GiveWell aims to save and improve lives as cost-effectively as possible. That mission has an urgency, and we put a lot of effort into finding and funding high-impact giving opportunities quickly. But we also want to maximize our impact over time, and have found that high-impact interventions can take years of investment to discover, vet, launch, and scale.

As a result, we’ve begun to deploy funds across a longer time period in order to (a) avoid a scenario where we want to make cost-effective grants but can’t due to lack of funds, (b) aid long-term planning for our research team, and (c) communicate consistent expectations to grantees and potential grantees about our cost-effectiveness threshold.

Specifically, we previously aimed to allocate all funds within the same year they were raised, targeting a year-end balance of zero. Now, we plan to enter each year with sufficient funds to fully cover our grantmaking activities for that year without accounting for new donations.¹We don’t expect to fully spend down our starting balance at the end of each year. This approach creates greater financial stability, which we think will allow us to plan better and to achieve greater impact over time.

If you donate to our Top Charities Fund (TCF), nothing has changed. We still expect to commit TCF donations in the quarter after they are received. These changes will only apply to our unrestricted and All Grants Fund (AGF).²A note on fungibility: this year, we expect to grant more to top charities than we hold in the TCF. In that scenario, we’d use flexible funds (e.g., the AGF). Under those circumstances, additional donations to TCF, which would fill part of that gap, would have the effect of freeing up some of those AGF funds to be used for other grants. In other words, the literal dollars you donated would have gone to a Top Charities grant and flow in and out of our bank account quicker, but the impact of your donation may be felt further in the future. Our best guess is that this will also likely be the case in 2025. Of course, if we raise more money for TCF than we currently expect or find smaller funding gaps than we currently expect, this note on fungibility wouldn’t hold. We’re significantly more uncertain about how this will balance out further in the future.

While the timeframe over which we intend to spend funds has changed, our goal to raise more funds remains. Raising more money is our top organizational priority, and the single most important lever to unlocking more impact over time. Huge needs remain and the more funds we raise, the more we’ll be able to help people in the future.

What’s the impetus for this change?

This is a fairly big change for us, but reflects GiveWell’s evolution over the past few years. We think it’s helpful to chart our organizational development to understand these shifts.

From 2007 to 2020, GiveWell’s work focused primarily on researching and recommending “Top Charities.” We announced them at the end of each year and recommended specific allocations for the public and for Open Philanthropy (e.g., X% to charity A and Y% to charity B). About 10% of our funding during this period went to non-Top Charities.

Starting in 2021, we became a more active grantmaker, significantly expanding the number of programs we supported and the amount of funding directed to those programs. For example, in 2021 we allocated about 37% of the donations we received to non-Top Charity programs. We also shifted to making most grants on a rolling basis throughout the year. This was possible during 2021 and 2022 because even though most donors still gave at year-end, funding committed by Open Philanthropy was sufficient to support any grants we recommended.

More recently, Open Philanthropy shifted its plans (as laid out in this update).³Although this announcement was released in 2023, Open Philanthropy and GiveWell discussed this topic in 2022. It committed $300 million to GiveWell over three years (2023, 2024, and 2025) with the intention of providing a few years of predictable support so we could plan effectively. Open Philanthropy has not committed funding past that point, and we are not factoring further funding from them into our plans in light of these updates; we’re grateful for this commitment as we shift to a grantmaking model (as laid out in this post) that is more compatible with unpredictable income.

What are some advantages of this change?

• We can engage in long-term planning. A lot of the work we’re doing today is building towards a stronger grantmaking machine for the future. This work involves supporting research that may inform our future giving (e.g., “value of information (VOI) grants”) and supporting the development of new programs via partnerships like the CHAI Incubator. We think these efforts will result in a pipeline of future giving opportunities that are more cost-effective than the opportunity set available today. For example, the $100 million we have directed to New Incentives (where we currently estimate the program’s cost-effectiveness ranging from 10x to 42x cash, our benchmark for comparing programs) is the result of pursuing a long-term research agenda: We first made a grant in 2014, funded a randomized-controlled trial of its program beginning in 2017, and named it a Top Charity in 2020.
• We can continue to prioritize the quality of our grantmaking. Conducting research deliberately improves the quality of our grantmaking. Entering the year knowing that we can fully cover our grantmaking agenda means that our research leaders can plan and staff investigations better to ensure we make good decisions about which grants to make or not make.
• We will be a better funding partner to grantees. Our grant investigations often take months to complete. It’s important that we’re able to communicate clearly with our grantees about our cost-effectiveness threshold and that the threshold does not change in the middle of an investigation. Additionally, our grantees plan their budgets and programs across multiple years, and having a stable cost-effectiveness threshold makes it easier for them to know whether a program is likely to have continued support.
• We will avoid a scenario where we can’t make the grants we want. We would like to avoid a situation where the grants we’ve invested in building can’t be supported.

How do other funders navigate this problem?

We don’t think many other funders have our exact set of constraints.

Many other big foundations have large endowments that they draw from, so they have a fairly predictable income stream. From that starting point, many of these foundations then choose to organize their grantmaking by program and allocate set budgets to each of them—for example, $X to the health portfolio and $Y to the education portfolio.

