Disclaimer: Iâm funded by EA for animal welfare work.
Some thoughts:
a. So much of the debate feels like a debate on identities and values. Iâd really love to see people nitpicking into technical details of cost-effectiveness estimates instead.
b. I think itâs worth reminding that animal welfare interventions are less cost-effective than they were when Simcikas conducted his analysis.
c. I generally feel much more comfortable standing behind Givewellâs estimates but Givewell doesnât analyse cost-effectiveness of advocacy work. My biggest misgivings about cost-effectiveness estimates are due to the difficulty of assessing advocacy work. I think we should make a lot more progress on this.
d. People seem to keep forgetting that uncertainty cuts both ways. If the moral worth of animals is too uncertain, that is also a reason against confidently dismissing them.
e. I donât think we have made much progress on the question of âHow much important is cage to cage-free transition for a chicken in terms of human welfare?â. I donât think Rethink Priorities Welfare ranges answer that question. In general Iâm confused about the approach of trying to find overall welfare capacities of different species rather than just focusing on comparing specific experiences of different individuals. In RPâs report, hereâs how the question of welfare comparison was addressed:
âI estimated the DALY equivalent of a year spent in each type of pain assessed by the Welfare Footprint Project by looking at the descriptions of and disability weights assigned to various conditions assessed by the Global Burden of Disease Study in 2019 and comparing these to the descriptions of each type of pain tracked by the Welfare Footprint Project.â
I think this is the core question on this issue and it merits a much longer and thorough analysis. I would love to see a team of biologists, animal behaviour experts and human health experts coming together to produce a more detailed report on this.
f. I think there should be more concrete examples for PR costs of animal welfare work. Animal welfare has been around for sometime and I donât see that it has created notorious enemies for EA that try to drag down the movement. On the contrary it has even brought in new donors for some of the non-animal welfare parts of the movement(The Navigation Fund). EA-supported interventions on animal welfare are generally pretty moderate and popular. Cage-free referendums were always won by over 60% support(78% support in Massachusetts!). End The Cage Age petition got 1.4 million signatures in the EU. EA-supported Nähtamatud Loomad got the NGO of the year award from Estonian president. Animal welfare work has its enemies, but they donât seem to have affected EA that much.
g. On the contrary I found animal welfare quite useful for EA community building. Open Philanthropy donating an additional 5 million dollars to AMF doesnât create new entry opportunities to EA. Whereas many of the EA organisers in Turkey got involved in the movement through the local EA supported animal advocacy organisation. Animal advocacy offers localised, effective and non-monetary ways to contribute. That is pretty useful in low trust or middle income countries.
But overall I think animal welfare spending should be evaluated primarily according to its impact on animals. If someone thinks that some positive or negative side-effect is significant enough they should concretely show it and provide an estimate for it.
h. I feel similarly about ripple effects. If someone is attempting to maximise that kind of outcome, they should choose an intervention that maximises ripple effects. Otherwise both animal advocacy work and global health work have loads of side effects on peopleâs values, ideas, population growth, economic growth and itâs an extremely ambitious effort to sum these all up and have a verdict on the overall direction of them. Iâm also surprised that people think animal advocacyâs effect is isolated on animals only. Itâs a mass communications work that leaves an impact on millions of people. That is a whole load of ripples.
d. People seem to keep forgetting that uncertainty cuts both ways. If the moral worth of animals is too uncertain, that is also a reason against confidently dismissing them.
Uncertainty also means a higher cost-effectiveness of animal welfare research which tries to decrease the uncertainty, given the high value of information.
b. I think itâs worth reminding that animal welfare interventions are less cost-effective than they were when Simcikas conducted his analysis.
Admittedly I havenât been following work on animal welfare cost-effectiveness analysis closely, but this is news to me; can you point me to further readings on this?
c. I generally feel much more comfortable standing behind Givewellâs estimates but Givewell doesnât analyse cost-effectiveness of advocacy work. My biggest misgivings about cost-effectiveness estimates are due to the difficulty of assessing advocacy work. I think we should make a lot more progress on this.
I agree with the need for the latter; Iâm thinking in particular of Animal Askâs systematic review finding âinsufficient evidence to break down overall policy success into the baseline rate of success and the counterfactual impact of lobbyingâ. I default to the evaluative framework in Founders Pledgeâs guide to evaluating policy advocacy organisations but would be keen to learn how to improve upon it.
re: the former, here are some GiveWell policy advocacy-related CEAs:
For both BOTECs GiveWell explicitly mentioned that they relyâon a number of difficult best-guess assumptions and judgment calls about modeling structure. It therefore contains less information value than cost-effectiveness estimates for our top charities, which limits its comparabilityâ, so Iâm not sure youâd feel as comfortable standing behind these estimates as with the top charity CEAs. And none of the models address the counterfactual estimation issue Animal Ask identified, again at a quick skimâcorrect me if Iâm wrong on this.
(None of this changes my general sense that funding top animal welfare interventions are more cost-effective on the margin than GHW.)
This seems to be a representative publicly available estimate from 4 years ago by Lewis Bollard:
âThis is a major question for us, and one we continue to research. Our current very rough estimate is that our average $ spent on corporate campaigns and all supporting work (which is ~40% of our total animal grant-making) achieves the equivalent of ~7 animals spared a year of complete suffering. We use this a rough benchmark for BOTECs on new grants, and my best guess is this reflects roughly the range we should hope for the last pro-animal dollar. â
I think several more up to date estimates will be available soon.
For advocacy evaluation, a concrete area for improvement is the following. Sauliusâs analysis has a really nice section titled âWays this estimate could be misleadingâ. Other advocates cite concerns similar to those when they argue against corporate welfare campaigns. They usually donât have empirical evidence, but I donât have super strong evidence to show them wrong either. Iâm not very happy about that.
What did you think of the GiveWell policy advocacy CEAs & BOTECs I linked? I shared them in response to your â...but Givewell doesnât analyse cost-effectiveness of advocacy workâ so I wondered if you had a different take.
I appreciate the correction. When I said âI generally feel much more comfortable standing behind Givewellâs estimatesâ that was for their main page recommendations. I currently wonât prioritise reviewing these BOTECS in detail in the short term but as a future exercise I will look into the linked analyses and compare them to animal welfare ones.
Disclaimer: Iâm funded by EA for animal welfare work.
Some thoughts:
a. So much of the debate feels like a debate on identities and values. Iâd really love to see people nitpicking into technical details of cost-effectiveness estimates instead.
b. I think itâs worth reminding that animal welfare interventions are less cost-effective than they were when Simcikas conducted his analysis.
c. I generally feel much more comfortable standing behind Givewellâs estimates but Givewell doesnât analyse cost-effectiveness of advocacy work. My biggest misgivings about cost-effectiveness estimates are due to the difficulty of assessing advocacy work. I think we should make a lot more progress on this.
d. People seem to keep forgetting that uncertainty cuts both ways. If the moral worth of animals is too uncertain, that is also a reason against confidently dismissing them.
e. I donât think we have made much progress on the question of âHow much important is cage to cage-free transition for a chicken in terms of human welfare?â. I donât think Rethink Priorities Welfare ranges answer that question. In general Iâm confused about the approach of trying to find overall welfare capacities of different species rather than just focusing on comparing specific experiences of different individuals. In RPâs report, hereâs how the question of welfare comparison was addressed:
âI estimated the DALY equivalent of a year spent in each type of pain assessed by the Welfare Footprint Project by looking at the descriptions of and disability weights assigned to various conditions assessed by the Global Burden of Disease Study in 2019 and comparing these to the descriptions of each type of pain tracked by the Welfare Footprint Project.â
I think this is the core question on this issue and it merits a much longer and thorough analysis. I would love to see a team of biologists, animal behaviour experts and human health experts coming together to produce a more detailed report on this.
f. I think there should be more concrete examples for PR costs of animal welfare work. Animal welfare has been around for sometime and I donât see that it has created notorious enemies for EA that try to drag down the movement. On the contrary it has even brought in new donors for some of the non-animal welfare parts of the movement(The Navigation Fund). EA-supported interventions on animal welfare are generally pretty moderate and popular. Cage-free referendums were always won by over 60% support(78% support in Massachusetts!). End The Cage Age petition got 1.4 million signatures in the EU. EA-supported Nähtamatud Loomad got the NGO of the year award from Estonian president. Animal welfare work has its enemies, but they donât seem to have affected EA that much.
g. On the contrary I found animal welfare quite useful for EA community building. Open Philanthropy donating an additional 5 million dollars to AMF doesnât create new entry opportunities to EA. Whereas many of the EA organisers in Turkey got involved in the movement through the local EA supported animal advocacy organisation. Animal advocacy offers localised, effective and non-monetary ways to contribute. That is pretty useful in low trust or middle income countries.
But overall I think animal welfare spending should be evaluated primarily according to its impact on animals. If someone thinks that some positive or negative side-effect is significant enough they should concretely show it and provide an estimate for it.
h. I feel similarly about ripple effects. If someone is attempting to maximise that kind of outcome, they should choose an intervention that maximises ripple effects. Otherwise both animal advocacy work and global health work have loads of side effects on peopleâs values, ideas, population growth, economic growth and itâs an extremely ambitious effort to sum these all up and have a verdict on the overall direction of them. Iâm also surprised that people think animal advocacyâs effect is isolated on animals only. Itâs a mass communications work that leaves an impact on millions of people. That is a whole load of ripples.
Nice points, Emre!
Uncertainty also means a higher cost-effectiveness of animal welfare research which tries to decrease the uncertainty, given the high value of information.
Admittedly I havenât been following work on animal welfare cost-effectiveness analysis closely, but this is news to me; can you point me to further readings on this?
I agree with the need for the latter; Iâm thinking in particular of Animal Askâs systematic review finding âinsufficient evidence to break down overall policy success into the baseline rate of success and the counterfactual impact of lobbyingâ. I default to the evaluative framework in Founders Pledgeâs guide to evaluating policy advocacy organisations but would be keen to learn how to improve upon it.
re: the former, here are some GiveWell policy advocacy-related CEAs:
2017 CEA of the Centre for Pesticide Suicide Prevention (grant writeup, 2018 blog post where they explained their reasoning in considering policy advocacy orgs in general), and 2021 skeleton BOTEC of the same org (grant writeup)
2021 BOTEC of Vital Strategies (grant writeup) to reduce harms of excessive alcohol consumption in LMICs.
For both BOTECs GiveWell explicitly mentioned that they relyâon a number of difficult best-guess assumptions and judgment calls about modeling structure. It therefore contains less information value than cost-effectiveness estimates for our top charities, which limits its comparabilityâ, so Iâm not sure youâd feel as comfortable standing behind these estimates as with the top charity CEAs. And none of the models address the counterfactual estimation issue Animal Ask identified, again at a quick skimâcorrect me if Iâm wrong on this.
(None of this changes my general sense that funding top animal welfare interventions are more cost-effective on the margin than GHW.)
This seems to be a representative publicly available estimate from 4 years ago by Lewis Bollard:
âThis is a major question for us, and one we continue to research. Our current very rough estimate is that our average $ spent on corporate campaigns and all supporting work (which is ~40% of our total animal grant-making) achieves the equivalent of ~7 animals spared a year of complete suffering. We use this a rough benchmark for BOTECs on new grants, and my best guess is this reflects roughly the range we should hope for the last pro-animal dollar. â
I think several more up to date estimates will be available soon.
For advocacy evaluation, a concrete area for improvement is the following. Sauliusâs analysis has a really nice section titled âWays this estimate could be misleadingâ. Other advocates cite concerns similar to those when they argue against corporate welfare campaigns. They usually donât have empirical evidence, but I donât have super strong evidence to show them wrong either. Iâm not very happy about that.
Thanks for the pointers, much appreciated.
What did you think of the GiveWell policy advocacy CEAs & BOTECs I linked? I shared them in response to your â...but Givewell doesnât analyse cost-effectiveness of advocacy workâ so I wondered if you had a different take.
I appreciate the correction. When I said âI generally feel much more comfortable standing behind Givewellâs estimatesâ that was for their main page recommendations. I currently wonât prioritise reviewing these BOTECS in detail in the short term but as a future exercise I will look into the linked analyses and compare them to animal welfare ones.