Matching donations supposedly generate more charitable giving. That raises the question whether EAs should give matching donations to non-EA donations, instead of giving direct donations themselves. The matching donation could be both above or below 1:1.
There is one problem, though, which has to do with counterfactual trust (see Ord’s Moral Trade paper). Lots of the time, EAs will have a certain amount that they will want to donate. Therefore, if the donation drive generates less non-EA donations than expected—so that some of the matching funds aren’t used—the EAs will be inclined to donate the rest of the matching funds anyway. But that will mean that they donate the same amount regardless of how much was given to the charity in question by non-EAs. If this gets known, it will of course destroy the incentive to give more that matching is supposed to create.
One solution is this. Suppose that this group of EA donors value radically different charities C1,...,Cn roughly equally—X-risk charities, animal charities, AMF, etc. (They might not do that individually, but as a group they value them equally.) Then they could match donations to all of these charities, setting the ratio so high so that all of the money would certainly be spent. This would create a sort of competitive situation, where donors would compete to give as soon as possible to their favourite charity, before the matching funds ran out.
Here it is of course important that these non-EA donors don’t value C1,...,Cn equally, but I believe that would normally be the case.
I’m not sure I follow your solution. If I do, here’s a possible worry:
Suppose that the group of EA donors agree to match donations to all the n charities at m :1 (up to some limit). Given your setup, many individual members of the group may end up donating their $X to charity B instead of their preferred charity A. For this to be worth it for them, they must presumably think that a donation of $(1+1/m)X to B is better than a donation of $X to charity A, but I suspect that this would often not be the case.
Is that right, or did I misunderstand your proposal?
Thanks for this very good comment. You’re right. You could adjust the ratio differently to C1,...,Cn, to make sure that all EA contributors to the matching fund have reason to expect more than one dollar in total going to their top charity for every dollar they contribute to the matching fund. E.g., the fund would match donations to charities that non-EAs find less attractive 3:1, whereas charities that non-EAs find attractive could be funded 1:3.
I should have been clearer in my comment about this.
Regarding illusory matching, my hope is that the idea in the OP could solve the problem of illusory matching.
This GiveWell blog post by Holden Karnofsky is critical of donation matching, for the same reason that you mention:
We’ve discussed whether we might be able to provide “true” donation matching – finding a donor who would give to our top charities only on condition that others did – but not surprisingly, everyone we could think of who would be open to making a large gift to our top charities would be open to this whether or not we could match them up with smaller donors. Ultimately, the only match we can offer is illusory matching.
Matching donations supposedly generate more charitable giving. That raises the question whether EAs should give matching donations to non-EA donations, instead of giving direct donations themselves. The matching donation could be both above or below 1:1.
There is one problem, though, which has to do with counterfactual trust (see Ord’s Moral Trade paper). Lots of the time, EAs will have a certain amount that they will want to donate. Therefore, if the donation drive generates less non-EA donations than expected—so that some of the matching funds aren’t used—the EAs will be inclined to donate the rest of the matching funds anyway. But that will mean that they donate the same amount regardless of how much was given to the charity in question by non-EAs. If this gets known, it will of course destroy the incentive to give more that matching is supposed to create.
One solution is this. Suppose that this group of EA donors value radically different charities C1,...,Cn roughly equally—X-risk charities, animal charities, AMF, etc. (They might not do that individually, but as a group they value them equally.) Then they could match donations to all of these charities, setting the ratio so high so that all of the money would certainly be spent. This would create a sort of competitive situation, where donors would compete to give as soon as possible to their favourite charity, before the matching funds ran out.
Here it is of course important that these non-EA donors don’t value C1,...,Cn equally, but I believe that would normally be the case.
I’m not sure I follow your solution. If I do, here’s a possible worry:
Suppose that the group of EA donors agree to match donations to all the n charities at m :1 (up to some limit). Given your setup, many individual members of the group may end up donating their $X to charity B instead of their preferred charity A. For this to be worth it for them, they must presumably think that a donation of $(1+1/m)X to B is better than a donation of $X to charity A, but I suspect that this would often not be the case.
Is that right, or did I misunderstand your proposal?
Thanks for this very good comment. You’re right. You could adjust the ratio differently to C1,...,Cn, to make sure that all EA contributors to the matching fund have reason to expect more than one dollar in total going to their top charity for every dollar they contribute to the matching fund. E.g., the fund would match donations to charities that non-EAs find less attractive 3:1, whereas charities that non-EAs find attractive could be funded 1:3.
I should have been clearer in my comment about this.
Regarding illusory matching, my hope is that the idea in the OP could solve the problem of illusory matching.
This GiveWell blog post by Holden Karnofsky is critical of donation matching, for the same reason that you mention: