Our Planned Allocation to GiveWell’s Recommendations for the Next Few Years

Last year, we committed $350 million to GiveWell’s evidence-backed, cost-effective recommendations in global health and development.

This year, we’re committing $300M to support GiveWell’s recommendations over the next three years. This returns our annual support to its 2020 level ($100M/year). And it brings us to more than $1 billion in funding for GiveWell’s recommendations over the course of Open Philanthropy’s existence.

We’re committing all of this funding now, but we anticipate that GiveWell will spend it at their discretion in whatever way they think helps beneficiaries the most. As they explain in this post, they’ve been spending carefully in anticipation of a potential decrease in our funding; we’ve been discussing this with them actively as we developed our plan for the next few years.

The rest of this post:

  • Shares examples of recent GiveWell recommendations we’ve funded. (More)
  • Discusses reasons we’ve increased our bar for funding in our Global Health and Wellbeing portfolio. (More)
  • Explains how GiveWell’s impressive growth and our higher funding bar has influenced our plans for GiveWell funding. (More)
  • Shows why we think the marginal GiveWell funding opportunity will be at least as strong in absolute terms as in past years. (More)
  • Concludes with why we think GiveWell is still the gold standard for donors seeking outstanding evidence-backed giving opportunities. (More)

 

Where our funding to GiveWell’s recommendations ultimately goes

These are some of the most impressive grants GiveWell has recommended to us over the past two years:

  • $67.5 million to Helen Keller International to support vitamin A supplementation programs throughout West and Central Africa. These programs are inexpensive (~$2 per child, per year), reduce susceptibility to infections, and may be among the most cost-effective ways to reduce child mortality.
  • $61.4 million to Malaria Consortium to support seasonal malaria chemoprevention programs in Nigeria, Burkina Faso, Chad, and Togo. These programs distribute preventive antimalarial drugs to young children during the malaria season, in areas where transmission rates are high; they are a highly efficient way to combat malaria. GiveWell’s funding to Malaria Consortium represents a large portion of the global scaleup of SMC.
  • $13.2 million to Evidence Action to support its work providing technical assistance in Zambia and Cameroon to scale up syphilis testing and treatment in pregnancy. Evidence Action will work in partnership with governments to facilitate a switch from HIV rapid tests to dual HIV/syphilis rapid tests and to scale up syphilis treatment in routine antenatal care nationally.
  • $10.4 million to the Clinton Health Access Initiative (CHAI) to fund an incubator to scope and pilot potentially cost-effective interventions, aiming to leverage CHAI’s global footprint, established relationships with governments, and technical expertise to create new opportunities to support promising programs at large scale.

 

Our higher bar for Global Health and Wellbeing giving

We’ve reduced the annual rate of our funding for GiveWell’s recommendations because our “bar” for funding in our Global Health and Wellbeing (GHW) portfolio has risen substantially. In July 2022, it was roughly in the range of 1100x-1200x; we recently raised it to slightly over 2000x. That means we need to be averting a DALY for ~$50 (because we value DALYs at $100K) or increasing income for 4 people by ~1% for a year for $1 (because we use a logarithmic utility function anchored at $50K).[1]

The increase in the bar comes from a few sources:

  • Our available assets are down ~1/3rd from EOY 2021, which has pushed us to be more conservative with our spending (though our main funders still plan to spend down virtually all of their assets within their lifetimes).
  • Prompted by the asset decline, growth in our giving opportunities to combat global catastrophic risks (GCRs), and events in the world, we’ve been revisiting how we will allocate resources between our GHW and GCR portfolios. Our current plans are within the range we’ve communicated publicly in the past, but represent a lower GHW funding trajectory than we had internally anticipated over the last couple of years. We now expect our GHW spending over the next couple of years to be at roughly 2021 levels. We’ve had a lot of volatility on this front over the last few years, and much of our current thinking is focused on trying to make sure we preserve optionality over the near term.
  • Our push since 2019 to identify causes that “[combine] multiple sources of leverage for philanthropic impact (e.g., advocacy, scientific research, helping the global poor) to get more humanitarian impact per dollar (for instance via advocacy around scientific research funding or policies, or scientific research around global health interventions, or policy around global health and development)” has led to new programs in South Asian air qualityglobal aid policyinnovation policyeffective altruism with a GHW focus, and global health R&D. Those programs have significantly increased our ability to find high-ROI opportunities, which in turn means that for a given amount of GHW funding, the “bar” will be higher (because we have found more opportunities at higher levels).

