Note: This page summarizes the rationale behind a GiveWell agreement with Good Judgment Inc. Good Judgment Inc. staff reviewed this page prior to publication.
In a nutshell
In March 2025, GiveWell paid Good Judgment Inc. (GJI) $72,000 to provide forecasts about future U.S. government foreign aid funding levels, particularly for global health programs.
We're providing this funding because:
- The forecasts could help inform our grantmaking strategy given the changing U.S. funding situation.
- We view GJI's superforecasters as uniquely positioned to make well-calibrated predictions about complex policy questions.
- We think making the forecasts public could benefit others in the field.
Our main reservations are:
- The current political environment may make historical patterns less reliable for prediction.
- The gap between appropriated amounts and actual spending could be significant in the current political environment, making forecasting more challenging.
GiveWell is providing 50% of the funding ($72,000), with Open Philanthropy providing the remaining 50%.
GJI has already provided us with the forecasts, along with the reasoning underlying them. All forecasts can be found on their website here.
Note: This spending is structured as a service contract, not a grant. This means it is funded through our unrestricted funding, not via donations made to our All Grants Fund or Top Charities Fund. We have commissioned work like this throughout our history and will continue to do so for our own learning. We don’t always publish about service contracts. In this case, we think it is of particular value for the public to know about this spending.
Published: May 2025
Background
Understanding future funding levels for global health and development programs is important for GiveWell's grantmaking decisions. If we anticipate that funding for cost-effective interventions will be reduced, we may want to consider looking into those funding gaps in future grantmaking. We also may want to consider raising our cost-effectiveness bar for grants if we think there will be many more highly cost-effective opportunities in the near future.
The current political environment creates substantial uncertainty about future U.S. government funding for global health and development programs.1 Given that the U.S. is one of the largest bilateral donors to global health,2 significant changes to these funding streams would affect the landscape for our grantmaking.3
The organization
Good Judgment Inc. provides forecasting services using their network of "superforecasters":4 individuals with demonstrated track records of making accurate predictions about future events.5
GJI works with clients across various sectors to provide forecasts on complex questions.6 Their superforecasters are selected based on their demonstrated ability to make accurate predictions7 and continuously refine their approaches based on new information.8
What we think this funding will do
This $72,000 agreement funded Good Judgment Inc. to deliver six forecasts about U.S. government foreign aid funding, with a specific focus on global health programs. The forecasts were made by GJI's "superforecasters".9 The total project cost is $144,000. GiveWell is providing 50% of the funding and Open Philanthropy is providing the remaining 50%.
The project includes:10
- Two year-long "dashboard" forecasts that will be updated regularly:11
- US global health funding levels in FY2026 vs FY2024
- Total US foreign assistance in FY2026 vs FY2024
- Four one-time "snapshot" forecasts:12
- PEPFAR/HIV funding in FY2026 vs. FY2024
- Malaria/President's Malaria Initiative (PMI) funding in FY2026 vs. FY2024
- Global Fund replenishment (2026-2028)
- Gavi replenishment (2026-2030)
The "dashboard" forecasts will be updated regularly over the course of a year as new information becomes available, while the "snapshot" forecasts are one-time predictions.13
We focused on these questions because they cover some of the major categories of global health funding. We prioritized FY2026 to understand longer-term trends beyond immediate budget negotiations.
The project engaged 25 contracted Superforecasters to work on these questions, with additional superforecasters able to contribute informally through GJI's platform.14
All forecasts are publicly available on the Good Judgment website here.
The case for the funding
- These forecasts will help us understand whether and where to expect funding gaps in global health programs, which will inform our research and grantmaking priorities.
- Our impression is that GJI's superforecasters have a strong track record of making accurate predictions on complex policy questions.15 While we're already gathering input from subject matter experts, we think the forecasters may provide additional useful perspectives.
- Making these forecasts public could help other organizations working in global health better plan for different funding scenarios.
Risks and reservations
- We’re uncertain how useful these forecasts will be. The current political environment may make forecasting particularly challenging, since historical patterns (which forecasters often rely on) may be less reliable than usual.
- There are tradeoffs in how we structure the forecasting questions. While it's easier to get data on budgeted amounts, what actually gets spent may end up being quite different.16 We've tried to address this by having forecasters predict both budgeted amounts and likely spending levels through an additional survey, but this makes the forecasting task more complex.17
Plans for follow up
We plan to review forecasts as they come in. We’ve received initial forecasts here. For the dashboard questions, we will receive regular updates and alerts about major forecast changes. We may also decide to commission additional “snapshot” forecasts on other areas.
Internal forecasts
For this funding, we are recording the following forecasts:
Confidence | Prediction | By time | Resolution |
---|---|---|---|
20% | The forecasts will lead us to make at least one different grantmaking decision than we would have without them. | March 2026 | |
Our process
We provided this funding after conducting a shallow investigation. This included:
- Discussions with Open Philanthropy staff who had previously worked with GJI
- Working with GJI staff to refine the questions and methodology
- Conversations with staff from the Kaiser Family Foundation, whose data on US government spending will be used to resolve predictions,18 to better understand their budget tracking methodology and data availability
Sources
- 1
“The chaos began with a stop-work order for employees and contractors of the United States Agency for International Development and a freezing of all funds, including reimbursements for hundreds of millions of dollars already spent. That was followed by a process allowing organizations that provided lifesaving medical treatment and food aid to seek a waiver allowing them to continue their work.
