Photo by Frames For Your Heart via Unsplash
In 2011, when we began what would eventually become Coefficient Giving, we set out to explore a big question: If someone had billions of dollars to give, what should they fund to make the greatest positive difference?
Previously called Open Philanthropy, we started as a partnership between Good Ventures, the foundation of Cari Tuna and Dustin Moskovitz, and GiveWell, a nonprofit dedicated to finding and supporting evidence-backed charities that save or improve lives the most per dollar. Our goal was to expand GiveWell’s focus on cost-effectiveness beyond the areas that already had a lot of evidence, with a greater openness to cause areas beyond global health and to high-risk, high-reward opportunities.
From the beginning, our vision wasn’t just to maximize the impact of Cari and Dustin's giving, but to make the journey easier for other philanthropists who might want to follow a similar approach. We recently changed our name from Open Philanthropy to Coefficient Giving to signal our growing work with other donors. And after distributing more than $4 billion across a wide range of issues, now is an appropriate moment to reflect on some lessons we've learned about both practical tactics and philosophical approaches to doing the most good.
1. Choosing causes is the most important decision you can make.
The problems you work on as a philanthropist determine almost everything about your work: who you hire, who your peers are, the organizations in your scope of funding, the places you work, and more. Your programs become the streetlight under which you look for the keys of your impact. But despite their nearly all-encompassing importance, many philanthropists don’t take time to consider a wide variety of problems they could address with their philanthropy; they focus instead on problems near or dear to them. This is important because just as some specific tactics can be much more effective than others, some problem spaces can have systematically more cost-effective giving opportunities than others.
We’ve used the “importance, neglectedness, tractability” (INT) framework from our earliest days to guide our decisions about which cause areas to focus on:
- Importance: How many individuals does this issue affect, and how deeply?
- Neglectedness: All else equal, is this cause receiving less attention from other actors, particularly other major philanthropists?
- Tractability: Are there clear ways in which a funder could contribute to progress in this area?
Causes can vary hugely across these criteria, especially importance and neglectedness. For instance, many philanthropists overlook the importance of factory farming: There are more livestock animals living under miserable conditions in factory farms than humans alive today. An example of neglectedness is Strep A, a disease that causes nearly as many deaths as HIV/AIDS (which is itself underfunded), but receives around 100 times less funding for vaccine research and development.
In recent years, we’ve come to place greater emphasis on importance and neglectedness over tractability, because windows of opportunity that determine how tractable a problem is can open and close unexpectedly, and we’ve been surprised by the progress we’ve seen in causes that we initially expected to be intractable.
2. Field-building can pay off enormously, but it requires patience.
Over the years, we’ve observed that the first dollars into a truly neglected field can go especially far. We have often worked in areas where there aren’t many established funders, and as a result, we have taken several early bets to build up the number of people and organizations working in various areas.
For instance, in the early 2010s there was a broad academic consensus that many zoning laws artificially restricted housing construction, causing higher rents. But politicians did not treat zoning reform as a priority because it wasn’t salient to voters. In 2015, we began supporting the nascent Yes In My Backyard (YIMBY) movement to organize around these issues, including a grant that allowed Brian Hanlon, then a US Forest Service employee who moonlit as a volunteer, to go full-time on YIMBY organizing. Based on his initial experience in Sacramento, Hanlon went on to found California YIMBY, an advocacy group that helped secure a series of major victories this year. This includes perhaps the largest YIMBY victory of all time, SB 79, which legalized multifamily housing near high-frequency transit in California and could lead to the construction of hundreds of thousands more homes.
Another example is AI safety and security. In 2015, we thought that the risks of advanced AI were underappreciated, but there were very few people thinking deeply about these topics. We funded several research centers and talent pipelines for people to build relevant skills, develop research agendas, and start new groups. It was uncertain how effective this work would be at the time, but a decade later, the field of AI safety and security has grown enormously, and is better placed than it would have been otherwise to meet the demands from companies and policymakers. For example, we supported the program that would eventually become ML Alignment and Theory Scholars (MATS) in 2019, which has now trained hundreds of people in full-time 10-12 week technical research fellowships. Over 80 percent of these scholars have gone on to work in AI safety and security, including at governments and leading AI companies.
In both housing and AI, early investments in nascent communities ended up having a surprisingly large return on investment, showing that it can be a risk well worth taking.
