Do you think this is an insurmountable problem with the model, or just one that is practically difficult to overcome? I can think of two possible ways to fix it:
In theory, committed donors’ willing to donate should depend on the cost-effectiveness of their donation. As chicken-welfare years (CWY) can be purchased more cheaply, the portion of their income that donors are willing to spend on CWY should increase. Thus, its plausible that for donors already in this space, the ability to counterfactually “buy” CWY at a ~2:1 discount would increase their donation amount. If they donated the amount of this increase to FarmKind, I think the donation would count as truly counterfactual.
You could also have a large donor (e.g., Open Phil) who is pre-committed to using the money for effective charities, but is open as between several different cause areas. Suppose they set aside $10MM in their budget for matching to approved effective charities, not charged to any existing cause area. From the donor perspective, this may be nearly full counterfactual matching. But for the matched donor’s choice, the bulk of the funds would have likely flowed to causes the non-EA donor probably cares little about. (I don’t get the sense that most non-EA people open to farmed animal welfare donations would ascribe much if any value to an AI safety donation, so it wouldn’t be that different in their eyes to burning the money.)
Do you think this is an insurmountable problem with the model, or just one that is practically difficult to overcome? I can think of two possible ways to fix it:
In theory, committed donors’ willing to donate should depend on the cost-effectiveness of their donation. As chicken-welfare years (CWY) can be purchased more cheaply, the portion of their income that donors are willing to spend on CWY should increase. Thus, its plausible that for donors already in this space, the ability to counterfactually “buy” CWY at a ~2:1 discount would increase their donation amount. If they donated the amount of this increase to FarmKind, I think the donation would count as truly counterfactual.
You could also have a large donor (e.g., Open Phil) who is pre-committed to using the money for effective charities, but is open as between several different cause areas. Suppose they set aside $10MM in their budget for matching to approved effective charities, not charged to any existing cause area. From the donor perspective, this may be nearly full counterfactual matching. But for the matched donor’s choice, the bulk of the funds would have likely flowed to causes the non-EA donor probably cares little about. (I don’t get the sense that most non-EA people open to farmed animal welfare donations would ascribe much if any value to an AI safety donation, so it wouldn’t be that different in their eyes to burning the money.)