Thanks to everyone who voted and commented! It was helpful to learn more about how EAs think about multiplier orgs, and I hope it was helpful to hear my perspective from inside one of those orgs.
Here are my biggest takeaways from the discussion, apologies that it took me so long to post this:
Somebody is likely wrong: either multiplier orgs are pursuing strategies that don’t work, or those strategies do work and donors are missing out on leverage. If EAs can learn more about which statement is true, there’s a real opportunity for improvement. In theory there could be a middle ground where multiplier orgs are pursuing strategies that work at their current scale and marginal multiples are close to one, but it seems unlikely that all/most multiplier orgs fall into that category.
If you look at the history of multiplier orgs in EA, I think it’s clear that at least for some of them the model has worked. REG seems like a pretty clearcut case study of an organization that has spent very little money to move a lot of money to highly effective charities from donors who almost certainly never would have heard of those charities without REG.
By a large margin, people’s biggest objection to multiplier orgs is that they don’t trust the multipliers those organizations are reporting. People mentioned a variety of concerns about the reported numbers, including that they aren’t counterfactually adjusted and that they think the marginal return on new donations is likely to be lower than the historical average return. And this may create a reinforcing dynamic, where EAs aren’t excited about multiplier orgs because they get the sense that other EAs aren’t excited about them.
There are very high barriers to entry to getting a better understanding of the multiplier space (In other words, I think this area is quite vetting constrained.) There are a lot of different organizations, lots of different methodological considerations, and inconsistent reporting across multiplier orgs (e.g. money moved numbers may or may not be counterfactually adjusted). It would be hard enough for one person learn enough to make an informed decision; it’s completely unreasonable to expect everyone to learn this much on their own. (Note: it’s not clear to me that comparing multiplier orgs is harder than, say, comparing movement building orgs; my sense is that both areas are vetting constrained.)
The most sophisticated potential funders for meta orgs (Open Phil and the EA Infrastructure Fund) have actually supported about a dozen different multiplier organizations. However, it’s unlikely that these two funders alone can support a thriving ecosystem of multiplier orgs, given that Open Phil only considers a narrow subset of multiplier orgs and that the Infrastructure Fund’s funding capacity is quite low relative to the what a thriving ecosystem would need.
The multiplier ecosystem is diverse enough such that there may well be an organization out there that addresses your biggest concerns about the space. For example, if you’re worried that people might not keep the GWWC pledge, Founders Pledge’s legally binding pledge could be interesting. Conversely, if you’re worried that FP’s founders will give to ineffective charities, GWWC could be a good option. REG is extremely efficient but is relatively hard to scale, while TLYCS has a more scalable model but is willing to sacrifice efficiency in pursuit of that scale (i.e. TLYCS is more likely to pursue a strategy we think will cost $10 million and move $100 million [10x leverage] than we would be to pursue a strategy that would cost $1k and move $100k [100x leverage].) I’m not trying to argue here that one model is superior to the other, my point is that the universe of multiplier orgs is broad enough to appeal to a lot of different types of donors.
I get the sense many EAs don’t realize just how terrible a “minimum viable product” is, and how significantly an MVP can be improved with dedicated work. In the context of multiplier organizations, I think there’s a misperception that a basic website with limited maintenance is good enough, and that this misperception drives the concerns people have about marginal impact being lower than average impact. The empirical evidence is at odds with the preference for MVP models, based on the experiences of orgs like RC Fwd, GWWC, and TLYCS. TLYCS is a pretty clear example because we essentially used the website + volunteer model from ~2009 through mid-2013, before switching to paid full time staff. When we made the switch, web traffic (which had been flat for 4 years) started increasing and continued to do so, and money moved has steadily grown ever since (and is probably >25x higher than it was in the MVP years).
Outcomes I’d like to see going forward:
I’d love to see someone write up an overview of the multiplier space, similar to Larks’ annual AI Alignment Literature Review and Charity Comparison. Consolidating information would make it much easier for donors to engage with the space. Something as simple as a list of organizations with a few sentences about their work, their multiplier data, and links to more info would go a long way. (Ideally this would be done by someone who doesn’t work at a multiplier org; I’ll post this as a volunteer project on EA Work Club.)
I’d hope that overview would encourage more EAs dip their toes in the water by making a small donation to one or more multiplier orgs and/or subscribing to their mailing lists (I just did this to put some skin in the game). This is less about the actual money, and more about making it more likely you’ll stay informed about their work going forward. The more you do that, the more you’ll be able to make your own informed decision about whether their model is working.
A final note… While I’d love to see more people donating to multiplier orgs, I’d hate to see donors naively donating to the organization with the highest multiplier or otherwise incentivizing multiplier orgs to prioritize maximizing their short term multiplier. Ideally, both donors and organizations will prioritize strategies that maximize long run impact, and prioritize the magnitude of that long run impact (money moved – expenses) rather than the efficiency of that impact (money moved / expenses). For donors, I’d recommend asking 1) “do I believe in the strategy?” and 2) “do I believe the team can execute the strategy?”
