.e what to do with the ma­te­rial and la­bor re­sources I have ac­cu­mu­lated to most pos­i­tively im­pact hu­man­ity. I agree that this is im­por­tant, but add that the “good per­son” is defined by mul­ti­ple re­la­tion­ships, not just of the one they have to cap­i­tal. For ex­am­ple, I would ar­gue a truly good per­son would be a good child, good par­ent, good friend, good older/​younger (de­pend­ing on the age of the op­pos­ing ac­tor), good stranger, good cit­i­zen, good char­ac­ter, and po­ten­tially much more. To max­i­mize the mean­ing and pos­i­tivity of all crit­i­cal re­la­tion­ships is not done through eco­nomic effi­ciency. And while I can­not make any ab­solute claims that the so­cial im­pact a per­son makes is more benefi­cial than the way they use their cap­i­tal, per­son­ally I be­lieve it to be so. Now if your origi­nal state­ment about already be­ing char­i­ta­ble was meant to mean that you are already a very hu­mane per­son (mean­ing your re­la­tional im­pact in your com­mu­nity is max­i­mized) , then sure, I think max­i­miz­ing char­i­ta­ble ac­tion is great. But I think to max­i­mize your role within a so­cial net­work is re­ally hard, if not im­pos­si­ble to some ex­tent. I also think that most peo­ple are not as em­pa­thet­i­cally de­vel­oped as they would like to think. I would go as far to as to say that a perfect em­pa­thetic aware­ness is as un­reach­able as Truth with a cap­i­tal T. I apol­o­gize if I sound ar­gu­men­ta­tive, I just was not sure if my ques­tion was already dealt with in your minds/​blogs and this is a fur­ther point. • That story sounds sus­pi­ciously nice. Given the choice be­tween be­ing a ‘good per­son’ and fos­ter­ing lo­cal re­la­tion­ships of var­i­ous fuzzy im­pacts, or sav­ing the lives of ten thou­sand peo­ple, would you re­ally choose the former over the lat­ter? Do you think that ac­tu­ally makes you a good per­son? Note that this is not merely a hy­po­thet­i­cal; that is effec­tively the choice the lawyer is mak­ing when he works in soup kitchen in­stead of donat­ing money. • Well given the way you word it, yes, it does seem sus­pi­cious. There are sev­eral things I would change about your retel­ling of my po­si­tion. 1.) I ad­vo­cate for proper and effi­cient re­la­tion­ships. This idea is lo­cal if you mean think­ing of me­chan­i­cal soli­dar­ity be­fore or­ganic soli­dar­ity, but in this day in age with telecom­mu­ni­ca­tion and a globally mo­bile work­force I would not call re­la­tion­ship cul­ti­va­tion “lo­cal” in the tra­di­tional sense. For ex­am­ple, my self-net­work spans mul­ti­ple con­ti­nents. The po­ten­tial for im­pact is huge. 2.) Proper re­la­tion­ships are by no means “fuzzy,” I would say that the fact that you would de­scribe re­la­tion­ship cul­ti­va­tion as fuzzy shows a se­ri­ous lack of men­tal effort. Since it is some­thing I think about a lot, I will give you an ex­am­ple. First let me say I am cur­rently try­ing to define all core re­la­tion­ships of the so­cial self. The so­cial self is the idea that hu­man iden­tity, mo­ti­va­tion, ac­tion, cog­ni­tion, do not arise from au­tonomous agents, but from, a net­work of hu­man, non-hu­man, and cul­tural re­la­tion­ships. One such re­la­tion­ship is the re­la­tion­ship be­tween child and par­ent/​ child and guardian. It is pos­si­ble to not have par­ents, or to not have a guardian, but it is not pos­si­ble to avoid the con­se­quences of this fact. The dy­nam­ics of the child to par­ent/​ guardian re­la­tion­ship is fun­da­men­tal to a per­son’s ac­tions, thoughts, and feel­ings. If my mom or dad were to die, no mat­ter how happy, satis­fied, com­plete I felt im­me­di­ately prior to this, it would com­pletely re­ar­range my feel­ings and thoughts. I would even­tu­ally re­cover, but I would be a differ­ent per­son, one who had to figure out how to be happy, satis­fied and com­plete know­ing my mother was not al­ive. So far I have been try­ing to show the im­pact of a core re­la­tion­ship. The point I origi­nally wanted to make was that cul­ti­vat­ing re­la­tion­ships is not “fuzzy.” Frankly speak­ing it is hard be­ing a good son. If your par­ents are racist, re­li­gious zealots, un­healthy, in­se­cure, it is not your job to fix that. You think it is your job, be­cause your par­ents raised you, fixed you in a sense, and at some point to val­i­date your own ma­tu­rity you want to do the same. And hon­estly in a perfect world you should be able to. I have far more ed­u­ca­tion than my par­ents about health, psy­cholog­i­cally, and so­cial­ity. I am pos­i­tive that if I know what my par­ents are do­ing wrong in cer­tain as­pects of their life, and that I could do bet­ter. There is noth­ing wrong with tel­ling your par­ents you think they should change in some way; the prob­lem arises when they do not want to. You can­not force your par­ents to change. You can cut them out of your life, but that is de­stroy­ing a re­la­tion­ship not cul­ti­vat­ing it. Now I am not talk­ing about ex­tremes here. There might be some cases where they choice comes be­tween those two op­tions, but the ma­jor­ity of the time it is not. The ma­jor­ity of the time, the choice is to ei­ther ac­cept your par­ents for their im­perfec­tion, ig­nore it, or aban­don them. The proper choice be­ing the former. It is a hard thing to do. Proper re­la­tion­ships are not fuzzy. If a re­la­tion­ship is fuzzy all the time, gen­er­ally you are not main­tain­ing it well. 3.) I see cul­ti­vat­ing good peo­ple as mak­ing trans­for­ma­tion change. Mean­ing that it is a trans­form­ing change that does not just stop at ini­tial im­pact. It is per­pet­ual. If you model proper re­la­tions in your so­cial net­work, then the net­works con­nected periph­er­ally will be im­pacted. In the short run pour­ing money on the prob­lem might help, but I do not see this as a solu­tion. A perfect ex­am­ple of this is Aris­to­tle’s ap­peal for the need of prac­ti­cal wis­dom to com­ple­ment laws. You can make laws to reg­u­late, but if peo­ple do not have an in­ter­nal com­mit­ment to the spirit be­hind the laws then the laws will be­come per­pet­u­ally less effec­tive. How many thou­sands of pages of new laws does the United States pro­duce each year? The byproduct of which is that nor­mal peo­ple can no longer un­der­stand the law be­cause it has be­come so com­plex. If nor­mal peo­ple can­not un­der­stand it the re­sult is two-fold. The masses do not in­ter­nal­ize it, and the elite figure out how to take ad­van­tage of it. I would ar­gue this prob­lem of defi­cient prac­ti­cal wis­dom is di­rectly re­lated to a lack of proper re­la­tion­ships and knowl­edge of how to cul­ti­vate them. 4.) I do not think you can save 10,000 peo­ple with any one ac­tion. Nor do I think just be­cause your in­ten­tion is to save peo­ple that is what you ac­tu­ally do. If you get 10,000 peo­ple malaria nets that does not save them from a. be­ing able to get malaria, b.) liv­ing in an en­vi­ron­ment where malaria is preva­lent, c.) the poor con­di­tion of their lives, d.) be­ing able to sus­tain their lineage for mul­ti­ple gen­er­a­tions. Dam­bisa Moyo has a book called “Dead Aid” the ar­gu­ment is that the mil­lions of dol­lars in aid sent to africa is ac­tu­ally do­ing more dam­age than good. There are sev­eral rea­sons for this, if you are in­ter­ested in hear­ing them I would be happy to share. • You are equiv­o­cat­ing on the word fuzzy. There’s a con­trast be­tween do­ing some­thing be­cause it feels good and do­ing some­thing be­cause it ac­tu­ally helps oth­ers. Con­trast serv­ing food at a soup kitchen on Thanks­giv­ing, which makes one feel good vs. serv­ing food on some ran­dom day in June, which is prob­a­bly more helpful to the soup kitchen. The first act pro­vides “fuzzy.” The sec­ond pro­vides more so­cial util­ity. None of this as­serts that main­tain­ing re­la­tion­ships is not valuable or real. The ar­gu­ment is that trans­for­ma­tional re­la­tion­ships have less pay­out per effort than other so­cial im­prove­ment acts (like donat­ing lots of money). And one point of anti-Aid groups is that in­ter­na­tional donors are so con­sumed with “trendy” types of aid that they crowd out both Afri­can self-im­prove­ment and for­eign aid that might help. For more on Dead Aid in par­tic­u­lar, you might find this de­vel­op­men­tal economist’s take in­ter­est­ing. • ″ The ar­gu­ment is that trans­for­ma­tional re­la­tion­ships have less pay­out per effort than other so­cial im­prove­ment acts (like donat­ing lots of money).” I re­al­ize this is the ar­gu­ment, it is what I am dis­agree­ing with. • I apol­o­gize if I am re­hash­ing some­body else’s post; I find that skim­ming the com­ments and then putting my own .02 in is a more valuable use of my time than thor­oughly read­ing the com­ments (and thus al­lo­cat­ing less time to an English pa­per I have com­ing up) and try­ing to sound like I’d ex­haus­tively re­searched the topic (which would take way too much time). The pay­off in terms of lives saved per work unit ex­pended (ei­ther di­rectly through vol­un­teer­ing or in­di­rectly via donat­ing money) varies from per­son to per­son. Even among those who con­sider them­selves ra­tio­nal­ists, there may be vari­a­tions in which char­ity is most “effi­cient”. For ex­am­ple, if one is in­ca­pable of be­com­ing a high-pow­ered lawyer for some rea­son, one may well have a differ­ent pay­off ma­trix in terms of “fuzzies” and ways to go about donat­ing. In ad­di­tion, a high-pow­ered lawyer who quits his$1000-per-hour job to work at a non­profit may in­spire oth­ers to donate to said non­profit, which might in­crease the amount of lives saved. Per­son­ally, I am not con­cerned with world-op­ti­miza­tion; in my opinion, perfec­tion is unattain­able in any dis­ci­pline (Godel un­de­cid­ablilty gen­er­al­ized) and, as such, we should be con­cerned with im­prove­ment; any im­prove­ment over the baseline is “good” and should be ac­cepted. My goal is not to leave the world as close to perfec­tion as I can; my goal is to max­i­mize my hap­piness. As such, I donate to char­i­ties that al­ign with my be­liefs and vol­un­teer at places I en­joy vol­un­teer­ing at. This may not be strictly ra­tio­nal, but a strict ra­tio­nal­ist is much like a work­ing com­mu­nis­tic gov­ern­ment: only at­tain­able in fic­tion. I apol­o­gize if I have offended any­one; I am rel­a­tively new to LW and have much to learn about the com­mu­nity and ra­tio­nal­ity.

