Sure—admittedly I don’t put as much thought into these as I probably should, and have deferred a reasonable amount to GiveWell.
At a high level I generally donate to GHD causes because I’m more risk averse with my donations than my personal time (i.e. I’m very happy to think and contribute to things that are potentially high payoff but more speculative in my work time, but I prefer my donations to be making tangible, measurable impact with short feedback loops, even if that means “only” getting ~$100/DALY).
More specifically, I chose Malaria Consortium (MC) over AMF because MC is marginally more cost effective at the moment, and because there was a recent paper about the benefits of combined chemoprevention alongside the recently released RTS,S malaria vaccine.
I chose New Incentives because it’s also similarly cost-effective, and I have a soft spot for improving vaccine coverage.
I chose GiveDirectly because that seems “best-in-class” in terms of optimising for preferences of recipients / beneficiaries. I don’t have fully formed views about the extent to which we should defer to recipients VS prioritise other measures, but until I do, I opt to donate a bit to GiveDirectly for this reason despite it being less cost effective by GiveWell metrics.
And then I chose my local women’s refuge because I think sexual assault / intimate partner violence (IPV) is terrible (I get pretty riled up by abuse of power / trust generally), and I care a lot about it for other reasons that I won’t get into here. I’m mindful that this probably means I’m not maximising DALYs averted in expectation, but I’m okay with this because doing so appeals to the less maximising / less utilitarian parts of me. I donate to the local refuge instead of some LMIC version that might have higher EV in part because it’s more convenient to, in part because NZ have among the worst domestic violence / IPV rates in the developed world, in part because I care about my local community, and in part because I’m fully aware I’m not donating to this because I think it’s the most cost-effective by EA standards.[1]
That being said, I do think the harm/suffering from intimate partner violence, both to the victim as well as family members is really difficult to capture. I wouldn’t be surprised if something like IPV was considerably undervalued, conditional on the DALY-maximising approach being wrong. (Something like better access to palliative care / relieving terrible end of life care suffering is probably also something that fits in this kind of category). It’s not primarily intended as a hedge against a DALY-maximising approach, but I guess it functionally acts as one.
I have also donated to global health and development interventions in the past, but now focus on longtermist ones (see here). FWIW, one thing which made me wonder about the cost-effectiveness of GiveWell’s top charities was their effects on the population size:
According to Wilde 2019, it seems that the population increases in the near-term, but decreases in the long-term:
Abstract: “The effect on fertility is positive only temporarily – lasting only 1-3 years after the beginning of the ITN distribution programs – and then becomes negative. Taken together, these results suggest the ITN distribution campaigns may have caused fertility to increase unexpectedly and temporarily, or that these increases may just be a tempo effect – changes in fertility timing which do not lead to increased completed fertility”.
Conclusion: “In contrast, our findings do not support the contention that erosion of international funding for malaria control, specifically of ITNs, would lead to higher fertility rates in the short-run. While our results are suggestive that this may be the case for long-run fertility, we show the exact opposite for the short-run”.
If the results are suggestive that decreasing ITNs leads to higher long-run fertility, AMF would tend to decrease longterm fertility.
I do not know whether increasing the population size is good or bad, but I think it may well be the major driver for the total effect of GiveWell’s top charities.
Even neglecting long-term effects, the sign of the effect of GiveWell’s top charities seems unclear if we account for the impact on animals:
“A plausible estimate is that the average person on Earth prevents ~1.4 * 107 insect-years by his/her environmental impact each year”.
Assuming each insect has 316 k neurons, as I did here, the mean human prevents 4.42 T neuron-years each year.
Each human has about 0.0875 T neurons (see here), so the scale of the insect welfare affected by the mean human is arguably 50.5 (= 4.42/0.0875) times as large as the scale of the welfare of the mean human.
I guess the above factor is not much affected in the quality of life per neuron. According to Charity Entrepeneurship’s Weighted Animal Welfare Index, in a scale from −100 to 100, the total welfare of the wild bug is −42, and that of a human in a low middle-income country is 32.
