GiveWell, and the EA community at large, often emphasize the "cost of saving a life" as a key metric, $5,000 being the most commonly cited approximation.
As you point out, there is a lot more to EAs' foundational assumptions than "helping people". They sometimes claim that that's all that's important, but they sneak in a lot more in order to get to the utilitarian calculus at the next level up. The hours you put in nurturing your sick grandparents are very much hours spent helping people, but the EA bullet-biting suggests that that's a worse use of your time than making more money that you could then send to South Sudan.
I remember reading or listening to something where they mentioned a PhD thesis about this topic, possibly done by someone at Open Philanthropy (I tried to find it before posting but could not easily). I think at least some of the conclusion of the thesis was that saving a life in a developed country would be more valuable, but that the differences in cost strongly outweigh that. Just thought I would mention it in case it helps you look into the topic more. If I manage to find the paper, I’ll try to link it.
As someone who considers themselves an EA, this is a very good critique and should definitely be taken seriously. Consider posting it to the EA forum for it to be on the radar of EAs that may be able to look more deeply into these empirical questions: https://forum.effectivealtruism.org.
Thank you. I intend to post it on the forum but wanted to see if I can find time to do some back of the envelopes before posting it but you're right - could be worth just having others build on it if it's interesting.
So instead of buying mosquito nets and providing other medical aid, maybe hire mercenaries to take over the Congo and implement effective legal institutions.
I was being a little facetious. That said, if you take the whole Singer Drowning Child thing seriously, then the conquest and benevolent rule of Subsaharan Africa would seem to be the only choice. I’ll let you decide if this is a reducto ad absurdum or a serious policy proposal.
A while back I saw a documentary on an election in Kenya, and one of the candidates was campaigning against foreign aid for Kenya. He was saying that the country had received however many billions over the years, but there were still many problems, which indicated bad government. He said that foreign aid enables bad government, because it reduces governnment accountability to its people. He wanted it cut off, so the people would demand better government.
This makes sense to me. But it's difficult to be hard on that when you're talking about human lives. Foreign aid has undoubtedly helped keep tyrants in power in places like DPRK and Gaza - but can we just watch people die?
The "EA Tower" doesn't rest on the assumption that all lives have equal value. I think most EAs would readily admit that for the same price they would definitely choose to save an American over an African.
The assumption is that helping Africans is more effective per dollar, and I think that assumption is correct:
- Donating to GiveDirectly is more effective than donating to American charities. Like you said, extreme poverty is really bad, so giving directly to people in extreme poverty is extremely effective at alleviating suffering!
- Regardless, the latter claim belongs in the highest level of the "tower" and is not a foundational assumption like "some methods are more effective than others, and we should do those."
Feel free to read the updated piece. I'm guessing you read the previous version. My argument doesn't apply to Give Directly, it applies to charities that draw most of their justification from saving lives in poor countries. Trying to minimize the cost of lives saved will systematically bias you towards buying the lowest quality of life. So if the EA community wants to save lives, it should either loudly disclose this or make the adjustments (the form of which I allude to )
1) buy a bunch of new lives in sub Saharan Africa. These lives remain stuck in something resembling Malthusian poverty generation after generation and contribute nothing to human advancement.
2) get some smart sperm and eggs via IVF (your local fertility clinic already has done this for you) mix in some polygenic screening, and create the next generation of smart healthy achievers that drive forward human progress
#2 would definitely maximize utility more then #1, the calculation isn’t even close.
Heck you can maybe make #2 even better if you don’t bother to raise the children! You could dump em in an orphanage right after surrogacy and their good genes would probably mean they turn out alright. Odds are they would even be adopted by some random UMC couple (the demand for healthy white babies with good genes for adoption vastly exceeds supply).
So if the utility calculations are so obvious and the lives created will probably be of very high quality with positive externalities even if you literally throw them to the wolves, why aren’t EAs talking about it?
Because it’s all just blathering nonsense. Nobody cars about utility. EA always just boils down to “here is my slightly different take on acceptable center left causes my social circle might give me an attaboy for”. Bed Nets sound cool at a party, that’s it.
I certainly think that recognizing Africa is something of a black hole where resources go in and human development doesn’t come out is a good revelation for EAs to have, but let’s take it further shall we. It turns out that maximizing utility for future generations mostly means doing a lot of boring ROI maximizing stuff that looks a lot like making profits and investing and smart people having more kids.
People need to get out of the business of helping Africans. They don't deserve it. Foreign aid, IMF and World Bank has created a survival of the weakest in Africa. All of these global institutions need to be dismantled and the continent needs to be left alone.
This seems broadly reasonable. I think the overwhelming cos effectiveness of African public health charities (and the argument that public health problems are destabilizing so PH charities would also indirectly help improve institutions) means they still win out even with this adjustment though.
