One reason why we are moving more slowly is that our current estimates of the gap between marginal animal and human funding opportunities is very different from the one in your post – within one order of magnitude, not three. And given the high uncertainty around our estimates here, we think one order of magnitude is well within the “margin of error” .
I assume that even though your answers are within one order of magnitude, the animal-focused work is the one that looks more cost-effective. Is that right?
Assuming so, your answer doesn’t make sense to me because OP funds roughly 6x more human-focused GHW relative to farm animal welfare (FAW). Even if you have wide uncertainty bounds, if FAW is looking more cost-effective than human work, surely this ratio should be closer to 1:1 rather than 1:6? It seems bizarre (and possibly an example of omission bias) to fund the estimated less cost-effective thing 6x more and justify it by saying you’re quite uncertain.
Long story short, should we not just allocate our funding to the best of our current knowledge (even by your calculations, more towards FAW) and then update accordingly if things change?
I personally agree with your reasoning, but it assumes the marginal cost-effectiveness of the human-focussed and animal-focussed interventions should be the same. Open Phil is not sold on this:
We’re also unsure conceptually whether we should be trying to equalize marginal returns between FAW and GHW or whether we should continue with our current approach.
I do not know what the “current approach” specifically involves, but it has led to Open Phil starting 6 new areas with a focus on human welfare in the last few years[1]. So it naively seems to me like Open Phil could have done more to increase the amount of funding going to animal welfare if there was a desire to do so. These areas will not be turned off easily. If Open Phil was in the process of deliberating how much funding animal-focussed interventions should receive relative to human-focussed ones, I would have expected a bigger investment in growing the internal cause-prioritisation team, or greater funding of similar research elsewhere[2].
I assume that even though your answers are within one order of magnitude, the animal-focused work is the one that looks more cost-effective. Is that right?
Assuming so, your answer doesn’t make sense to me because OP funds roughly 6x more human-focused GHW relative to farm animal welfare (FAW). Even if you have wide uncertainty bounds, if FAW is looking more cost-effective than human work, surely this ratio should be closer to 1:1 rather than 1:6? It seems bizarre (and possibly an example of omission bias) to fund the estimated less cost-effective thing 6x more and justify it by saying you’re quite uncertain.
Long story short, should we not just allocate our funding to the best of our current knowledge (even by your calculations, more towards FAW) and then update accordingly if things change?
Nice point, James!
I personally agree with your reasoning, but it assumes the marginal cost-effectiveness of the human-focussed and animal-focussed interventions should be the same. Open Phil is not sold on this:
I do not know what the “current approach” specifically involves, but it has led to Open Phil starting 6 new areas with a focus on human welfare in the last few years[1]. So it naively seems to me like Open Phil could have done more to increase the amount of funding going to animal welfare if there was a desire to do so. These areas will not be turned off easily. If Open Phil was in the process of deliberating how much funding animal-focussed interventions should receive relative to human-focussed ones, I would have expected a bigger investment in growing the internal cause-prioritisation team, or greater funding of similar research elsewhere[2].
South Asian air quality, global aid policy, innovation policy, effective altruism with a GHW focus, global health R&D, and global public health policy.
Open Phil has made grants to support Rethink’s moral weight project, but this type of work has apparently not been fully supported by Open Phil: