OP currently uses the welfare ranges that Luke Muehlhauser produced as part of his 2018 moral patienthood report. He lists species’ ranges here, though we use point estimates he produced internally. Luke’s numbers are steeper /​ more hierarchical than Rethink’s.
I actually fitted distributions to the moral weights conditional on moral patienthood Luke shared in that post, and multiplied them by Luke’s probabilities of moral patienthood given in his 2017 report, and got expected moral weights pretty close to 1 (i.e. similar to that of humans):
Species
Mean moral weight relative to humans
Uniform
Normal
Loguniform
Lognormal
Pareto
Logistic
Chimpanzees
0.900
0.900
0.490
3.27
+∞
0.900
Pigs
1.40
1.40
0.765
13.1
+∞
1.40
Cows
2.00
2.00
1.14
132
+∞
2.00
Chickens
4.00
4.00
2.41
1.50 k
+∞
4.00
Rainbow trouts
4.55
4.55
3.00
28.4 k
+∞
4.55
Fruit flies
2.50
2.50
1.95
2.46 M
+∞
2.50
Would it be possibe to share the specific point estimates you are using, and how Luke (@lukeprog) obtained them?
We sometimes test the sensitivity of species-specific grants to using Luke or Rethink’s welfare ranges. So far this hasn’t often been action-guiding, since we’re already primarily funding work focused on the most numerous farmed vertebrates (chicken and fish) and our funding on invertebrate welfare is more limited by other factors.
I see. That being said, welfare ranges can still affect cause prioritisation?
Sorry, I can’t share our internal numbers. To date, we haven’t focused on making direct comparisons between GHW and FAW. Instead, we’ve focused on trying to equalize marginal returns within each area and do something more like worldview diversification to determine allocations across GHW, FAW, and Open Philanthropy’s other grantmaking. Luke has written about moral weights in the past, we’ve commissioned more recent work by Rethink Priorities, and we hope to do more research ourselves in the future—on moral weights and also on other components of BOTECs that would allow comparisons between animal and human focused work (e.g., welfare range, the difference between the number of humans and chickens, respectively, affected by a marginal intervention in each area), as well as our overall framework for making this decision. That said, the timing is TBD.
Thanks, Lewis!
I actually fitted distributions to the moral weights conditional on moral patienthood Luke shared in that post, and multiplied them by Luke’s probabilities of moral patienthood given in his 2017 report, and got expected moral weights pretty close to 1 (i.e. similar to that of humans):
Would it be possibe to share the specific point estimates you are using, and how Luke (@lukeprog) obtained them?
I see. That being said, welfare ranges can still affect cause prioritisation?
Sorry, I can’t share our internal numbers. To date, we haven’t focused on making direct comparisons between GHW and FAW. Instead, we’ve focused on trying to equalize marginal returns within each area and do something more like worldview diversification to determine allocations across GHW, FAW, and Open Philanthropy’s other grantmaking. Luke has written about moral weights in the past, we’ve commissioned more recent work by Rethink Priorities, and we hope to do more research ourselves in the future—on moral weights and also on other components of BOTECs that would allow comparisons between animal and human focused work (e.g., welfare range, the difference between the number of humans and chickens, respectively, affected by a marginal intervention in each area), as well as our overall framework for making this decision. That said, the timing is TBD.