 

How GiveWell’s impressive growth and our higher Global Health and Wellbeing bar influence our current plans

As we discussed last year, changes in our bar disproportionately affect funding for opportunities that are scalable and close to our bar:

“Some portfolios or categories of interventions will have many grants with cost-effectiveness near the bar. Those grants could be ruled in or out by small fluctuations in the bar, so our giving in those categories will be especially sensitive to changes in asset values. Others will have most of their grants far above or far below the bar, which means our giving in those categories will not be very sensitive to changes in asset values [or the bar].”

The marginal GiveWell recommendations tend to be more “elastic” (i.e., decline less in cost-effectiveness with more funding) and closer to our bar, for a couple of reasons:

  • GiveWell has grown and matured impressively over the last few years, and has built a strong base of donors: in April their median projection for funding raised in 2024 from sources other than Open Philanthropy was $421 million, compared to ~$117 million five years earlier. This makes GiveWell’s projected support from non-OP donors roughly twice as large as OP’s projected non-farm animal welfare, non-GiveWell GHW giving in 2024.
    • Having that funding available allows GiveWell to cover many of its highest-ROI opportunities regardless of our support. Given that independent funding base, even if the distribution of opportunities GiveWell was selecting from was the same as the distribution of other GHW opportunities, the existence of other donors would mean that our support for GiveWell was on the flatter part of the curve, and therefore should be more elastic than the underlying opportunity set. (Note that we think in terms of marginal funding. We expect GiveWell to find opportunities far above our bar each year, but as noted above, we expect those to be covered by other donors, such that our funding on the margin would be close to our bar. This isn’t as true for most of our other GHW areas.)
    • By contrast, we are a much larger portion of the funding available in many other areas where we work. Relative to the size of the opportunity space in farm animal welfare (to name one example), the amount of funding from other donors is quite low, and some of our highest-ROI opportunities would often go unfunded without our support.
  • Separately, as we’ve said before, GiveWell’s marginal recommendations are more often (though not exclusively) “scalable, commodities-driven global health interventions… than opportunities like R&D or advocacy that are more people-intensive (where we have a prior that returns tend to be more like logarithmic, which is more steeply declining [in cost-effectiveness with more funding] than our model of the GiveWell top charities).” By contrast, our other GHW programs tend to investigate opportunities involving research, advocacy, or other relatively inelastic activities. These tend to have high returns from initial funding, which more steeply decline with additional funding, even separate from their different funding bases.

So, reversing the order of the points above, we both think that the abstract GiveWell returns curve is flatter than in other spaces where we work, and even conditional on that, GiveWell’s large base of other donors pushes our marginal support into an even flatter part of the curve. Accordingly, as we’ve noted previously, our support for GiveWell’s recommendations should be (and has been) disproportionately responsive to changes in assets and the bar.

More mechanically, in the near term, we expect a small wedge between our estimates for our GHW bar and GiveWell’s bar, possibly on the order of 0-30% but with a wide margin of error. We think that means less than one might naively expect for a couple of reasons:

  • While a small difference in actual expected impact could obviously be very substantial at this scale, we don’t think you can take these expected value estimates literally, and we see a small difference in estimated/projected ROI as well within the margin of error of these estimates. And as we note below, our estimate of this wedge could shrink (or the sign could change) as we continue to engage with GiveWell.
  • GiveWell and GHW use our bars differently. Our perception is that GiveWell has more unified decision-making — it’s much closer to the platonic ideal of having a single list of grants, sorting it by estimated cost-effectiveness, and then funding grants until the budget is exhausted. For GHW, while centralized decision-makers review every grant, we think much more in terms of annual budgets for programs, and tend to be more (but far from absolutely) deferential to program staff within their budget envelopes. This is related to differences in grant size and number (GiveWell makes fewer bigger grants) and staff composition (OP has historically been more likely than GiveWell to hire subject matter experts as program officers). As a result, we think GiveWell’s bar binds its grant decisions much more than GHW’s does in practice. Additionally, roughly ~75% of our internal GHW spending projected for 2024 is based on budget commitments we made to program officers in prior years; the place where we expect the GHW bar in particular to weigh most heavily in the near term is around the decision on whether to launch new programs (which usually come with large multi-year budget commitments to be able to attract top candidates), and then over time it will also weigh in annual budget setting processes. Accordingly, we’d still expect there to be individual GHW grants over the next few years that OP leadership inside-view thinks are below this bar.

So our support for GiveWell’s recommendations should be more responsive to changes in the bar than our other GHW giving, and we expect a small wedge in estimated/projected cost-effectiveness between our bar and GiveWell’s. We’re making a large multi-year commitment in spite of that because:

  • As discussed above, we’re not sure how real the gaps in expected cost-effectiveness will actually turn out to be. We’re excited to continue collaborating closely with GiveWell over the next few years to get their critical engagement on assumptions we’re making, to see whether the current projected differences in fact manifest, and then, if they do materialize, to look for ways on both sides to reduce them and make sure that the best opportunities get funded regardless of organizational boundaries. As we undergo this process, it’s possible the wedge in marginal cost-effectiveness will turn out to be smaller than we think or even not materialize. But if the wedge turns out to be real and persistent, it’s possible this could be our final large-scale commitment to GiveWell’s recommendations.
  • We committed $350M to GiveWell’s recommendations last year. We feel an obligation to them and to the organizations they support given our long-term partnership; we want to make sure that we handle large changes in our funding trajectory in a reasonably gradual and responsible way. As we said above, we’ve been in conversation with them about these potential changes over the course of this year, and the prospect has shaped their gradual spending out of our 2022 commitment. But even with that, we didn’t want to overreact to a potential wedge in cost-effectiveness, and wanted to give GiveWell some support to help transition in the event that the wedge does turn out to be real and persistent.
  • Finally, given the volatility of the last few years and uncertainty about the future, we want to give GiveWell a few years of predictable support from us so they can plan effectively and so we have plenty of time to sync up on the first point above.
  • We think this will be an impactful grant, even if we aren’t yet sure how it will compare to other opportunities we could support with our uncommitted funding. GiveWell estimates that its top charities each prevent a death for around $5,000, so we expect the programs funded by this grant to save tens of thousands of lives, mostly among young children.

 

This means that the marginal GiveWell opportunity will be at least as strong as in past years

All of the changes in the GiveWell allocation here are coming from changes on our side. Both we and GiveWell expect that GiveWell’s cost-effectiveness margin over the next year or two will be as strong or potentially stronger than in past years. So for donors who have supported GiveWell in the past or been on the margin of doing so, we think the marginal GiveWell opportunity today looks at least as promising as it did last year or the year before.

 

Closing thoughts

Open Philanthropy began as a project of GiveWell, and one of us (Alexander) began his career there. Elie Hassenfeld, GiveWell’s CEO, serves on Open Phil’s board, and Cari Tuna, our main funder, serves on GiveWell’s. The two organizations continue to see our missions and outlooks as extremely aligned.

And in our personal charitable giving, both of us continue to use, and aggressively evangelize, GiveWell’s recommendations. When it comes to evidence-backed, scalable global health interventions, we don’t know of another resource for donors that is remotely comparable; GiveWell continues to set the gold standard in our eyes. Regardless of what happens with our respective bars over the next few years, we’re confident that GiveWell’s recommendations will remain a stellar resource for donors looking for outstanding giving opportunities, and we’re proud to be able to support their work.

Footnotes
  1. [1] Note that we and GiveWell express cost-effectiveness in different units. GiveWell compares opportunities to the cost-effectiveness of giving unconditional cash transfers to people living in extreme poverty, which it describes in multiples of “cash.”