• Nice post. You could write a similar one on helping the en­vi­ron­ment. How of­ten do you hear peo­ple say, about helping the en­vi­ron­ment, that “ev­ery lit­tle bit helps”?

• Good guide, in­deed hav­ing more money to spend through what­ever ca­reer may al­low for be­ing more use­ful for char­ity.

The ex­pe­di­tion anal­ogy is good. I’ll get into dis­cussing the spe­cific goal or util­ity func­tion. What is the goal we’re head­ing to?

I’d say the goal as I see it is to in­crease the in­tel­li­gence (or cure the lack of it) to make the agents of this world able to will­ingly solve their prob­lems, and thereby reach a state of tech­nolog­i­cal ad­vance­ment that al­lows them to get rid of all prob­lems for good, and start do­ing bet­ter things such as spend­ing time in par­adise and ex­plor­ing the uni­verse.

We shouldn’t med­i­cate our prob­lems in the short-term, we should think in the long-term in cur­ing them for good. How to do that? Scien­tific re­search into in­tel­li­gence, ar­tifi­cial in­tel­li­gence and hu­man in­tel­li­gence aug­men­ta­tion.

How does “sav­ing” (should I say, pro­long­ing?) Afri­can lives help with that? Not at all, in my view. Africa re­ceives many billions of dol­lars in dona­tions, there’s clearly some­thing wrong with the way it works, and you’re not go­ing to fix it by adding a mil­lion dol­lars to that sea of re­sources that doesn’t end up chang­ing any­thing in the long-term. It’s like a car that leaks fuel, you can keep adding more and more fuel, or you should try and fix it, and that is what I sug­gest. You should rather spend a mil­lion dol­lars in a vi­tal area that is badly lack­ing fund­ing, such as in­tel­li­gence aug­men­ta­tion and ar­tifi­cial in­tel­li­gence.

I don’t think that we want to “save lives”. Save suffer­ing in­stead. If you pro­long an Afri­can life you’re prob­a­bly pro­long­ing suffer­ing, which is a waste. A life of suffer­ing and mis­ery is not worth sav­ing. Peo­ple have no souls. This is a phys­i­cal world, if you lose con­scious­ness some­where you still got plenty of it all around.

• I think you’re falsely as­sum­ing that “Africa” is a sin­gle mono­lithic re­cip­i­ent for that “sea of re­sources”—that ig­nores both the spec­tac­u­lar vari­a­tion be­tween and within Afri­can na­tions, and the differ­ence be­tween re­sources given to a cor­rupt gov­ern­ment aand re­sources ap­plied by non-gov­ern­ment or­gani­sa­tions for the benefit of peo­ple there.

I think it is fair to say that the stag­ger­ing sums of money given by Western na­tions to Afri­can gov­ern­ments has been at best a com­plete waste of money—in fact I con­sider that money to have caused sig­nifi­cant net harm. It props up cor­rupt regimes, in­creases and strength­ens class differ­ences, and gen­er­ally re­sults in in­creased op­pres­sion and wide­spread mis­ery of var­i­ous kinds. Your ar­gu­ment ap­plies very well to this—“Africa” does in­deed re­ceive billions of dol­lars, and there is in­deed some­thing bro­ken (most of the gov­ern­ments re­ceiv­ing the money).

This ar­gu­ment does not ap­ply to the in­ter­na­tional NGO’s work­ing in Africa. Some of those or­gani­sa­tions are short-term ori­ented and thus ar­guably pointless in the long term, but some are not. A clas­sic ex­am­ple would be Kiva, which offers micro-loans for peo­ple to start small busi­nesses to sup­port them­selves and fam­ily (in­ci­den­tally not just in Africa) - there are a fair few or­gani­sa­tions do­ing things like this, and it is “teach a man to fish” rather than “give a man fish”. Th­ese ini­ti­a­tives, when they work right (which they of­ten do) help lift Afri­cans out of poverty and put them in a po­si­tion to do some­thing about their own fu­ture (and Africa’s fu­ture). A lot of worth­while ini­ti­a­tives cen­tre around ed­u­ca­tion, for in­stance, for fairly ob­vi­ous rea­sons.

I think you’re con­flat­ing “in­tel­li­gence” with other con­cepts such as ed­u­ca­tion and good judge­ment (which are what’s ac­tu­ally needed here). Rephrased like that, it be­comes ob­vi­ous that a much more prac­ti­cal ac­tion is to fund and or­ganise ed­u­ca­tion of Afri­can peo­ple—give them the means with which to figure out the solu­tions to their own prob­lems, but now rather than post-Sin­gu­lar­ity. Add di­rect fi­nan­cial sup­port (eg. by Kiva or Grameen etc) in or­der that these now-ed­u­cated peo­ple have the means to im­ple­ment their ideas, and we have to­mor­row’s solu­tion to­day. This is cur­rently hap­pen­ing, but all we tend to know about Africa’s cur­rent situ­a­tion is an as­sort­ment of dra­matic bad news merged to­gether into a highly mis­lead­ing nar­ra­tive. To give you some idea of how sig­nifi­cantly our per­cep­tions differ from re­al­ity on this mat­ter, here’s a TED talk from from the in­com­pa­rable statis­ti­cian Hans Rosling 4.5 years ago: http://​​www.youtube.com/​​watch?v=RUwS1uAdUcI—feel free to poke around for more re­cent pre­sen­ta­tions and data, of course, but even this old one is an eye-opener.

I’m not say­ing that in­vest­ment in ed­u­ca­tion and en­trepreneur­ship in Africa is nec­es­sar­ily the most effec­tive use of re­sources from a strictly util­i­tar­ian stand­point—what I am say­ing is that you have not pre­sented a strong case for Afri­can in­vest­ment not be­ing a worth­while use of re­sources. Per­son­ally I re­gard your ar­gu­ment largely as an ex­cuse to not feel guilty about dis­tant suffer­ing, but that is just an un­sup­ported opinion.

• Ut­ter crap.

I’m read­ing this much later than it was writ­ten, but feel I must re­spond.

1. “Africa re­ceives many billions of dol­lars in dona­tions, there’s clearly some­thing wrong with the way it works, and you’re not go­ing to fix it by adding a mil­lion dol­lars...” Even if it were 100% true that there’s some­thing wrong with the way it works, it DOES NOT FOLLOW that the an­swer is to give less (or none). It may be the case that we have not given enough.

2. “It’s like a car that leaks fuel, you can keep adding more and more fuel, or you should try and fix it...” Even if the anal­ogy were 100% true and ap­pli­ca­ble, you offer a false di­chotomy (add fuel OR fix it). It may be that we need to work on both si­mul­ta­neously. To fur­ther the anal­ogy, it may be that you have to add fuel to the car in the short term (know­ing some will be lost), in or­der to keep it run­ning long enough to get it to the me­chanic who can do the long term fix.

3. “If you pro­long an Afri­can life you’re prob­a­bly pro­long­ing suffer­ing, which is a waste. A life of suffer­ing and mis­ery is not worth sav­ing.” This is at best, hor­ribly wrong-headed, and at worst, dis­gust­ingly elitist. I sus­pect it’s a troll. Only a fool thinks a life that con­tains suffer­ing and mis­ery is not worth sav­ing. Some peo­ple in­deed are suffer­ing so much that they may pre­fer death. But that is THEIR choice, not yours or mine. If you’re re­ally se­ri­ous about triage, and how best to spend one’s money to re­lieve suffer­ing, a lit­tle hu­mil­ity and com­pas­sion (qual­ities I see very lit­tle of in this com­ment) might go a long way to­ward achiev­ing hu­mane (and yet ra­tio­nal) solu­tions. I sus­pect JonatasMuel­ler be­longs to a set of peo­ple about whom no one will ever have to ask these hard ques­tions, and I sus­pect his an­swer is in­fluenced by that fact. And: there’s a sub­tle mis­take here: the idea that re­duc­ing suffer­ing does not in­clude pre­serv­ing or pro­long­ing life. Says who?

4. ″...the goal as I see it is to in­crease the in­tel­li­gence (or cure the lack of it) to make the agents of this world able to will­ingly solve their prob­lems, and thereby reach a state of tech­nolog­i­cal ad­vance­ment that al­lows them to get rid of all prob­lems for good.” That’s a crap goal, if it crowds out other wor­thy goals. I be­lieve we’re already as smart as we need to be, so any efforts to try to in­crease in­tel­li­gence are a waste of re­sources. I agree ad­vanc­ing tech­nol­ogy is im­por­tant, but I be­lieve shar­ing what we have is much much more im­por­tant. Our prob­lems are mainly poli­ti­cal, not tech­nolog­i­cal. What good will it do, it tech­nol­ogy im­proves the lives of some, but the fruits of that tech­nol­ogy are not shared?

• I be­lieve we’re already as smart as we need to be, so any efforts to try to in­crease in­tel­li­gence are a waste of re­sources.

You’ll find that this is a meme many on LW dis­agree with profusely.

• Look how many sci­en­tific ad­vance­ments that we have suc­cess­fully made in the past 100 years, can you hon­estly tell me that you don’t think we can con­tinue to in­crease our sci­en­tific method fast enough to solve in­tel­li­gence as well?