Some of our largest regrets are not investing in neglected areas sooner. Lead poisoning is a good example of this: we missed several opportunities to fund this area faster even though it seemed very promising on our initial investigations. And movement building doesn’t always mean that every opportunity will be captured: We supported biosecurity field-building prior to the COVID-19 pandemic, and though this work has led to some meaningful accomplishments, the global COVID-19 response was deeply suboptimal, and we largely haven’t seen the necessary reforms implemented since.
3. High-impact philanthropy is possible anywhere on the risk spectrum.
When we first spun out of GiveWell and looked at grantmaking across a wider range of areas, I initially thought the risk profile of our most cost-effective grants would look like a barbell, being either low-risk (e.g. highly-studied interventions recommended by GiveWell, which have the virtue of measurability and tight feedback loops) or high-risk (e.g. hits-based opportunities that have a low probability of working but a massive pay-off if they do). Tight feedback loops allow you to learn, iterate, and generate confidence that your grantmaking is highly cost-effective. Conversely, the enormous potential upside of hits-based opportunities, especially grants focused on reducing global catastrophic risks, can more than make up for the lack of learning and adjustment over time.
At the time, I thought the in-between space was to be avoided because it had some risk and worse feedback loops, but not the upside to offset it. I’ve since come to think that high-impact grants can be made anywhere along the risk spectrum, including in the middle. The feedback loops are looser for many middle-risk-profile grants like policy advocacy and scientific research than for direct interventions like bed nets to prevent the spread of malaria, but the longer track record of other funders in global health research and development and the impact we’ve been able to have in policy fields like housing suggest these can be competitive with the cost-effectiveness we’ve seen from selected direct global health charities.
4. Balancing moderation and maximization is hard.
The causes we work on each represent an extraordinary opportunity to do good: In global health, highly studied interventions can save a life for around $5,000; for animals, a single dollar can avert 10 years of suffering; and many promising initiatives are trying to reduce overall risks to the survival of humanity.
Any one of these considerations can be all-consuming: If you can really save a child’s life for $5,000, how can you take that vacation? If a single dollar can reduce years of suffering for animals, or if your giving could move the needle on literally saving humanity, how can you do anything else? And how do you choose between these different compelling goals?
In a world of scarcity, the logic of maximizing cost effectiveness pushes to have some shared unit of comparison, so we can say how to split our budget between helping animals and humans, or how many lives today are worth some degree of risk to the future of humanity overall. Most potential answers to those comparisons will lead to all-or-nothing allocations: If you think averting a year of animal suffering is at all similar in value to a year of human life, you’ll end up only working to avert animal suffering; if you think no amount of animal suffering could compare to a year of human life, you’ll do nothing.
Within our portfolio areas we try to have shared units so we can optimize as much as possible. But for these highest-level philosophical comparisons between very different types of impact, instead of trying to choose one imperfect unit to use to compare everything, we instead try to divide up the portfolio and budget more based on how much weight we put on the underlying worldview or value in question.
As my one of cofounders articulated years ago, truly maximizing for any one thing is perilous: "If you’re maximizing X, you’re asking for trouble by default. You risk breaking/downplaying/shortchanging lots of things that aren’t X, which may be important in ways you’re not seeing. Maximizing X conceptually means putting everything else aside for X—a terrible idea unless you’re really sure you have the right X."
So, even though we strive to be rigorous and quantitative in our giving in order to deliver as much impact as possible for our beneficiaries, some of our toughest and most uncertain decisions end up being these high-level philosophical judgment calls that we don’t expect to ever fully resolve. We’ve always tried to strike a balance between maximizing impact on some specific frame and maintaining a commitment to pluralism across worldviews, and I expect that specific balance to continue to evolve as we learn more.
Looking Ahead
We still have many questions about the best ways to give to help others, but we strive to act quickly in the face of uncertainty while learning and iterating over time. Many of our early wins took time to mature, and we’ve learned important lessons from bets that didn’t succeed.
However, giving with urgency is still too rare. According to data from Forbes, nearly three-quarters of billionaires have given away less than 5 percent of their wealth. In some fast-moving fields like AI or biotechnology, there’s a narrow window where new funders can play a crucial role in reducing risks. And for other immediate challenges like global health, there are scalable and shovel-ready initiatives that could save millions of lives in the coming years. My hope is that sharing these lessons will inspire others in a position to give to do so strategically and ambitiously, and to share their own lessons in turn.
Read more stories by Alexander Berger.