Lingering thoughts on the talk āHow to Handle Worldview Uncertaintyā by Hayley Clatterbuck (Rethink Priorities):
The talk proposed several ways that altruists with conflicting values can bargain in mutually beneficial ways, like loans, wagers, and trades, and suggested that the EA community should try to implement these more in practice and design institutions and mechanisms that incentivize them.
I think the EA Donation Election is an example of a community-wide mechanism for brokering trades between multiple anonymous donors. To illustrate this, consider a simple example of a trade, where Alice and Bob are donors with conflicting altruistic priorities. Aliceās top charity is Direct Transfers Everywhere and her second favorite is Pandemics No More. Bobās top charity is Lawyers for Chickens, and his second favorite is Pandemics No More. Bob is concerned that Aliceās donating to Direct Transfers Everywhere would cancel out the animal welfare benefits of his donating to Lawyers for Chickens, so he proposes that they both donate to their second choice, Pandemics No More.
The Donation Election does this in an automated, anonymous, community-wide way by using a mechanism like ranked-choice voting (RCV) to select winning charities. (The 2024 election uses RCV; the 2023 election used a points-based system similar to RCV.) Suppose that Alice and Bob are voting in the Donation Electionāand for simplicity, weāll pretend that the election uses RCV. If their first-choice charities (Direct Transfers Everywhere and Lawyers for Chickens) are not that popular among the electorate, those candidates will be eliminated, and Alice and Bobās votes reallocated to Pandemics No More. This achieves the same outcome as the trade in the previous example automatically, even though Alice and Bob may not have ever personally met and agreed to that trade.
Lingering thoughts on the talk āHow to Handle Worldview Uncertaintyā by Hayley Clatterbuck (Rethink Priorities):
The talk proposed several ways that altruists with conflicting values can bargain in mutually beneficial ways, like loans, wagers, and trades, and suggested that the EA community should try to implement these more in practice and design institutions and mechanisms that incentivize them.
I think the EA Donation Election is an example of a community-wide mechanism for brokering trades between multiple anonymous donors. To illustrate this, consider a simple example of a trade, where Alice and Bob are donors with conflicting altruistic priorities. Aliceās top charity is Direct Transfers Everywhere and her second favorite is Pandemics No More. Bobās top charity is Lawyers for Chickens, and his second favorite is Pandemics No More. Bob is concerned that Aliceās donating to Direct Transfers Everywhere would cancel out the animal welfare benefits of his donating to Lawyers for Chickens, so he proposes that they both donate to their second choice, Pandemics No More.
The Donation Election does this in an automated, anonymous, community-wide way by using a mechanism like ranked-choice voting (RCV) to select winning charities. (The 2024 election uses RCV; the 2023 election used a points-based system similar to RCV.) Suppose that Alice and Bob are voting in the Donation Electionāand for simplicity, weāll pretend that the election uses RCV. If their first-choice charities (Direct Transfers Everywhere and Lawyers for Chickens) are not that popular among the electorate, those candidates will be eliminated, and Alice and Bobās votes reallocated to Pandemics No More. This achieves the same outcome as the trade in the previous example automatically, even though Alice and Bob may not have ever personally met and agreed to that trade.
Update: The 2024 Donation Election is using straight-up ranked-choice voting; details here.