If you do, then please read some of the se­quences that fo­cus on this topic and try to un­der­stand it bet­ter. Poli­tics is not a field that will de­stroy our world if not treated cor­rectly, hu­man­ity will sur­vive and learn from our poli­ti­cal mis­takes even­tu­ally but its not an easy prob­lem to solve quickly. How­ever, not solv­ing the crisis of uFAI is the great­est cur­rent ex­is­ten­tial risk to hu­man­ity ac­cord­ing to my val­u­a­tion with a con­fi­dence fac­tor of 90%.

I don’t dis­agree with donat­ing to Africa(via givewell.org) i just think that if you en­joy donat­ing to Africa you should donate and equal amount to SIAI where they will put it to a bet­ter use in terms of the long term prob­lems af­fect­ing them. In fact when Eliezer finished the se­quences he said that what­ever you feel like do­ing, go out­side and do it. If donat­ing to Africa makes you happy then do that and spend more time earn­ing money to do that. How­ever if you read more from LW you might de­cide that there is a faster solu­tion to the prob­lem in Africa :) and you might help us achieve it.

• I fi­nally re­mem­bered to post this here

Good timing, though: now this is fresh in our minds dur­ing the challenge.

• Would any­one of av­er­age in­tel­li­gence who wants to do as much good as pos­si­ble fail fla­grantly in giv­ing where it max­i­mizes welfare? Why then the dras­tic differ­ence be­tween per­sonal buy­ing and dona­tion? Some­one plan­ning an Arc­tic trip will not make the con­tem­plated mis­takes, which is why you used per­sonal buy­ing to con­trast with dona­tion prac­tices. Whether the de­sire to benefit mankind is pow­ered by warm fuzzies or some other ex­pres­sion of al­tru­is­tic mo­ti­va­tion, if con­trib­u­tors were hu­man-welfare max­i­miz­ers they would do a lot bet­ter at max­i­miz­ing welfare. Peo­ple aren’t that ir­ra­tional!

Dona­tions have lit­tle to do with welfare max­i­miza­tion and warm fuzzies. They pri­mar­ily func­tion as a sig­nal­ing de­vice. Isn’t that ob­vi­ous?

• Isn’t that ob­vi­ous?

Yes. It should also be ob­vi­ous that it suc­ceeds be­cause “ap­pear­ing to de­sire welfare max­i­miza­tion and warm fuzzies” is the sig­nal. If you donated to a char­ity that burned your money to re­duce in­fla­tion your sig­nal would fail.

It is less ob­vi­ous but hope­fully in­ferred by many that one of the in­ten­tions of this post and posts like it (and one of the in­ten­tions of char­ity-rat­ing ini­ti­a­tives) is to break con­ven­tional giv­ing, turn it into a failed sig­nal, and re­place it with effi­cient giv­ing. That way, all the sig­nal­ling donors will con­tinue to donate as they did be­fore, sig­nal­ling the same things, achiev­ing the same sta­tus gain, and ac­ci­den­tally helping the world more.

• How does try­ing to pre­vent wars and/​or stop them rate as char­i­ta­ble ac­tivity? On the one hand, wars are tremen­dously de­struc­tive and on the other, it’s hard to be sure how effec­tive op­pos­ing a war is.

• I am not sure. One of my nodes for “char­i­ta­ble ac­tivity” is non-profit or­gani­sa­tions work­ing in the area in ques­tion, so I hadn’t even con­sid­ered pre­ven­tion/​stop­ping of wars as char­ity. Some very light re­search sug­gests that peace ac­tivism has had a mea­surable im­pact, al­though by im­part­ing the kind of pres­sures I think a lobby group would have more suc­cess ap­ply­ing.

This sug­gests to me some sort of lobby group should be formed, and made pow­er­ful through char­i­ta­ble dona­tions. Stop­ping the Iraq war, to work with an ex­am­ple, would have saved about 110,000 lives; if it cost a lobby group around 20 mil­lion dol­lars to achieve the non-in­va­sion of Iraq then you’re look­ing at 180 dol­lars a life—for com­par­i­son, VillageReach gets about $200 a life and Stop TB Part­ner­ship gets about 150-170 a life. The non-dis­rup­tion of cit­i­zens’ lives should beat the im­prove­ment in cit­i­zens’ lives when char­i­ties be­gin op­er­at­ing in a town; I don’t know how to treat it more rigor­ously though. My in­tu­ition is that for­mal­is­ing the dis­abil­ity-ad­justed life years lost due to in­va­sion and oc­cu­pa­tion should blow all the other char­i­ties out of the wa­ter (at 20 mil­lion). Could a lobby group stop a war with 20 mil­lion dol­lars? Maybe. The num­bers are in the same bal­l­park. Lob­by­ing against war might have some bet­ter lev­er­age with timing (they might no need con­stant pres­sure against, they might just need to can­cel out the high­est points of pres­sure for), but de­clared lob­by­ing might vastly un­der­rep­re­sent ac­tual lob­by­ing. Again, I don’t know enough to treat this pos­si­bil­ity prop­erly. So it rates as pos­si­bly up there with the very best of char­i­ties, definitely worth in­ves­ti­gat­ing, but not a very warm or fuzzy cause at all. Edit: GiveWell’s recom­men­da­tions on cost-effec­tive­ness give the up­per limits of$100 per DALY pre­vented and $1000 per life sig­nifi­cantly changed; with­out DALYs for war (I can find no figures) this gives us at most 110 mil­lion lobby dol­lars be­fore GiveWell would stop recom­mend­ing the lobby group. That much money seems like it could have made a sig­nifi­cant try at op­pos­ing the Iraq war, to be hon­est. • Thanks. I think stop­ping wars is a warm and fuzzy cause, but per­haps peace ac­tivists are apt to po­si­tion them­selves as out­siders and this has made it less likely for them to cre­ate a for­mal lobby. On the other hand, demon­stra­tions may have be­come less effec­tive, which may make a lobby more plau­si­ble. Not sure how you’d add this to the figures, but vet­er­ans make up 1 in 4 of the home­less and Home­less adults have an age-ad­justed mor­tal­ity rate nearly 4 times that of the gen­eral pop­u­la­tion their av­er­age life span is shorter than 45 years. I don’t have stats for the effects on chil­dren of hav­ing one or both par­ents with PTSD, but I ex­pect it to be se­ri­ous. • Hmm. Watch out, more num­bers. Try­ing to take into ac­count those and other fac­tors of war be­sides be­ing shot in bat­tle led me to this statis­ti­cal es­ti­mate of the effect of the Iraq War. The low­est bound was 400,000, and treat­ing ex­cess mor­tal­ity as hap­pen­ing on the av­er­age be­tween 16 and 67 (an overtly op­ti­mistic hope that peo­ple un­der 16 re­main the least af­fected-by-war group), we get 10.2 mil­lion ex­pected life years lost. As­sume that for ev­ery per­son kil­led, there is one per­son dis­abled in some way (a quick check sug­gested 3:1 se­ri­ous in­jury:death ra­tio for sol­diers and 1.8:1 for civili­ans, I went with 1:1 be­cause dis­abil­ity doesn’t always fol­low from se­ri­ous in­jury, but con­cerns like PTSD can bring the ra­tio up to equal) and the figure is just above 20 mil­lion DALYs lost to the Iraq war. GiveWell’s figure of$100 per DALY be­ing the up­per bound of effi­cient char­ity means that a lobby group or char­ity that could have stopped the Iraq War given 2 billion dol­lars would have been a gold-stan­dard effi­cient char­ity.

