Proposal: Funding Diversification for Top Cause Areas
Epistemic status: Exploratory, just one proposal for how we might mitigate certain funding risks in the future in light of the collapse of FTX. Probably not a novel idea overall, but it’s timely and I think some parts of this proposal are new and worth considering.
Summary
Funding Diversification is something that could be explicitly evaluated for priority cause areas in EA. Working to have more diverse funding sources for top cause areas could make highly valued work more resilient to the sudden loss of any individual funder.
Introduction
It’s common in EA analysis to evaluate cause areas based on Importance, Tractability and Neglectedness. In light of the recent bankruptcy of FTX and the resulting funding crunch for several top EA causes, we might consider a fourth factor: Funding Diversification.[1]
Funding Diversification is similar to Neglectedness in that it evaluates the current resources going toward a given cause area. Except instead of considering the total amount of funding going into the cause area, it considers the diversity of funding sources the cause area draws from.
For example, if 70% of the funding for AI Safety came from the foundation of a single billionaire[2], then we would rate this cause area as having low funding diversification.
Benefits of funding diversification
High funding diversification is desirable because it provides financial stability in the event of a sudden loss in funding. For example, we might strive for there to be no single funding source which accounts for more than 10%[3] of the total funding in a top cause area. This way, even if the largest funder suddenly collapsed, 90% of funding for that cause area would remain.
In addition to financial stability, increased funding diversification means less exposure to moral, legal and reputational hazards from funder failures. If it turns out the largest funder in a space committed a crime, it’s a lesser crisis all around if the cause area was only 10% dependent on that funder rather than 70% dependent.
Another benefit of funding diversification is it can help with diversity of thought/approaches within a cause area. In case funders have strong opinions about how things should be done in the cause area they’re funding, funding diversification provides people working in that area with more opportunity to blaze a different trail. If their current funder has a blindspot which prevents them from seeing the merits of their new intervention or research direction, they are more likely to find another funder who will support them.
How do we increase funding diversification in a cause area?
People working on fundraising in a given cause area could take its funding diversification into account. If their area has low funding diversification, then they could prioritize raising funds from new donors, rather than going back to the largest funder to try and raise additional funds.
Similarly, grantmakers could take funding diversification into account when they are evaluating cause areas to donate to. If there is a cause area they believe is important but which has low funding diversification and they are not currently funding, then they could see it as a benefit to fund that area in order to help increase its funding diversification.
Arguments against
It’s hard enough just to get sufficient funding for a cause area. Sculpting the makeup of funding sources for a cause area isn’t a luxury we have.
Philanthropists are looking for opportunities to stand out and build a legacy. It’s exciting to be the majority funder in an important cause area. “Thanks to your grant, funding diversification in this cause area was increased by 5%” is a less exciting prospect.
Related ideas
Setting up financial endowments for top cause areas could be another way to provide them with financial security and help them to better weather funding shocks.[4]
In addition to funding diversification, cause areas could be evaluated in terms of the number/diversity of people, organizations, types of interventions, active research directions, etc.
Thanks to EA UBC and the people who attended tonight’s meetup for giving me some feedback on this idea.
- ^
Funding Diversification would be a slightly different kind of factor than Importance, Tractability and Neglectedness. I argue that it’s an important consideration, but it should probably influence a cause area’s overall priority calculation less than ITN does. So it would be more of a secondary consideration that comes into play in tactical funding decisions, rather than a primary criteria for evaluating a cause area’s priority.
- ^
This is a rough approximation of what I found to be true from looking at TylerMaule’s funding data. In this data, the 2022 total est. funding for Longtermism and Catastrophic Risk Prevention was $223M, with Future Fund providing $157M or 70.4% that. The data isn’t granular enough to see the funding for AI Safety specifically, but since AI Safety falls under the umbrella of Longtermism and Catastrophic Risk Prevention, its funding probably has a similar breakdown (but it might be different!).
- ^
10% is just a made up number, but it illustrates the point. There is probably a clever way to calculate optimal targets for funding diversification, but I haven’t gone to the trouble to figure that out here.
- ^
Credit to James Deng at the EA UBC meetup tonight who suggested this idea.
