Zooming out for a bit, if you’re trying to evaluate GWWC’s historical impact, there’s five categories of impact:
Money pledgers have already donated
Money pledgers will donate in the future
People who will take the pledge in the future due to past investments (e.g. GWWC’s website will keep creating pledgers even if they closed tomorrow).
Money influenced by people who didn’t take the pledge
Impact that isn’t in the form of donations to top charities (creating the EA movement, founding CEA, their research contributions, spinning off 80k, substantial policy influence)
GWWC are very modest and only focus on (1) and (2) :)
To evaluate the total impact, you need to calculate the expected value for each component. You should make your own estimates given the best information you have.
Only counting (1) would be a very poor estimate. (2) to (5) are very unlikely to be zero. My personal guess is that:
(2) is about 10x (1) [because I think GWWC’s internal estimates are reasonable]
(3) and (4) are similar to (1)
(5) is much larger than (2)
However, even if you only cared about (1), GWWC’s multiplier would still be about 10x, which would make them more cost-effective than GiveWell recommended charities, and equal to or more cost-effective than the other EA fundraising orgs (in the past). So I’m not sure it’s even that decision relevant.
How do should you weigh the information?
Zooming out for a bit, if you’re trying to evaluate GWWC’s historical impact, there’s five categories of impact:
Money pledgers have already donated
Money pledgers will donate in the future
People who will take the pledge in the future due to past investments (e.g. GWWC’s website will keep creating pledgers even if they closed tomorrow).
Money influenced by people who didn’t take the pledge
Impact that isn’t in the form of donations to top charities (creating the EA movement, founding CEA, their research contributions, spinning off 80k, substantial policy influence)
GWWC are very modest and only focus on (1) and (2) :)
To evaluate the total impact, you need to calculate the expected value for each component. You should make your own estimates given the best information you have.
Only counting (1) would be a very poor estimate. (2) to (5) are very unlikely to be zero. My personal guess is that:
(2) is about 10x (1) [because I think GWWC’s internal estimates are reasonable] (3) and (4) are similar to (1) (5) is much larger than (2)
However, even if you only cared about (1), GWWC’s multiplier would still be about 10x, which would make them more cost-effective than GiveWell recommended charities, and equal to or more cost-effective than the other EA fundraising orgs (in the past). So I’m not sure it’s even that decision relevant.