I be­lieve this an­swers your query, Nancy. Stop­ping wars most definitely rates as some of the best char­ity around.

• I’ve been amazed for a long time how much peo­ple don’t add up the costs of war.

• There’s an­other level, I think. Afaik, wars are less likely be­tween democ­ra­cies than for other per­mu­ta­tions of gov­ern­ment types, so char­i­ties which spread democ­racy might also be a good choice, de­pend­ing on one’s es­ti­mate of their effec­tive­ness.

• In­deed. To the ex­tent that a char­ity spreads democ­racy, and to the ex­tent that de­moc­ra­ti­za­tion re­duces like­li­hood of war, these democ­racy char­i­ties are stop-war char­i­ties.

• “Ex­cess mor­tal­ity” is a difficult con­cept. Most es­ti­mates I’ve seen calcu­late ex­cess mor­tal­ity based on a pre-hos­tilities baseline be­cause this is rel­a­tively easy to calcu­late and pro­duces the high­est pos­si­ble figure for ex­cess mor­tal­ity. But the num­ber we re­ally want would com­pare that mor­tal­ity to the ex­pected post-war mor­tal­ity. In the case of Iraq, this would provide a higher ex­pected value than the rate im­me­di­ately pre-war.

• I aimed for the lower bound. If you go by strictly what has been con­firmed then some­thing like 400 mil­lion dol­lars is the effi­ciency cut-off.

• This ar­gu­ment is based on com­pletely ig­nor­ing fu­ture costs and benefit anal­y­sis and the available al­ter­na­tives. To ac­cept this as a (im­plicit?) ax­iom seems un­nat­u­ral. Imag­ine a pow­er­ful lobby group stopped Amer­i­can in­volve­ment in the Korean war and all of South Korea ended up like the North. Imag­ine NATO did not strike Ser­bia and Milo­se­vic con­tinued to reign. Even the Iraq war did have some pos­i­tive effect—Hus­sein was evil, and po­ten­tially the new gov­ern­ment in Iraq would lead to less suffer­ing, both in­ter­nally and be­cause of other—lo­cal and global—con­flicts avoided. The par­tic­u­lars of these ar­gu­ments are de­bat­able (the Iraq gov­ern­ment may col­lapse into chaos; even if it does not, we will never know the ul­ti­mate costs of keep­ing or not keep­ing Sad­dam in power), but the larger point stands. Other com­ments men­tioned pro­mot­ing democ­racy as a means of pro­mot­ing peace. War can be a rad­i­cal mean of pro­mot­ing democ­racy (at least in Ser­bia it seems to have worked), and this should not be ig­nored.

• Even the Iraq war did have some pos­i­tive effect—Hus­sein was evil, and po­ten­tially the new gov­ern­ment in Iraq would lead to less suffer­ing, both in­ter­nally and be­cause of other—lo­cal and global—con­flicts avoided.

If you take the af­ter-in­va­sion Iraq gov­ern­ment and sub­tract that from Hus­sein’s gov­ern­ing, you get the im­prove­ment in gov­er­nance. Is that im­prove­ment go­ing to save 400,000 lives in its re­duc­tion of lo­cal and global con­flicts? Please keep in mind it is not a di­chotomy of “In­vade and fix Iraq XOR aban­don coun­tries to the whims of evil dic­ta­tors”. It is closer to “Im­prov­ing Iraq: Mili­tary in­ter­ven­tion, or other means?”.

As­sas­si­nat­ing Hus­sein and back­ing a demo­cratic coup is a much bet­ter way of rad­i­cally pro­mot­ing democ­racy in Iraq. I ac­cuse you of com­pletely ig­nor­ing available al­ter­na­tives.

• Good, now we are talk­ing.

Shall I con­tribute to char­i­ties pro­mot­ing as­sas­si­na­tions of evil for­eign lead­ers (there are still a few left) and back­ing demo­cratic coups in­stead of the blan­ket pro-peace move­ments?

• Char­i­ties aren’t well-suited to rad­i­cal poli­ti­cal tasks. You would be bet­ter off pur­su­ing a ca­reer in in­ter­na­tional diplo­macy or states­man­ship, fo­cus­ing on net­work­ing with es­pi­onage rather than the mil­i­tary-in­dus­trial com­plex, if you wish to achieve these kinds of changes.

• It doesn’t have to be all one thing or the other.

• http://​​jackt123.blogspot.com/​​2012/​​02/​​effi­cient-char­ity-or-is-that-oxy­moron.html

The essence of this ar­gu­ment is that char­i­ties which save hu­man lives are more valuable than char­i­ties which don’t. Fur­ther­more, it as­sumes that char­i­ties which save the most lives per dol­lar donated are the most wor­thy char­i­ties and the most de­serv­ing of our con­tri­bu­tions.

How valid is this philos­o­phy? Is sav­ing hu­man lives re­ally the most cru­cial goal to­wards which we all should ad­vance? If we ex­plore the au­thor’s con­tention that the frivolous waste of money on a paint­ing would have been bet­ter spent in the sav­ing of one-thou­sand lives in Africa, then we must also ex­plore the re­sults of this bi­nary de­ci­sion made by the con­tri­bu­tion giver.

The thou­sand saved Afri­cans would al­most cer­tainly agree that they are the most de­serv­ing of the char­i­ta­ble con­tri­bu­tion. This how­ever ig­nores an un­for­tu­nate fact. The amount of life-sus­tain­ing re­sources in Africa―food, medicine, shelter, clean wa­ter―are sadly in­suffi­cient to sus­tain the lives of the Afri­cans already liv­ing on this des­per­ately poor con­ti­nent. By fund­ing a non-Afri­can out­side force to travel thou­sands of miles car­ry­ing the cargo and re­sources and man-power nec­es­sary to save all these thou­sands of lives makes the dona­tor in large part re­spon­si­ble for putting fur­ther strain on an already un­sus­tain­able econ­omy. The idea that these thou­sand “saved” Afri­cans will now live hap­pily-ever-af­ter in peace and har­mony along with the other six-hun­dred mil­lion starv­ing Afri­cans is so ab­surd that it liter­ally bog­gles the mind.

