This is interesting, but only 3x marginal cost-effectiveness of GiveWell’s top charities is not that impressive given that GiveWell’s are supported by RCTs and careful adjustments, whereas there aren’t going to be very relevant RCTs on affecting policy or supporting innovation, so the evidence for direct impact is weaker. When you loosen evidentiary standards and are willing to make more bets with weaker links in your causal chain, I suspect you can do even better than 3x, at least if you’re moving much less money than GiveWell and its top charities.
I wouldn’t be surprised if fundraising for GiveWell’s top charities (via GWWC, TLYCS, OFTW, Founders Pledge) beat 3x GiveWell’s top charities, although presumably you could fundraise for GIF instead and do even better. And then there are different cause areas entirely, instead of global health and development.
On the other hand, Open Phil seems to have found it hard to beat GiveWell’s top charities for helping humans alive today even while loosening these standards, so maybe it is very hard, and either GIF is in fact doing really well, or we should be more skeptical that they’ve succeeded in beating GiveWell’s top charities 3x: https://www.openphilanthropy.org/research/givewells-top-charities-are-increasingly-hard-to-beat/
Skimming the report, but especially pp. 30-31, I’d probably adjust their estimates significantly downward as relying on informed speculation about future events, and probably on unclear assumptions about counterfactual funding and grantee/investee growth. I find such projections optimistic until proven otherwise.
That being said, I think it’s a good idea to have a variety of effective options for prospective donors to capture as large a slice of the philantrophic pie as feasible. If someone doesn’t find saving the lives of children under age 5 -- the means by which GiveWell’s top charities produce the bulk of their assessed value—compelling enough to attract the bulk of their donations, I still would like their money for another effective intervention even assuming that the alternative doesn’t score quite as well as GiveWell’s best.
(I haven’t read the report.)
This is interesting, but only 3x marginal cost-effectiveness of GiveWell’s top charities is not that impressive given that GiveWell’s are supported by RCTs and careful adjustments, whereas there aren’t going to be very relevant RCTs on affecting policy or supporting innovation, so the evidence for direct impact is weaker. When you loosen evidentiary standards and are willing to make more bets with weaker links in your causal chain, I suspect you can do even better than 3x, at least if you’re moving much less money than GiveWell and its top charities.
I wouldn’t be surprised if fundraising for GiveWell’s top charities (via GWWC, TLYCS, OFTW, Founders Pledge) beat 3x GiveWell’s top charities, although presumably you could fundraise for GIF instead and do even better. And then there are different cause areas entirely, instead of global health and development.
On the other hand, Open Phil seems to have found it hard to beat GiveWell’s top charities for helping humans alive today even while loosening these standards, so maybe it is very hard, and either GIF is in fact doing really well, or we should be more skeptical that they’ve succeeded in beating GiveWell’s top charities 3x: https://www.openphilanthropy.org/research/givewells-top-charities-are-increasingly-hard-to-beat/
Skimming the report, but especially pp. 30-31, I’d probably adjust their estimates significantly downward as relying on informed speculation about future events, and probably on unclear assumptions about counterfactual funding and grantee/investee growth. I find such projections optimistic until proven otherwise.
That being said, I think it’s a good idea to have a variety of effective options for prospective donors to capture as large a slice of the philantrophic pie as feasible. If someone doesn’t find saving the lives of children under age 5 -- the means by which GiveWell’s top charities produce the bulk of their assessed value—compelling enough to attract the bulk of their donations, I still would like their money for another effective intervention even assuming that the alternative doesn’t score quite as well as GiveWell’s best.