Gavi Marginal Funding and Fundraising Support 2023

Executive Summary

This report is a historical, illustrative assessment, produced in October 2023, ahead of Gavi’s 2025 (Gavi 6.0) replenishment, of how cost-effective marginal philanthropic funding to Gavi would have been measured as a Social Return on Investment (SROI) in Coefficient Giving (CG) units. The 2025 replenishment has since concluded: Gavi raised roughly $9 billion against a ~$11.9 billion target, falling short largely because the United States zeroed out its contribution. We share the analysis as a worked example of the method and intuitions, but our estimates should be considered rough and not necessarily reflective of current cost-effectiveness.

To summarize our findings:

• We estimated, with moderate-low confidence, that Gavi would face a net funding gap of around $0.7 billion in 2025, driven more by rising fundraising targets than by donors withdrawing.
• We expected that, under a shortfall, Gavi would protect active commitments and instead slow or postpone less-committed rollouts, making vaccines in four categories the most likely candidates to be saved with marginal funding:
  • newly approved longlist vaccines
  • recently approved programmes such as the RTS,S
  • malaria vaccine
  • previously paused COVID-era vaccines, and stricter approval of new country proposals
• Using Gavi’s own cost and impact data, we estimated the SROI of marginal funding at roughly 200 to 1400x for the vaccine categories Gavi was most likely to treat as marginal at the time. These figures rest on 2023 data and several major assumptions. A larger or differently structured funding gap could move the relevant margin to larger or more cost-effective programs, so the estimates should be read as historical evidence about one plausible margin, not as a current estimate of Gavi’s marginal cost-effectiveness.

Introduction

This report, commissioned by Coefficient Giving (then Open Philanthropy) in 2023, examined Gavi, The Vaccine Alliance. At the time of its initial writing, this report did two things: 1) estimate how much of a shortfall might be expected ahead of Gavi’s 2025 replenishment (Gavi 6.0; Gavi, 2026) and 2) produce the cost-effectiveness model of marginal funding to Gavi in 2025. Since writing this report, various relevant funding decisions have been made (Ridgy, 2025), and Gavi has received $9 billion in commitments, significantly short of its target, mostly because of the zeroing out of US contributions under the Trump administration. We have therefore adapted this version of the report to focus on the cost-effectiveness analysis.

Our central question was: if a philanthropic funder gave additional money to Gavi around the 2025 replenishment, how cost-effective would that marginal dollar be? Answering it required three steps: a) settling on an expected budget shortfall, b) identifying which vaccines Gavi would most plausibly fund or scale back at the margin, and c) estimating the SROI of those marginal vaccines using Gavi’s own cost and impact data, converted into Coefficient Giving (CG) units. We drew on Gavi’s internal estimates, historical funding trends, and Vaccine Investment Strategy documents, supplemented by interviews with five global-health experts.

We centered the analysis on a roughly $0.7 billion funding gap. The shortfall that actually materialized was substantially larger, which could move the relevant margin to more cost-effective programs, potentially increasing both the scale and effectiveness of additional support. However, we have not updated the analysis to reflect this larger gap.

Before sharing our headline findings, a few caveats are in order:

• We have not updated our thinking related to these cost-effectiveness models. This was produced in 2023 for the 2025 replenishment window, and should therefore be considered an illustrative, historical example of the cost-effectiveness of marginal funding to Gavi.
• Gavi’s approach to shortfalls seems to be proceeding differently from what we assumed (e.g., headquarters cost-cutting, structural reforms, loans from multilateral development banks rather than entirely postponing or scaling back the rollout of certain vaccines).
• Finally, our exact inputs have not changed. Our estimates rest on the numbers we collected in 2023 regarding lives saved, vaccine prices, and costs.

Our main findings are as follows:

• We expected a large budget shortfall to be unlikely, but assumed that a $0.7 billion budget shortfall was possible. This estimate was highly uncertain.
• We anticipated, with medium confidence, that Gavi would slow down or postpone rollout for vaccines for which it has few commitments due to a funding shortfall, before cutting down on existing programs, suggesting marginal vaccines from several categories (COVID delayed, newly board-approved, long-listed future candidates).
• We estimated, with medium confidence, that the SROI for marginal funding to Gavi targeting the 2025 replenishment vaccines could be in the region of 200–1400x.