That money spent on san­i­ta­tion and sewage treat­ment could have pur­chased per­haps one-mil­lion con­doms, which would have de­creased the birth-rate and al­lowed those hun­dreds of mil­lions of Afri­cans to en­dure for per­haps an­other gen­er­a­tion. But here again the logic trips us up. For per­haps in those one-mil­lion ba­bies never born, was the Afri­can leader who would have led the Afri­can peo­ple out of dark­ness and sav­agery into a bet­ter more com­pas­sion­ate fu­ture. And who’s to say to what use those con­doms would have gone? Per­haps all the wise Afri­cans would have used the con­doms while all the fool­ish Afri­cans would have con­tinued mak­ing ba­bies. Your big-hearted and thought­ful gift could wind up turn­ing a hellish bru­tal land into some­thing so much worse that there’s no way it could even be imag­ined.

In­deed even among the origi­nal thou­sand saved lives might be a great one with des­tiny rid­ing his shoulders. So per­haps sav­ing the most lives pos­si­ble is the best an­swer. Un­less… what if among those thou­sand saved lives is an­other Hitler or Stalin or Pol Pot! No, the only thing I’m cer­tain of is that try­ing to pre­dict what the fu­ture holds based on a spur of the mo­ment im­pul­sive char­i­ta­ble gift is enough to drive a per­son to drink...

• Do you have a point, or are you just throw­ing up a smoke­screen? Africa has prob­lems, sure, but not dy­ing of malaria is still a good thing (in ex­pec­ta­tion.) Malthus has not been proven right yet, and in­deed Africa does usu­ally have enough food (famines are rare, and usu­ally the re­sult of a com­pletely fscked up gov­ern­ment. Effi­cient aid does con­sider the gov­ern­ment.)

• A good ar­ti­cle, if your goal is to save as many lives as pos­si­ble from per­ish­ing. But I’m go­ing to say, for most peo­ple, this is not their goal. Yes, if you ask some­one di­rectly “would you save a paint­ing, or save 1000 lives”, they would al­most all say “lives of course”. But in re­al­ity, peo­ple don’t have an emo­tional at­tach­ment to 1000 peo­ple they have no idea about.

In my case, I re­ally don’t care if 1000 lives are lost if I don’t do some­thing. I know that makes me sound like a bad per­son. But what are peo­ple re­ally? We’re a self-repli­cat­ing gene ma­chines. There’s 6 billion of us in the world. There is in­her­ently very lit­tle value in sav­ing 1000 lives out of 6 billion. It’s like say­ing “if you give us money we can pre­vent a loon from de­stroy­ing 1000 iPads just be­cause he feels like it”. Now, if in this world there were only 5000 iPads left, then I might con­sider pre­serv­ing those doomed 1000 iPads.

I think you should ap­ply this ar­gu­ment to­wards some­thing else, and it re­ally needs it: an­i­mal con­ser­va­tion efforts. The amount of dis­pro­por­tionate money and fo­cus spent on cer­tain an­i­mals over oth­ers is highly un­fair. It hap­pens be­cause peo­ple give money ac­cord­ing to their emo­tional at­tach­ment. Hence, whales over bask­ing sharks, or seal pups over frogs, for ex­am­ple. Bio­di­ver­sity is an as­set with un­limited po­ten­tial, but be­fore we can gene se­quence it all, we’re los­ing this di­ver­sity in front of our lives. It’s like, na­ture has served to us on a plate amaz­ing de­signs pat­terns and strate­gies, but we don’t care.

• So, clearly, the best way to op­ti­mize your util­ity func­tion is to start a gene bank for freez­ing tis­sue sam­ples from ev­ery species. You can clone them back if they turn out to be use­ful. It’s a lot cheaper than con­ser­va­tion­ism, I as­sure you.

• GiveWell is eth­i­cally ques­tion­able and tak­ing them (and their met­rics) at face value is du­bi­ous wis­dom. Here’s why. http://​​mssv.net/​​wiki/​​in­dex.php/​​Givewell

They are not just a “site that col­lects and in­ter­prets data”, they col­lect ac­tual money and dis­burse less of that money, to char­i­ties who are rated ac­cord­ing to a highly ques­tion­able sys­tem, which was made up by peo­ple with lit­tle ex­pe­rience in char­ity and lots of ex­pe­rience in hedge funds.

Seems to me they are a self-in­serted mid­dle-man, whose busi­ness model is to lev­er­age the hu­man char­i­ta­ble im­pulse into an op­por­tu­nity to scoop off a lit­tle or a lot of cream for their pre­cious lit­tle selves. Ac­cord­ing to their fi­nan­cial state­ments in 2009 they took in ~$750K and gave out ~$110K. That’s quite some over­head.

• I re­quest more in­for­ma­tion. Your link goes to an “as­tro­tur­fing scan­dal”, which was stupid but doesn’t ob­vi­ously cast doubt on their met­rics or sincer­ity (also, wow; re­mind me never to do any­thing wrong on the In­ter­net where those peo­ple can see it).

As far as I can tell, GiveWell al­lows you to donate ei­ther to them or to their recom­mended char­i­ties, and makes it very clear which is which. I don’t know if they’re do­ing enough good to jus­tify their op­er­at­ing ex­penses, but they don’t seem to be do­ing any­thing de­liber­ately dishon­est in that re­gard.

But If you can point me to other peo­ple who do the same sort of thing GiveWell does but bet­ter, and are eas­ily ac­cessible on­line, I’ll switch the links to them.

• I’d clas­sify it as an in­di­ca­tor of Holden Karnofsky’s sense of ethics, per­son­ally.

Char­ity Nav­i­ga­tor also makes the claim of an­a­lyz­ing char­i­ties’ perfor­mance and I can’t speak to the rel­a­tive qual­ity of the two sites’ met­rics, but Char­ity Nav­i­ga­tor ap­par­ently only takes dona­tions di­rectly for it­self. This is more trans­par­ent - $1 given to them is$1 to them, and $1 given to Char­ity X at Char­ity Nav­i­ga­tor’s recom­men­da­tion is$1 to Char­ity X. On the other hand, $1 to “Givewell and what­ever char­i­ties Givewell recom­mends”, given to Givewell, will be di­vided up as Givewell pleases, and un­less I’m miss­ing some­thing fun­da­men­tal, it pleased it­self to di­vide it$640K (Givewell) : $110K (donated) in 2009. I’ll ad­mit to a com­plete lack of sur­prise at that. • Char­ity Nav­i­ga­tor ad­mits “We do not cur­rently eval­u­ate the qual­ity of the pro­grams and ser­vices a char­ity pro­vides. As soon as we de­velop a method­ol­ogy for do­ing so, we will. For now, how­ever, we limit our rat­ings to an anal­y­sis of a char­ity’s fi­nan­cial health.” As such, I don’t see them as re­ally in the same busi­ness as GiveWell. They’re use­ful for avoid­ing get­ting scammed, but not for max­i­miz­ing the effi­ciency of your char­i­ta­ble giv­ing. I’ve awk­wardly added a link to Nav­i­ga­tor in the ar­ti­cle, but think I’ll con­tinue to link GiveWell. • I don’t see your point.$1 given to Char­ity Nav­i­ga­tor is $1 given to Char­ity Nav­i­ga­tor;$1 given to char­ity X at Char­ity Nav­i­ga­tor’s recom­men­da­tion is $1 given to char­ity X.$1 given to GiveWell is 85 cents given to GiveWell and 15 cents given to char­i­ties Y and Z. $1 given to char­ity X at GiveWell’s recom­men­da­tion is$1 given to char­ity X.