I agree that some public health improvements will help long run institutional quality. But in that case, i would focus on lead removal etc and things that impact human capital in the medium term, not simply m aximize lilves saved in the worst parts of the worst countries by distributing bednets.
IIRC the top rated EA charities now are mostly deworming, which reduce physical (and plausibly mental) disability more than mortality, which I think has more effect than lead removal.
Huh, you're right, I hadn't checked in a while. The vitamin A deficiency one seems closer to the effects of deworming.
(Even for malaria though, I suspect the effects of malaria prevention are more likely to be increased stability than actually higher population growth).
Projecting value systems on other people and then disturbing whatever they are doing to forcefully "uplift" them is generic narcissism
"They are just like me but worse" rejects existence of other cultures and values, donating any amount and observing that they are still "worse" proves the elite status of the donator (possibility of target becoming better than donator is obviously excluded)
Hmm. While I generally agree, I think what the EA movement has done (albeit imperfectly) is align personal incentives for status and recognition to amount of social welfare generated through work or donations. Do you think efforts in this direction are always misguided? It's not that i disagree but I'm trying to understand how strong the public choice objections are relative to the benefit of trying
If you mean maximizing donated dollar amount (regardless of target or effect), that effort to "get rid" of income implies the belief that the income is partially or wholly earned wrongfully or incidentally, and a wish to make someone else responsible for (spending) it, removing the pressure to responsibly conserve, use to fuel action or otherwise spend it with care.
Whatever the reasons (and whether they support specific value systems), the act of "helping" is destructive - less if done as a conscious response to a personal plea with conscious expectation of certain value exchange in "win-win" outcome.
Purely altruistic help does not exist and thus whatever hides under the label is malicious, indulging in the repressed consciously-unacceptable "win" of the inevitably obligatory "win-win".
As you point out, there is a lot more to EAs' foundational assumptions than "helping people". They sometimes claim that that's all that's important, but they sneak in a lot more in order to get to the utilitarian calculus at the next level up. The hours you put in nurturing your sick grandparents are very much hours spent helping people, but the EA bullet-biting suggests that that's a worse use of your time than making more money that you could then send to South Sudan.
I elaborated on the point a bit here: https://loveofallwisdom.substack.com/p/you-dont-have-to-drop-philosophy
I remember reading or listening to something where they mentioned a PhD thesis about this topic, possibly done by someone at Open Philanthropy (I tried to find it before posting but could not easily). I think at least some of the conclusion of the thesis was that saving a life in a developed country would be more valuable, but that the differences in cost strongly outweigh that. Just thought I would mention it in case it helps you look into the topic more. If I manage to find the paper, I’ll try to link it.
Here’s the post : https://forum.effectivealtruism.org/posts/7WKiW4fTvJMzJwPsk/adverse-selection-in-minimizing-cost-per-life-saved
That’d be great
I was going to post on the forum but will refrain if someone already looked into this
Posting on the forum might also get you a link to the paper more quickly, though I get why you’d want to have it beforehand.
As someone who considers themselves an EA, this is a very good critique and should definitely be taken seriously. Consider posting it to the EA forum for it to be on the radar of EAs that may be able to look more deeply into these empirical questions: https://forum.effectivealtruism.org.
Thank you. I intend to post it on the forum but wanted to see if I can find time to do some back of the envelopes before posting it but you're right - could be worth just having others build on it if it's interesting.
Here’s the post: https://forum.effectivealtruism.org/posts/7WKiW4fTvJMzJwPsk/adverse-selection-in-minimizing-cost-per-life-saved
Awesome! Thanks for contributing!
So instead of buying mosquito nets and providing other medical aid, maybe hire mercenaries to take over the Congo and implement effective legal institutions.
How about do nothing in Congo ? Thats an option …
I was being a little facetious. That said, if you take the whole Singer Drowning Child thing seriously, then the conquest and benevolent rule of Subsaharan Africa would seem to be the only choice. I’ll let you decide if this is a reducto ad absurdum or a serious policy proposal.
Nice post, Vaish! Somewhat relatedly, you may like mine on "Helping animals or saving human lives in high income countries is arguably better than saving human lives in low income countries?" (https://forum.effectivealtruism.org/posts/Gazt9dKDD4isGcupy/helping-animals-or-saving-human-lives-in-high-income).
A while back I saw a documentary on an election in Kenya, and one of the candidates was campaigning against foreign aid for Kenya. He was saying that the country had received however many billions over the years, but there were still many problems, which indicated bad government. He said that foreign aid enables bad government, because it reduces governnment accountability to its people. He wanted it cut off, so the people would demand better government.
This makes sense to me. But it's difficult to be hard on that when you're talking about human lives. Foreign aid has undoubtedly helped keep tyrants in power in places like DPRK and Gaza - but can we just watch people die?
It's a difficult issue.
The "EA Tower" doesn't rest on the assumption that all lives have equal value. I think most EAs would readily admit that for the same price they would definitely choose to save an American over an African.
