At the same time, this only works if more “disinterested” donors recognize groups that are for any reason isolated from such networks, or where the entire network received a correlated negative shock, so that “impersonal” aid funds are directed to improve their situation (because after all, that’s where the highest returns are).
I think this is a more robust system than having all aid be impersonal, because after all, what if the cost benefit calculation is wrong? So MacAskill can of course donate to whomever he wishes, including not donating to those he wishes to give to give the most, but personally I find that an example I will admire through observation rather than emulation.
At the same time, this only works if more “disinterested” donors recognize groups that are for any reason isolated from such networks, or where the entire network received a correlated negative shock, so that “impersonal” aid funds are directed to improve their situation (because after all, that’s where the highest returns are).
I think this is a more robust system than having all aid be impersonal, because after all, what if the cost benefit calculation is wrong? So MacAskill can of course donate to whomever he wishes, including not donating to those he wishes to give to give the most, but personally I find that an example I will admire through observation rather than emulation.