This is a point in favour of GiveWell, by any mea­sure.

• Only if you as­sume that (a) donors are ac­tu­ally aware of an 85%:15% split in the char­i­ties’ dis­fa­vor; (b) ap­prove of that. I would ex­pect the naive as­sump­tion to be on the or­der of 90%:10% in char­i­ties’ fa­vor, but maybe that’s just me.

Now, their dona­tion pages for sep­a­rate char­i­ties eg http://​​www.givewell.org/​​in­ter­na­tional/​​top-char­i­ties/​​villagereach/​​donate do state that the dona­tion is di­rect to the char­ity, which is .a good thing.

So it’s “I’m will­ing to take your money for me, but if you want to give it to X, give it to X di­rectly” vs “I’m will­ing to take your money for me to split be­tween me and X, or you can give it to X”. Now on the face of it, that looks like X would get more money in the sec­ond sce­nario, as you point out. How­ever there is an in­her­ent naive as­sump­tion there that the split will be fair to X. If Donor A wanted to give $50 to Char­ity Nav­i­ga­tor and$50 to char­i­ties through Char­ity Nav­i­ga­tor, A has to give those amounts sep­a­rately. If A wants to give $50 to Givewell and$50 to char­i­ties through Givewell, A may be tempted to just give $100 to Givewell un­der the as­sump­tion that Givewell will split it$50/​$50. I sug­gest that donors who as­sumed that Givewell will be split­ting at 50%/​50% or bet­ter, have been if not de­ceived, at least per­mit­ted to op­er­ate un­der a false as­sump­tion where the one who could cor­rect the as­sump­tion (ie, Givewell) benefits from not do­ing so. I think the split with po­ten­tial breach of trust is more eth­i­cally du­bi­ous than the known split. I’ll ad­mit that it’s pos­si­ble that Givewell have cleaned their act up since 2007. But they seem to have a sig­nifi­cantly higher on­line pro­file than Char­ity Nav­i­ga­tor, while also seem­ing to have a smaller num­ber of char­i­ties rated and smaller amount of money donated due to their in­fluence, which “smells funny” (or if you pre­fer, trig­gers heuris­tic es­ti­mates of sus­pi­cious­ness) to me. • I don’t know what met­ric you’re us­ing to de­ter­mine whether CN or GiveWell has a sig­nifi­cantly higher on­line pro­file, but “char­i­ty­nav­i­ga­tor.org″ re­turns ten times as many hits on Google as ”givewell.org No doubt about it, Char­ity Nav­i­ga­tor eval­u­ates more char­i­ties, but they’re able to do so be­cause they use a sub­stan­tially less rigor­ous method­ol­ogy. They carry out a fun­da­men­tally differ­ent func­tion: they’re a watch­dog group, aiming to avoid fraud, while GiveWell con­ducts re­search to try to find ex­cel­lent char­i­ties, a much more difficult task. (Looks like Yvain makes this point above). Be­cause it’s younger and ap­peals to a smaller group of peo­ple that want to max­i­mize their im­pact, GiveWell moves sub­stan­tially less money than Char­ity Nav­i­ga­tor (though it’s grow­ing). Edit: I’ve been a fan of the GiveWell pro­ject for quite some time, and have an in­for­mal agree­ment to join GiveWell as an em­ployee in mid-2011. I’m a stu­dent and was com­ment­ing sim­ply on my own be­half, with­out any dis­cus­sion with GiveWell. After Holden com­mented, I emailed him to say that I had com­mented, and he recom­mended that I dis­close my plan to work for them. • “we may use these funds for op­er­at­ing ex­penses or grants to char­i­ties, at our dis­cre­tion” (source: http://​​www.givewell.org/​​about/​​donate ) This does not im­ply “we will treat the char­i­ties fairly” at all. It im­plies noth­ing about the nu­mer­i­cal split. My quick at­tempt to find char­ity nav­i­ga­tor’s fa­vorite char­i­ties: http://​​www.char­i­ty­nav­i­ga­tor.org/​​in­dex.cfm?bay=topten.de­tail&listid=100 There are sev­eral uni­ver­si­ties drown­ing in money on this list. That does not sound very effi­cient. Look­ing for an ex­pla­na­tion, they seem to care too much about % op­er­at­ing ex­penses and so on, and not enough about real im­pact. Do­ing the same for GiveWell (look­ing at US to make it more com­pa­rable): http://​​www.givewell.org/​​united-states/​​top-char­i­ties We have two char­i­ties. Both are eval­u­ated based on marginal im­pact and mea­sured effec­tive­ness. The sec­ond, KIPP, is an ex­tremely “cool” pro­gram. Con­clu­sion: It ap­pears as though GiveWell recom­mends more effi­cient char­i­ties with bet­ter crite­ria. GiveWell also seems to be­have in a man­ner that sounds bet­ter to In­ter­net geeks, for ex­am­ple with their mis­take list. This is suffi­cient ex­pla­na­tion for their In­ter­net pop­u­lar­ity. • It is pos­si­ble that Donor A may choose to donate fully to GiveWell for many rea­sons, in­clud­ing a prior as­sump­tion that it’s 50:50 or bet­ter with­out check­ing eas­ily available facts. This re­flects badly on Donor A, not GiveWell, and does not in any way make a case for call­ing GiveWell “eth­i­cally ques­tion­able”. The most you could pos­si­bly say is “GiveWell does not overly pan­der to the low­est com­mon de­nom­i­na­tor enough” but these peo­ple are already donat­ing their money to Make a Wish foun­da­tion or some­thing equally silly. I be­labour this point be­cause char­i­ties run solely on their ap­pear­ance as eth­i­cal, and to the ex­tent that your com­ments de­prive GiveWell of pos­si­ble dona­tions on the ba­sis of spu­ri­ous claims, you’re do­ing a bad thing. • This is Holden Karnofsky, the co-Ex­ec­u­tive Direc­tor of GiveWell. As a fre­quent Less Wrong reader, I’m re­ally glad to see the thought­ful dis­cus­sion here. Thanks to Yvain for call­ing at­ten­tion both to GiveWell and to the gen­eral topic of effec­tive giv­ing. First off, much of this con­tent over­laps with our own, so peo­ple in­ter­ested in this thread might also find the fol­low­ing links in­ter­est­ing: I’m mostly post­ing to clar­ify a few things re­gard­ing the con­cerns that have been raised about GiveWell (by aeschenkarnos). • We re­gret the as­tro­tur­fing that aeschenkarnos brought up. This in­ci­dent is dis­closed, along with other mis­takes we’ve made, on our short­com­ings list , which is ac­cessible via a top-level link on our nav­i­ga­tion bar. • Re­gard­ing the split be­tween grants to char­i­ties and funds spent on our own op­er­a­tions: • Early in our ex­is­tence, we re­lied on mak­ing grants of our own to char­i­ties. We weren’t able to point them to any benefits that would come from our recom­men­da­tions (since we were new and had no track record of in­fluenc­ing dona­tions), so rather than invit­ing them to be re­viewed, we in­vited them to ap­ply for grants (sub­ject to cer­tain con­di­tions such as pub­lic dis­clo­sure of ap­pli­ca­tion ma­te­ri­als). Grant­mak­ing is no longer im­por­tant to our pro­cess and we no longer so­licit dona­tions to be re­granted, though we still oc­ca­sion­ally re­ceive them. That ex­plains why the % of our funds spent on grants has fallen a lot, though it hasn’t hit zero. • At this point, we ac­tively so­licit dona­tions to GiveWell only when deal­ing with in­sti­tu­tional fun­ders or with peo­ple who have a re­la­tion­ship with us. When deal­ing with the gen­eral pub­lic, we put the so­lic­i­ta­tion on be­half of recom­mended char­i­ties—rather than our­selves—front and cen­ter. Our top char­i­ties page, linked promi­nently from our front page and nav­i­ga­tion bar and in other places through­out the site, links to “donate” pages for top char­i­ties ( here’s the one for our top-rated char­ity VillageReach ) that al­low us to track dona­tions, but oth­er­wise take no part in the dona­tion pro­cess (the money does not touch our bank ac­count). Th­ese “donate” pages also are linked from char­ity re­views. The only way to get to the “Donate to GiveWell” page is un­der “About GiveWell.” If donors make a con­sid­ered de­ci­sion to sup­port us rather than our top char­i­ties, we want them to be able to do so, but our site is de­signed to push the ca­sual user to our top char­i­ties. • In 2009 we tracked ~$1 mil­lion in dona­tions to our top char­i­ties as as re­sult of our re­search, while our own op­er­at­ing (non-grant) ex­penses were un­der $300k. We ex­pect 2010 to have a higher “dona­tions to top char­i­ties” figure on similar op­er­at­ing ex­penses. We are still new and hope the ra­tio will im­prove sub­stan­tially over time. • We have a policy of re­grant­ing un­re­stricted funds if our re­serves go above a cer­tain level; we don’t be­lieve in build­ing a mas­sive en­dow­ment for our­selves. This is the only con­di­tion un­der which we re­grant un­re­stricted funds. We don’t want donors to fear that we might blindly pile up re­serves with­out limit (we won’t), but we don’t want to get into all the de­tails of our “Ex­cess re­serves” policy on the Donate page, so we went with the lan­guage: “we may use these funds for op­er­at­ing ex­penses or grants to char­i­ties, at our dis­cre­tion.” • Bot­tom line—grant­mak­ing used to be an im­por­tant part of what we do but it isn’t now; the % of our funds spent on grants is not a mean­ingful figure. • Re­gard­ing Char­ity Nav­i­ga­tor: • I be­lieve Yvain is cor­rect to say that Char­ity Nav­i­ga­tor does not eval­u­ate effec­tive­ness (and ad­mits this) and that GiveWell does. See also this re­cent New York Times ar­ti­cle on planned changes at Char­ity Nav­i­ga­tor and Char­ity Nav­i­ga­tor’s dis­clo­sure of the full de­tails of its cur­rent method­ol­ogy. • I agree with alexan­deris that “num­ber of char­i­ties rated” is higher for Char­ity Nav­i­ga­tor pri­mar­ily be­cause its re­search is not as in-depth. I be­lieve Char­ity Nav­i­ga­tor would agree with this as well. • I be­lieve that Char­ity Nav­i­ga­tor has a sig­nifi­cantly higher pro­file than GiveWell, over­all, and know of no ev­i­dence sug­gest­ing oth­er­wise. How­ever, GiveWell does have a higher pro­file within cer­tain com­mu­ni­ties, in­clud­ing Less Wrong. I at­tribute our higher pro­file on Less Wrong to spe­cific in­di­vi­d­u­als in­clud­ing Michael Vas­sar, Anna Salomon, Carl Shul­man, Razib at GNXP, and mul­ti­fo­li­aterose. I don’t be­lieve any of these in­di­vi­d­u­als have plugged GiveWell in ig­no­rance of Char­ity Nav­i­ga­tor (in fact I have prob­a­bly dis­cussed the differ­ences speci­fi­cally with each of them). We’ve worked to find the best, most cost-effec­tive char­i­ties (in terms of ac­tual im­pact per marginal dol­lar) and write up all the de­tails of our anal­y­sis. We wel­come more com­ments and ques­tions about our work, whether here, on our blog, or via email. • Alright. You’ve given an ex­pla­na­tion here that seems rea­son­able to me, and you’ve con­tinued to run GiveWell for sig­nifi­cantly longer than I would have ex­pected if you were just in it for your­selves. For what it’s worth, I’ll give you the benefit of the doubt and I wish you well in your mis­sion. • I don’t know what fi­nan­cial state­ments you’re look­ing at, but if you look at the 2009 IRS form 990, it shows that they raised$374K and spent $340K. Of that,$110K was re-granted.