[hastily written, apologies in advance if I don’t respond further]
Hi! Thanks for sharing. Unfortunately I don’t have the bandwidth to engage with the attention this comment deserves. Part of the reason I didn’t default to including my reasoning initially is because I didn’t want it to turn into a “neartermism vs longtermism” discussion thread. But it’s my fault for not clarifying this, and I appreciate you sharing your thoughts, which will hopefully be helpful for other readers who do have time to engage!
I think in short it sounds like I have larger uncertainties than you around both empirical and philosophical considerations, and because of these uncertainties, it’s not clear to me that funding things like AI alignment youtube content or longtermist researchers or forecasting infrastructure are clearly better than preventing malaria or getting more kids vaccinated, or cash transfers to the very poor.
If you have a compelling case that the cost-effectiveness of GiveWell’s top charities is much lower than they think it is, I’d encourage you to also reach out to GiveWell! They recently had a “Change our Minds” contest, and while the contest has closed, I’m confident they’ll be open to a forum post that strongly justified why they are too optimistic about their top charities.
Lastly, as mentioned earlier, I prefer to try and contribute to longtermist causes with my work time,[1] and donate to neartermist causes.
Things I’m involved with / have done that could plausibly be categorised as longtermist in nature: potentially writing a grand futures paper with Anders Sandberg @ FHI, taking part in a forecasting tourney organised by Tetlock, getting involved with advocating for mitigating existential risks for UN OCA process, moderating an AI safety panel at the Internet Governance Forum, and contributing to someone’s script submission for longtermist youtube content.
I didn’t go into this previously because I think it is irrelevant, but I’m just noticing a feeling of somehow needing to defend my choice of only donating to neartermist causes. But to be clear I currently think neartermist donation choices are justifiable even without any contributions from one’s work life to existential risk or the far future of humanity.
Thank you Bruce! If it’s easy to share, how did you decide on these nonprofits?
Sure—admittedly I don’t put as much thought into these as I probably should, and have deferred a reasonable amount to GiveWell.
At a high level I generally donate to GHD causes because I’m more risk averse with my donations than my personal time (i.e. I’m very happy to think and contribute to things that are potentially high payoff but more speculative in my work time, but I prefer my donations to be making tangible, measurable impact with short feedback loops, even if that means “only” getting ~$100/DALY).
More specifically, I chose Malaria Consortium (MC) over AMF because MC is marginally more cost effective at the moment, and because there was a recent paper about the benefits of combined chemoprevention alongside the recently released RTS,S malaria vaccine.
I chose New Incentives because it’s also similarly cost-effective, and I have a soft spot for improving vaccine coverage.
I chose GiveDirectly because that seems “best-in-class” in terms of optimising for preferences of recipients / beneficiaries. I don’t have fully formed views about the extent to which we should defer to recipients VS prioritise other measures, but until I do, I opt to donate a bit to GiveDirectly for this reason despite it being less cost effective by GiveWell metrics.
And then I chose my local women’s refuge because I think sexual assault / intimate partner violence (IPV) is terrible (I get pretty riled up by abuse of power / trust generally), and I care a lot about it for other reasons that I won’t get into here. I’m mindful that this probably means I’m not maximising DALYs averted in expectation, but I’m okay with this because doing so appeals to the less maximising / less utilitarian parts of me. I donate to the local refuge instead of some LMIC version that might have higher EV in part because it’s more convenient to, in part because NZ have among the worst domestic violence / IPV rates in the developed world, in part because I care about my local community, and in part because I’m fully aware I’m not donating to this because I think it’s the most cost-effective by EA standards.[1]
That being said, I do think the harm/suffering from intimate partner violence, both to the victim as well as family members is really difficult to capture. I wouldn’t be surprised if something like IPV was considerably undervalued, conditional on the DALY-maximising approach being wrong. (Something like better access to palliative care / relieving terrible end of life care suffering is probably also something that fits in this kind of category). It’s not primarily intended as a hedge against a DALY-maximising approach, but I guess it functionally acts as one.