Then came terminations, last Wednesday, of more than 5,000 projects and programs. Since then, some projects have been told they were fully restored, and others that they are restored only to the terms of their original waiver, which runs out next month. Almost none have seen any of the funds they are owed unfrozen.” Nolen, March 6, 2025. - 2
According to the Institute for Health Metrics and Evaluation, U.S. funding ($14.6 billion) made up 22.6% of total global health aid spending ($64.6 billion) for 2023 ($14.6 billion / $64.6 billion = 0.226). IHME, VizHub - Financing Global Health, 2023. See tab "DAH trends," which can be filtered by "United States" or "All".
- 3
For more context on the effects of changes to U.S. funding cuts, see our “USAID Funding Cuts: Our Response and Ways to Help” page.
- 4
"GJP research found compelling evidence that some people are exceptionally skilled at assigning realistic probabilities to possible outcomes—even on topics outside their primary subject-matter training. We call these people Superforecasters…Today, Good Judgment maintains a global network of elite Superforecasters who collaborate to tackle our clients’ forecasting questions with unparalleled accuracy." Good Judgment Inc., “About” page.
- 5
For information on the Superforecasters’ track records, see this GJI page.
- 6
See the “Who We Serve” section of their website here.
- 7
“We continue to grow this network by identifying and recruiting fresh talent from our public forecasting platform, Good Judgment Open.” Good Judgment Inc., “About” page.
- 8
"Central to the practice of Superforecasters is a mindset that encourages a continual reassessment of assumptions in light of new evidence." Good Judgment Inc., "Superforecasters’ Toolbox: Beliefs as Hypotheses."
- 9
For information on the Superforecasters’ track records, see this GJI page.
“For client-specific questions, Good Judgment recruits a core group of Superforecasters, who commit to working on those questions, before releasing the questions to our entire global Superforecaster network." Good Judgment Inc, "Custom superforecasts" page. - 10
Specific questions focus on the major funding sources for global health programs we support, including PEPFAR/HIV, Malaria/PMI, Global Fund, and Gavi:
- Q1: What will US Global Health Funding be in Fiscal Year (FY) 2026 as compared to FY 2024?
- Q2: How much will the US government obligate in total foreign assistance in Fiscal Year (FY) 2026 as compared to FY 2024, according to ForeignAssistance.gov?
- Q3: What will US PEPFAR/HIV funding be in Fiscal Year (FY) 2026 as compared to FY 2024?
- Q4: What will US malaria funding be in Fiscal Year (FY) 2026 as compared to FY 2024?
- Q5: How much funding will be contributed to the Global Fund to Fight AIDS, Tuberculosis and Malaria (Global Fund) in its 2026-2028 replenishment cycle?
- Q6: How much funding will be contributed to the Global Alliance for Vaccines and Immunization (Gavi) in its 2026-2030 replenishment cycle?
- 11
Dashboards allow superforecasters to update their predictions daily as new information becomes available.
“The Superforecasters post their initial probability estimates on Good Judgment’s private forecasting platform, share their sources and rationales with their colleagues, debate the merits of competing interpretations of the data, and then update their forecasts whenever new information or analysis justifies doing so.” Good Judgment Inc, “Custom superforecasts” page. - 12
"Snapshot" forecasts pose one-off questions to the superforecasters and lead to a final report for GiveWell. These won't be updated by superforecasters on an ongoing basis like the "dashboard" forecasts mentioned above.
- 13
"Between 7 March and 25 March 2025, a group of 24 Superforecasters provided their assessments and rationales on six forecasting questions, which were compiled into these reports. Forecasting on two of the questions, US global health funding and total foreign assistance obligations, will continue for a year." Good Judgment Inc., "Superforecasting the US Government’s International Health Funding."
- 14
"The formal task order had 25 Superforecasters contracted to work on the questions. However, another dozen or so have joined informally so far as well." Good Judgment Inc., email to GiveWell, 3/18/2025 (unpublished).
- 15
For more information on the Superforecasters’ track records, see this GJI page.
- 16
For example, even if the budget maintains similar line items to FY 2024, actual spending might be much lower due to impoundment, program terminations, administrative delays, or other executive actions preventing full disbursement.
- 17
"As for the expenditures for [questions] #1, #3, and #4, we believe our best option will be to ask Supers informally using our tool Delphineo where we are able to pose questions outside of the forecasting platform to gather those numbers along with any comments specific to that informal survey. (This is the same tool we use in our workshops.)" Good Judgment Inc., email to GiveWell.
- 18
"Resolution Criteria…The Kaiser Family Foundation (KFF) tracks US government spending on,
among other things, spending on international health care programs…The outcome will be determined using data as reported by KFF when first available for the whole of FY 2026." Good Judgment Inc, Superforecasting Global Health Spending, p. 21