They did raise $768K in 2008, but they only spent$155K of it, and saved the rest. \$250K of it was re­stricted for the their eco­nomic em­pow­er­ment grant, which was dis­tributed in early 2010.

Fur­ther­more, I find your char­ac­ter­i­za­tion of their busi­ness model mis­lead­ing. I don’t know what Holden and Elie made work­ing for a hedge fund, but I bet it’s a hell of a lot more than they’re mak­ing at GiveWell, so crit­i­ciz­ing their self­ish­ness strikes me as mis­taken.

I don’t know if GiveWell pub­lishes the data on where they get their money, but I’m fairly cer­tain that very lit­tle of it comes from the gen­eral pub­lic in the way you’re sug­gest­ing. Their mid-2010 bud­get up­date (DOC), for in­stance, calcu­lates when they would run out of money if they don’t get any dona­tions that they aren’t already an­ti­ci­pat­ing. Most of their fund­ing, on my un­der­stand­ing, comes from the Hewlett Foun­da­tion and their board.

Edit: I’ve been a fan of the GiveWell pro­ject for quite some time, and have an in­for­mal agree­ment to join GiveWell as an em­ployee in mid-2011. I’m a stu­dent and was com­ment­ing sim­ply on my own be­half, with­out any dis­cus­sion with GiveWell. After Holden com­mented, I emailed him to say that I had com­mented, and he recom­mended that I dis­close my plan to work for them.