It seems a little strange for a secondary donor to give to an org with the objective of increasing their funding diversity if the idea is this is good because it increases stability. If the org’s primary donor collapses, the secondary donor could fund them regardless of whether or not they had in the past.
This is not a fully convincing counterargument, as it could still be reassuring to the staff of the org that they have a credible signal of security, and it encourages the secondary donor to do due diligence ahead of time.
Separately, I think it would be useful to think about the magnitudes here. Suppose you are a small org that gets 90% of your funding from one donor. For what $X are you indifferent between an incremental $100 from your primary donor and $X from a new secondary donor?
Diversification would be nice, but I don’t see the mathematics of how that could work with a megadonor who wants to concentrate in one area and largely occupied the field.
If Joe Smith is funding $150MM being 75 percent of funding in area X vs $50MM from elsewhere, you’d need $200MM in new money to get Joe down to 37.5 percent of funding in the area (150/400). Or youd need Joe to reduce funding to $30MM to get to 37.5 percent. Joe and other funders are unlikely to be motivated to do this.
Great post—I think this is a really important meta-topic within EA that doesn’t get enough airtime. It might also be worth considering the “hidden zero problem” coined by Mark Budolfson and Dean Spears here. The thrust of their argument is that if a charity is funded by the ultra-rich or their foundations, small donations may have measurably 0 impact.
As an example: suppose NGO X wants $10M in funding for 2022. Foundation X has been NGO X’s largest donor for a few years running. If small donors give NGO X $8M in 2022, Foundation X will fully fund it to $10M, but if small donors give $8M, Foundation X will give $1M more and still fully fund it to $10M. This means that some of the small donations did 0 impact other than saving Foundation X some cash.
Of the top of my head, there are a few obvious problems with the hidden-zero problem:
Foundations having more money isn’t necessarily a bad thing, especially if they give their assets away relatively quickly and effectively.
How are we supposed to know how much a certain real foundation like Open Philanthropy plans to give certain organizations?
Many charities don’t have such cut-and-dried budgets and fundraising goals. E.g., if GiveDirectly gets more money in 2022, it will simply give away more money by expanding the number of recipients and/or its geographical operations.
Regardless, Budolfson and Spears did a lot of fancy math to show the hidden zero problem is worth taking seriously in many cases, especially within EA.
All that being said, it’s not clear to me how the hidden zero problem impacts your claim here. On one hand, if we intentionally diversify funding sources, charities might raises their budgets and demand the same amount from big foundations. However, if these foundations see that more money is coming in from more donors, they might decide the charity/cause is no longer “neglected” and choose to reduce the size of their grant.
Would love to hear thoughts on this from people more deeply entrenched in the grant-making world...
Another argument against: Without sufficiently good coordination within the EA community, funding diversification can exacerbate unilateralist’s curse dynamics; making it more likely for any impressive, impactful intervention that is judged by some as net-positive to get carried out. Thereby weakening the ability of the EA community to differentially carry out net-positive interventions in anthropogenic x-risk domains.
Notably, it seems that within the FTX Foundation it was considered reasonable to approve grants after less than 2 minutes of thought (see this comment).
Could you explain further why funding diversity would exacerbate unilateralist’s curse?
Suppose there are 100 funders who are not coordinated (i.e. each of them funds things unilaterally), and there is some potential intervention for mitigating anthropogenic x-risks that any one of the funders can unilaterally decide to fully fund. If 99 funders think the intervention is net-negative but one funder thinks it’s net-positive and decides to fund it, the intervention gets carried out.
Some examples of interventions to consider here (neverminded whether any particular version of them is net-negative or net-positive):
an effort to draw attention to a certain low-profile domain of anthropogenic x-risks.
an effort to publish analyses about the most promising approaches for developing AGI.
an effort to create an impact market.
a certain outreach campaign.
a certain regulation advocacy campaign.
Thanks for clarifying. That helps me understand your concern about the unilateralist’s curse with funders acting independently. But i don’t understand why the OP proposal of evaluating/encouraging funding diversification for important cause areas would exacerbate it. Presumably those funders could make risky bets regardless of this evaluation. Is it because you think it would bring a lot more funders into these areas or give them more permission to fund projects that they are currently ignoring?
I meant to describe an argument against causing there to be more unilateralist funders in cause areas that involve anthropogenic x-risks.