Thoughts on EA Funds

Hopefully helpful feedback

Photo by Diego PH on Unsplash

I spent two afternoons reading through information about every effective-altruism-related fund that I could find, and this has left me with thoughts. (“Fund” here is a term of art for when you donate to specific grantmakers, and they donate the money as they see fit.)

Make Basic Information Easy To Find

Here are some things that I think should be trivially easy to discover about a fund:

• The names of the grantmakers (if the fund has grantmakers).

• Why the grantmakers are qualified to make decisions about where money should be given.

• The amount of money that the fund has to work with.

• The last time the fund made a grant.

• What kind of grants the fund makes.

• Whether it is in fact possible for small donors to donate to the fund.

Not only should this information be available and reasonably up-to-date, it should ideally be on the homepage of the fund. I shouldn’t have to click through several pages to answer questions like “is this fund actually still making grants?” Further, the information should be consistent across the website: if a grantmaker has left the team, every page that describes the team should be updated to remove the grantmaker.

Write Up A Theory of Change

Funds usually do a good job of distinguishing themselves from other funds under the same umbrella organization. For example, Founders Pledge clearly explains the difference between its Climate Change, Global Catastrophic Risk, Global Health & Development, and Patient Philanthropy Funds—including things that might commonly confuse people, like whether the Global Catastrophic Risk Fund makes climate change grants (it doesn’t).

Funds generally do a very poor job of distinguishing themselves from funds under different umbrella organizations. For example, it’s mysterious how Founders Pledge’s Global Catastrophic Risks Fund differs in approach from Longview Philanthropy’s Emerging Challenges Fund or EA Funds’s Long-Term Future Fund. A low-information donor hoping to outsource their judgment to someone else might as well roll a die to figure out which one to pick.

Giving What We Can has begun a program of charity evaluator evaluation. Their writeups explain, in detail, each fund’s strategy. I am a huge fan of Giving What We Can’s work here and relied on it heavily for my own post. But these evaluations are nowhere linked on the webpages of the funds they’re evaluating. You have to already know about Giving What We Can’s charity evaluator evaluation program to know to check there.

Some funds do detailed writeups explaining their grants: for example, GiveWell writes thousands of words explaining each grant from their All Grants and Top Charities Funds. However, this is an unrealistic expectation for funds that make small grants or funds whose grantmakers are working on them part-time: writeups could easily double the amount of time spent per grant.¹ Grant writeups may also be an unrealistic expectation for funds making certain kinds of grants: for example, funds that grant to individuals whose privacy might be violated, or funds that grant in areas like nuclear risk or biorisk where grantmakers may have made use of information that shouldn’t be shared publicly.

Detailed grant writeups are also not necessarily very helpful for informing donors. If you were the kind of person who read all of GiveWell’s writeups about why they prioritize vitamin A supplementation over iron fortification, you’d probably be a high-information donor who didn’t want to use a fund in the first place. Further, it’s easier to figure out a fund’s general principles and how it reasons if it tells you, rather than having to extrapolate it from a list of grants.

My proposal is that grantmakers should write detailed and explicit theories of change for their funds. What kinds of grants do they specifically make? How does their approach differ from the standard effective altruist approach, and from the approaches of other major donors in the area? What is their position on major debates? Why do they think their approach works better?

For example, Giving What We Can’s writeup explains that the Long Term Future Fund tends to make smaller grants to individuals and newer organizations, and to prioritize grants that no one else would fund, while the Emerging Challenges Fund tends to make larger grants to bigger organizations.

This information can be somewhat discerned from their respective webpages. “You have a preference for supporting more established organizations” is on the list of reasons that someone might not donate to the Long Term Future Fund. But much of the list is made up of items that apply to every longtermist fund, such as “you don’t think that future or possible beings matter, or that they matter significantly less” and “you have identified projects or interventions that seem more promising to you than our recommendations.” An uncareful reader could easily miss it—and many people have.

Conversely, a good example is The Life You Can Save’s Maximize Your Impact Fund:

Extreme poverty encompasses various characteristics that profoundly impact individuals’ ability to meet their family’s needs and to lead the lives they aspire to. The Maximize Your Impact Fund is a reflection of our commitment to addressing multidimensional poverty, with characteristics identified in the Global Multidimensional Poverty Index (MPI). The charities benefiting from the Maximize Your Impact Fund cover a wide range of interventions in health, education, living standards, and barriers to women and girls’ well-being. When you donate to the Maximize Your Impact Fund, you support their efforts in having an impact across all the indicators of multidimensional poverty.

A table then explains what specific factors make up multidimensional poverty.

Reading this paragraph, it’s easy for me to understand the difference between GiveWell’s All Grants Fund and the Maximize Your Impact Fund. The All Grants Fund tends to prioritize global health (broadly construed) and makes only a tiny number of grants outside of that area. Conversely, the Maximize Your Impact Fund is attacking the problem of global poverty from multiple directions. To be sure, this writeup isn’t perfect—I’d like more discussion of why the Maximize Your Impact Fund is adopting the multidimensional poverty approach instead of a global health focus. But, in general, I think the Maximize Your Impact Fund is a good model for communication to donors, which I think more funds should follow.

Find More Diverse Grantmakers

Only about a quarter of grantmakers at effective altruist funds are women. While I don’t have the capacity to check, I suspect the numbers are equally dire with regards to people from developing countries, nonwhite people, political conservatives, people who didn’t go to an Ivy League or Oxbridge university, and people with resumes that are just, you know, kind of weird.

Diversity is not a nice-to-have. In a position like grantmaker, where people have to exercise judgment in situations of severe uncertainty, it’s a major source of bias if everyone thinks the same way. Here are a few hypothetical examples of things people might miss due to lack of diversity:

• An all-male grantmaking team underestimates the cost effectiveness of a family planning charity, because they don’t realize how much it limits a person’s opportunities to have an unwanted pregnancy or more children than they can care for.

• A grantmaking team from the Anglosphere approves a grant to a shrimp welfare charity in Vietnam that takes a culturally insensitive approach, thus poisoning the well for future animal advocates.

• A liberal grantmaking team has no connections among conservative activists, and so fails to reach out to a conservative think tank that would be sympathetic to an AI pause.

• Elite universities cause their students to think in a particular way, which means that a grantmaking team that’s all from elite universities misses some crucial consideration.

In my opinion, people hiring for grantmaking positions should tend to favor hiring people whose backgrounds are unusual for their organization, so that they can get a diversity of perspectives. I also think in this—as in so many other areas of effective altruism—a major cause of lack of diversity is marginalized people self-filtering out before they even apply, because of imposter syndrome. The problem should be attacked from both angles.

1. If you doubt this: consider just the amount of time it would take to get approval for your writeup from the grantee.