The assumption is that helping Africans is more effective per dollar, and I think that assumption is correct:
- Donating to GiveDirectly is more effective than donating to American charities. Like you said, extreme poverty is really bad, so giving directly to people in extreme poverty is extremely effective at alleviating suffering!
- A harder claim is that saving children is more effective than giving adults cash. Givewell puts a lot of thought into this question (https://blog.givewell.org/2012/05/30/giving-cash-versus-giving-bednets/), and has concluded that malaria nets > GiveDirectly.
- Regardless, the latter claim belongs in the highest level of the "tower" and is not a foundational assumption like "some methods are more effective than others, and we should do those."
Feel free to read the updated piece. I'm guessing you read the previous version. My argument doesn't apply to Give Directly, it applies to charities that draw most of their justification from saving lives in poor countries. Trying to minimize the cost of lives saved will systematically bias you towards buying the lowest quality of life. So if the EA community wants to save lives, it should either loudly disclose this or make the adjustments (the form of which I allude to )
EA tends to end in the repugnant conclusion.
Let’s take two ways of spending money.
1) buy a bunch of new lives in sub Saharan Africa. These lives remain stuck in something resembling Malthusian poverty generation after generation and contribute nothing to human advancement.
2) get some smart sperm and eggs via IVF (your local fertility clinic already has done this for you) mix in some polygenic screening, and create the next generation of smart healthy achievers that drive forward human progress
#2 would definitely maximize utility more then #1, the calculation isn’t even close.
Heck you can maybe make #2 even better if you don’t bother to raise the children! You could dump em in an orphanage right after surrogacy and their good genes would probably mean they turn out alright. Odds are they would even be adopted by some random UMC couple (the demand for healthy white babies with good genes for adoption vastly exceeds supply).
So if the utility calculations are so obvious and the lives created will probably be of very high quality with positive externalities even if you literally throw them to the wolves, why aren’t EAs talking about it?
Because it’s all just blathering nonsense. Nobody cars about utility. EA always just boils down to “here is my slightly different take on acceptable center left causes my social circle might give me an attaboy for”. Bed Nets sound cool at a party, that’s it.
I certainly think that recognizing Africa is something of a black hole where resources go in and human development doesn’t come out is a good revelation for EAs to have, but let’s take it further shall we. It turns out that maximizing utility for future generations mostly means doing a lot of boring ROI maximizing stuff that looks a lot like making profits and investing and smart people having more kids.
I think most EA supporters care about alleviating short-term suffering in the most efficient way possible instead of longtermism.
People need to get out of the business of helping Africans. They don't deserve it. Foreign aid, IMF and World Bank has created a survival of the weakest in Africa. All of these global institutions need to be dismantled and the continent needs to be left alone.
This seems broadly reasonable. I think the overwhelming cos effectiveness of African public health charities (and the argument that public health problems are destabilizing so PH charities would also indirectly help improve institutions) means they still win out even with this adjustment though.
I agree that some public health improvements will help long run institutional quality. But in that case, i would focus on lead removal etc and things that impact human capital in the medium term, not simply m aximize lilves saved in the worst parts of the worst countries by distributing bednets.
IIRC the top rated EA charities now are mostly deworming, which reduce physical (and plausibly mental) disability more than mortality, which I think has more effect than lead removal.
The top two charities still seem to be malaria ones
Huh, you're right, I hadn't checked in a while. The vitamin A deficiency one seems closer to the effects of deworming.
(Even for malaria though, I suspect the effects of malaria prevention are more likely to be increased stability than actually higher population growth).
Ya I’m not saying preventing malaria is not good. It’s certainly good. The question is how good ?
Projecting value systems on other people and then disturbing whatever they are doing to forcefully "uplift" them is generic narcissism
"They are just like me but worse" rejects existence of other cultures and values, donating any amount and observing that they are still "worse" proves the elite status of the donator (possibility of target becoming better than donator is obviously excluded)
Hmm. While I generally agree, I think what the EA movement has done (albeit imperfectly) is align personal incentives for status and recognition to amount of social welfare generated through work or donations. Do you think efforts in this direction are always misguided? It's not that i disagree but I'm trying to understand how strong the public choice objections are relative to the benefit of trying
If you mean maximizing donated dollar amount (regardless of target or effect), that effort to "get rid" of income implies the belief that the income is partially or wholly earned wrongfully or incidentally, and a wish to make someone else responsible for (spending) it, removing the pressure to responsibly conserve, use to fuel action or otherwise spend it with care.
Why do you think so many people risk their lives to come here ?
Whatever the reasons (and whether they support specific value systems), the act of "helping" is destructive - less if done as a conscious response to a personal plea with conscious expectation of certain value exchange in "win-win" outcome.
Purely altruistic help does not exist and thus whatever hides under the label is malicious, indulging in the repressed consciously-unacceptable "win" of the inevitably obligatory "win-win".
Plug the leak before you grab the bucket.