Thanks for this candid answer! I really appreciate hearing your thinking and more emotional motivations behind the decisions.
Thanks for sharing, Bruce!
I have also donated to global health and development interventions in the past, but now focus on longtermist ones (see here). FWIW, one thing which made me wonder about the cost-effectiveness of GiveWell’s top charities was their effects on the population size:
According to Wilde 2019, it seems that the population increases in the near-term, but decreases in the long-term:
Abstract: “The effect on fertility is positive only temporarily – lasting only 1-3 years after the beginning of the ITN distribution programs – and then becomes negative. Taken together, these results suggest the ITN distribution campaigns may have caused fertility to increase unexpectedly and temporarily, or that these increases may just be a tempo effect – changes in fertility timing which do not lead to increased completed fertility”.
Conclusion: “In contrast, our findings do not support the contention that erosion of international funding for malaria control, specifically of ITNs, would lead to higher fertility rates in the short-run. While our results are suggestive that this may be the case for long-run fertility, we show the exact opposite for the short-run”.
If the results are suggestive that decreasing ITNs leads to higher long-run fertility, AMF would tend to decrease longterm fertility.
I do not know whether increasing the population size is good or bad, but I think it may well be the major driver for the total effect of GiveWell’s top charities.
Even neglecting long-term effects, the sign of the effect of GiveWell’s top charities seems unclear if we account for the impact on animals:
Accordind to this article from Brian Tomasik:
“A plausible estimate is that the average person on Earth prevents ~1.4 * 107 insect-years by his/her environmental impact each year”.
Assuming each insect has 316 k neurons, as I did here, the mean human prevents 4.42 T neuron-years each year.
Each human has about 0.0875 T neurons (see here), so the scale of the insect welfare affected by the mean human is arguably 50.5 (= 4.42/0.0875) times as large as the scale of the welfare of the mean human.
I guess the above factor is not much affected in the quality of life per neuron. According to Charity Entrepeneurship’s Weighted Animal Welfare Index, in a scale from −100 to 100, the total welfare of the wild bug is −42, and that of a human in a low middle-income country is 32.
[hastily written, apologies in advance if I don’t respond further]
Hi! Thanks for sharing. Unfortunately I don’t have the bandwidth to engage with the attention this comment deserves. Part of the reason I didn’t default to including my reasoning initially is because I didn’t want it to turn into a “neartermism vs longtermism” discussion thread. But it’s my fault for not clarifying this, and I appreciate you sharing your thoughts, which will hopefully be helpful for other readers who do have time to engage!
I think in short it sounds like I have larger uncertainties than you around both empirical and philosophical considerations, and because of these uncertainties, it’s not clear to me that funding things like AI alignment youtube content or longtermist researchers or forecasting infrastructure are clearly better than preventing malaria or getting more kids vaccinated, or cash transfers to the very poor.
If you have a compelling case that the cost-effectiveness of GiveWell’s top charities is much lower than they think it is, I’d encourage you to also reach out to GiveWell! They recently had a “Change our Minds” contest, and while the contest has closed, I’m confident they’ll be open to a forum post that strongly justified why they are too optimistic about their top charities.
Lastly, as mentioned earlier, I prefer to try and contribute to longtermist causes with my work time,[1] and donate to neartermist causes.
Things I’m involved with / have done that could plausibly be categorised as longtermist in nature: potentially writing a grand futures paper with Anders Sandberg @ FHI, taking part in a forecasting tourney organised by Tetlock, getting involved with advocating for mitigating existential risks for UN OCA process, moderating an AI safety panel at the Internet Governance Forum, and contributing to someone’s script submission for longtermist youtube content.
I didn’t go into this previously because I think it is irrelevant, but I’m just noticing a feeling of somehow needing to defend my choice of only donating to neartermist causes. But to be clear I currently think neartermist donation choices are justifiable even without any contributions from one’s work life to existential risk or the far future of humanity.
Thanks for clarifying, and the nudge to engage with GiveWell!