We’ve chosen to prioritize cost-effectiveness, meaning that we try to compare all possible grants on the same terms. We don’t pre-allocate budgets across our portfolios because we don’t yet know how each investigation will shake out. It may turn out that one portfolio moves significantly less money than originally planned because the programs they investigate don’t end up being that cost-effective, so that funding becomes available for other grant areas.

This is a challenging set of constraints to operate under, but we operate this way because we think it serves our mission best.

Where do we stand today?

We think maintaining a stable cost-effectiveness threshold will lead to greater impact. Instead of increasing or decreasing our grantmaking to meet annual donation fluctuations, we will use a funding buffer to maintain a stable threshold and smooth our spending over time. This will add calendar time between when we receive and disburse funds, but crucially will allow us to base our spending on what we actually receive, rather than on our projections.

Entering 2024, we had the following funds:

• $227 million that we held at the end of 2023 that came from funds raised in 2023 and before that weren’t granted out in 2023. This included $100 million from Open Philanthropy.
• $300 million from Open Philanthropy that it committed at the end of 2023 for three years of funding (i.e., for 2024, 2025, and 2026). We view this as an “exit grant” because we aren’t counting on Open Philanthropy supporting our recommendations in the future. We’re counting this in our funds available at the beginning of 2024 because, strictly speaking, it is, but we and Open Philanthropy agreed that it’s appropriate to spread these funds out over three years to smooth our spending trajectory.

This means we had a total of $527 million available for granting at the start of 2024. We anticipate that this will enable us to fully fund the grants we make during 2024, which we expect to total around $300 million. Currently, our best guess is that our spending will slowly increase over time, and we plan to draw down previously committed funds to cover those grants. If we raise more money than we currently expect, we’ll aim to speed up the pace of our spending.

How we could be wrong

It’s possible we’re wrong to adopt this new approach, and we should instead lower our cost-effectiveness threshold to increase our grantmaking to match the funds that we’ve raised to date. We took this consideration seriously and tried to quantify the trade-off.

Let’s say we have $200 million available to spend and use it to fund programs with a cost-effectiveness of 5x cash, which is very roughly equivalent to about $15,000 per life saved.⁴This is a very rough approximation. We calculate cost-per-life-saved differently than cost-effectiveness, due to cost-effectiveness including non-mortality benefits such as long-term income increases that are not included in the cost-per-life-saved. Therefore, two grants that we estimate at 10x cash may have significantly different cost-per-life-saved. In that case, $200 million would result in about 13,000 deaths averted. If we instead funded programs with a cost-effectiveness of 10x cash, very roughly equivalent to about $7,500 per life saved, the same $200 million in funding would avert around 26,000 deaths.

This is a hypothetical example, and we don’t take these numbers literally, but they’re informative because they provide a sense of what’s at stake. Because we expect that we will continue to identify funding opportunities at our current 10x threshold, we believe that holding funds to smooth spending over time and maintain this high cost-effectiveness threshold allows us to avert many more deaths than we might be able to avert if we used a lower threshold today to increase our grantmaking.

Still, this is a new model for us. We could be making mistakes both big and small. If it turns out that we raise significantly more funds than we expect, or that the opportunities we’re trying to build aren’t more cost-effective than today’s available options, we will revisit and adjust. And of course, we will share any new major shifts publicly because we know these considerations are important to your philanthropic planning.

What this means for donors

To reiterate: If you’re a GiveWell donor, you may be wondering what this shift means for you and the impact of your donation. As described above, we believe our new model will help us maximize the impact that your donations have. However, we have different funding recommendations depending on whether it’s important to you that your donation be deployed quickly.

• If it’s important that your donation be put to work as soon as possible after you make it, we recommend that you donate to our Top Charities Fund (TCF). We expect to commit TCF donations in the quarter after they are received. Our cost-effectiveness bar for the Top Charities Fund is 8x.⁵A note on fungibility: this year, we expect to grant more to top charities than we hold in the TCF. In that scenario, we’d use flexible funds (e.g., the AGF). Under those circumstances, additional donations to TCF, which would fill part of that gap, would have the effect of freeing up some of those AGF funds to be used for other grants. In other words, the literal dollars you donated would have gone to a Top Charities grant and flow in and out of our bank account quicker, but the impact of your donation may be felt further in the future. Our best guess is that this will also likely be the case in 2025. Of course, if we raise more money for TCF than we currently expect or find smaller funding gaps than we currently expect, this note on fungibility wouldn’t hold. We’re significantly more uncertain about how this will balance out further in the future.
• Conversely, if you trust GiveWell to decide where to spend your donation and on what timeline, we recommend our Unrestricted Fund or our All Grants Fund (AGF). The AGF supports opportunities that are more uncertain or riskier than our Top Charities, but have potentially higher cost-effectiveness, with a current bar of 10x. The Unrestricted Fund supports GiveWell’s operations, but will be granted out using the same principles as the AGF if our excess assets policy is triggered.

We hope this helps you understand our operations and giving options more clearly, and how they interact with your preferences. Raising more money remains our single most important lever for maximizing impact. If we have more funding, we’ll be able to make more grants to cost-effective programs that save and improve lives.