Thanks to everyone who voted and commented! It was helpful to learn more about how EAs think about multiplier orgs, and I hope it was helpful to hear my perspective from inside one of those orgs.
Here are my biggest takeaways from the discussion, apologies that it took me so long to post this:
Somebody is likely wrong: either multiplier orgs are pursuing strategies that don’t work, or those strategies do work and donors are missing out on leverage. If EAs can learn more about which statement is true, there’s a real opportunity for improvement. In theory there could be a middle ground where multiplier orgs are pursuing strategies that work at their current scale and marginal multiples are close to one, but it seems unlikely that all/most multiplier orgs fall into that category.
If you look at the history of multiplier orgs in EA, I think it’s clear that at least for some of them the model has worked. REG seems like a pretty clearcut case study of an organization that has spent very little money to move a lot of money to highly effective charities from donors who almost certainly never would have heard of those charities without REG.
By a large margin, people’s biggest objection to multiplier orgs is that they don’t trust the multipliers those organizations are reporting. People mentioned a variety of concerns about the reported numbers, including that they aren’t counterfactually adjusted and that they think the marginal return on new donations is likely to be lower than the historical average return. And this may create a reinforcing dynamic, where EAs aren’t excited about multiplier orgs because they get the sense that other EAs aren’t excited about them.
There are very high barriers to entry to getting a better understanding of the multiplier space (In other words, I think this area is quite vetting constrained.) There are a lot of different organizations, lots of different methodological considerations, and inconsistent reporting across multiplier orgs (e.g. money moved numbers may or may not be counterfactually adjusted). It would be hard enough for one person learn enough to make an informed decision; it’s completely unreasonable to expect everyone to learn this much on their own. (Note: it’s not clear to me that comparing multiplier orgs is harder than, say, comparing movement building orgs; my sense is that both areas are vetting constrained.)
The most sophisticated potential funders for meta orgs (Open Phil and the EA Infrastructure Fund) have actually supported about a dozen different multiplier organizations. However, it’s unlikely that these two funders alone can support a thriving ecosystem of multiplier orgs, given that Open Phil only considers a narrow subset of multiplier orgs and that the Infrastructure Fund’s funding capacity is quite low relative to the what a thriving ecosystem would need.
The multiplier ecosystem is diverse enough such that there may well be an organization out there that addresses your biggest concerns about the space. For example, if you’re worried that people might not keep the GWWC pledge, Founders Pledge’s legally binding pledge could be interesting. Conversely, if you’re worried that FP’s founders will give to ineffective charities, GWWC could be a good option. REG is extremely efficient but is relatively hard to scale, while TLYCS has a more scalable model but is willing to sacrifice efficiency in pursuit of that scale (i.e. TLYCS is more likely to pursue a strategy we think will cost $10 million and move $100 million [10x leverage] than we would be to pursue a strategy that would cost $1k and move $100k [100x leverage].) I’m not trying to argue here that one model is superior to the other, my point is that the universe of multiplier orgs is broad enough to appeal to a lot of different types of donors.
I get the sense many EAs don’t realize just how terrible a “minimum viable product” is, and how significantly an MVP can be improved with dedicated work. In the context of multiplier organizations, I think there’s a misperception that a basic website with limited maintenance is good enough, and that this misperception drives the concerns people have about marginal impact being lower than average impact. The empirical evidence is at odds with the preference for MVP models, based on the experiences of orgs like RC Fwd, GWWC, and TLYCS. TLYCS is a pretty clear example because we essentially used the website + volunteer model from ~2009 through mid-2013, before switching to paid full time staff. When we made the switch, web traffic (which had been flat for 4 years) started increasing and continued to do so, and money moved has steadily grown ever since (and is probably >25x higher than it was in the MVP years).
Outcomes I’d like to see going forward:
I’d love to see someone write up an overview of the multiplier space, similar to Larks’ annual AI Alignment Literature Review and Charity Comparison. Consolidating information would make it much easier for donors to engage with the space. Something as simple as a list of organizations with a few sentences about their work, their multiplier data, and links to more info would go a long way. (Ideally this would be done by someone who doesn’t work at a multiplier org; I’ll post this as a volunteer project on EA Work Club.)
I’d hope that overview would encourage more EAs dip their toes in the water by making a small donation to one or more multiplier orgs and/or subscribing to their mailing lists (I just did this to put some skin in the game). This is less about the actual money, and more about making it more likely you’ll stay informed about their work going forward. The more you do that, the more you’ll be able to make your own informed decision about whether their model is working.
A final note… While I’d love to see more people donating to multiplier orgs, I’d hate to see donors naively donating to the organization with the highest multiplier or otherwise incentivizing multiplier orgs to prioritize maximizing their short term multiplier. Ideally, both donors and organizations will prioritize strategies that maximize long run impact, and prioritize the magnitude of that long run impact (money moved – expenses) rather than the efficiency of that impact (money moved / expenses). For donors, I’d recommend asking 1) “do I believe in the strategy?” and 2) “do I believe the team can execute the strategy?”