Table 1: SROI for marginal vaccines. Calculations and sources for numbers in this spreadsheet.

These figures rest on several major assumptions and simplifications and are highly uncertain; we flag the most important caveats throughout.

The remainder of the report sets out our main takeaways, explains how we identified the categories of marginal vaccines, and then details the SROI calculation process and assumptions for each category. Supporting detail and calculations are provided in the appendices.

We Estimated a Funding Shortfall of Around $0.7 Billion in 2025

(This section has been condensed to highlight how we arrived at our original estimate.)

We thought it likely that Gavi would face a net funding gap of around $0.7 billion at its 2025 replenishment, excluding investments in Vaccine Investment Strategy (VIS) 2024 vaccines. Our confidence here was moderate-low. The gap was driven less by major donors reducing their commitments, which we thought unlikely to fall by more than 10% of Gavi’s 2021–2025 budget in aggregate, than by Gavi’s fundraising targets rising to cover the (re)introduction of COVID-delayed vaccines, approved VIS 2024 vaccines, and other new vaccines.

We arrived at this conclusion from the following pieces of evidence:

• Gavi had stable funding over the past two decades: Generally, pledge fulfillment rates have been fairly stable overall, at least when looking at US, UK, and Gates Foundation figures, with total contributions being basically identical to pledged amounts from 2016 to 2023.^[3] While we have not tracked these pledges back to each replenishment target in turn, a quick look through recent replenishments indicates that Gavi did not have significant difficulty reaching its replenishment targets (Gavi, 2026).
• Exploring the three potential reasons for Gavi being underfunded:
  • Gavi and the Global Fund might have an overlapping replenishment round in 2025. We concluded that this overlap in replenishment rounds would have a limited impact on Gavi’s funding level, though this was with low confidence.
  • Changes in the commitments of major donors. We concluded that in a contracted fiscal space and post-COVID political shifts, global health donors would be likely to pull back some funding. We estimated that the UK might reduce its pledge by 30%. By contrast, we believed that contributions from the US and the Bill & Melinda Gates Foundation would remain stable, with the Foundation possibly stepping in to cover specific vaccination gaps if needed.
  • Gavi might need to catch up on lost ground due to COVID. Our best guess was that doing so would require fundraising targets that are significantly higher than current funding levels. We think this poses the biggest risk for a funding shortfall in the next replenishment period, based on donors failing to increase funding levels to meet these raised targets.
• Projected funding gap estimates:
  • Excluding COVAX funds, Gavi projected a funding shortfall of $2.1–2.5 billion over the 2026–2030 period. Meanwhile, it expected to have about $2.6 billion in COVAX funds left at the end of 2025 (Gavi, 2023). An expert told us it was unlikely Gavi could secure additional funding so long as it had some remaining COVAX funds.
  • With low confidence, we assumed it was in Gavi’s interest to overestimate the funding shortfall in its reporting. Therefore, we projected a base shortfall of $2.1 billion (the lower bound of its estimate), excluding COVAX funds and newly approved vaccines from the longlist in VIS 2024/2025.
  • If true, the $2.6 billion in unspent COVAX funds could be used towards Gavi’s core programs, which would mitigate its projected shortfall. While we thought it likely that some of the COVAX funds would become available for Gavi’s core work, we were very uncertain exactly how much. Assuming roughly half of these funds (approx. $1.4 billion) to become available, we estimated that Gavi could face a net funding gap of $0.7 billion ($2.1 billion – $1.4 billion), excluding investments in VIS 2024 vaccines.

Given a Funding Shortfall, We Expected Gavi Would Prioritize Active Commitments and Slow or Postpone Uncommitted Rollouts

(This section has been condensed to highlight how we arrived at our original conclusion about marginal vaccines.)

To understand the cost-effectiveness of future contributions to Gavi, it is critical to identify the marginal vaccine: the programs most likely to be cut or delayed during a funding shortfall. If underfunded, we expected Gavi to prioritize its active commitments and maximize overall health impact, opting to slow or postpone uncommitted rollouts rather than canceling existing programs. Our sense was that the most vulnerable funding areas, from most to least likely to be cut, are: 1) postponing newly approved longlist vaccines, 2) scaling back recently approved programs (like the RTS,S malaria vaccine or Ebola/Marburg preparedness), 3) delaying previously paused vaccines (such as RSV and MMCV), and finally, 4) applying stricter approval criteria for new country proposals for currently supported vaccines like Typhoid and oral cholera vaccine (OCV).

How Gavi Funds Vaccine Programs

To understand what we consider marginal vaccines, it is useful to understand how Gavi funded vaccine programs at the time of our analysis.^[4] Gavi supports a list of specific vaccines that is updated every five years through a prioritization process called the Vaccine Investment Strategy (VIS). It entails compiling a longlist of possible new vaccines to support. From this list, the board selects a few vaccines that Gavi will endorse based on several measurements, including value-for-money (which is the term Gavi uses for cost-effectiveness). Gavi might decide to approve fewer vaccines from the list, which corresponds to the first category of “marginal vaccines” described below.

For the vaccines that Gavi decides to support, an Independent Review Committee (IRC) approves or denies proposals from eligible countries to start vaccination programs, based on metrics such as technical soundness and feasibility. If Gavi experiences funding constraints, there is an extra step in which it can prioritize proposals based on criteria, including value-for-money of the country programs (Gavi, 2018).

An expert shared in an interview, with low confidence, that it is likely that slowing down the rollout of approved vaccines across countries would be a way that Gavi could cut spending. They added that it is unlikely that Gavi would cut down on programs that are already running or that it has explicitly committed to, and that the prioritization would most likely be based on health needs, similar to the decisions Gavi makes when vaccines are supply-constrained. For malaria, this framework prioritizes areas in which the most lives can be saved, among other factors (Gavi, 2023). We did not have time to look into that decision process in more detail.

We found a prioritization mechanism (Gavi, 2026) for Gavi’s new vaccine support proposals that was meant to be applied “when Gavi resources are insufficient to meet demand in a particular application round.” This appears to have been last approved in June 2013, with a requirement to review the mechanism for a potential update “after its use in a funding shortfall or at the Gavi Board’s request.” (Gavi, 2013). This process has never been used due to there being no qualifying funding shortfall.

The criteria for prioritization are weighted as follows:

• Health impact; 30% (ratio of future deaths averted to total population from first five years of vaccination)
• Value for money; 30% ($/death averted)
• Financial stability of immunization programs; 25% (number of years for which a country has not fulfilled its co-financing commitment; percentage of spending on vaccines used in routine immunization financed with Government funds)^[5]
• Countries with greatest need; 15% (Gross GNI / capita)
• Equitable distribution of Gavi resources (1 application funded per country per round)

We are not confident that this approach or weighting would be exactly what Gavi used to determine how to allocate resources at the global level in a funding shortfall, but it does somewhat increase our confidence that Gavi will attempt to maximize health impact in its decisions.

How We Expected Gavi to Prioritize Given a Shortfall

We were uncertain about what Gavi would do if it were underfunded, but our best guess was that it would prefer to slow down or postpone rollout for the vaccines for which it had few commitments, before cutting down on vaccines that have many programs already running. Gavi could do this for one or several vaccines, depending on the size of the funding shortfall.^[6] Hence, we identified four categories of vaccines in order of least to most commitments, based on the table in Appendix C. We thought that Gavi might cut down costs in the following categories (including potentially cumulatively^[7]):

1. Postpone rollout of newly approved vaccines on its longlist
2. Slow down rollout of recently approved vaccines (in particular, RTS,S malaria vaccine, see explanation below)
3. Slow down rollout of the programs that had been paused in 2021 due to the COVID-19 pandemic and were re-approved in July 2023
4. Be stricter on new proposals by countries for other supported vaccines. (in particular, Typhoid and OCV, see explanation below)

While the above ordering is intuitively logical, we recognized that complex decision-making in an underfunding scenario could easily drive different approaches to cutting costs. We updated our original best guesses of the marginal vaccine based on expert interviews, indicating this input in the following sections.

Category 1: Postpone Rolling Out Any Newly Approved Vaccines on Its Longlist

We considered this most likely since no commitments or investments had been made regarding these vaccines yet. During the COVID-19 crisis, Gavi decided to pause or postpone the rollout of five out of six newly approved vaccines from the longlist, and we thought with medium confidence that Gavi would do the same if there was a funding gap in the new replenishment cycle (for the vaccines that have been approved in 2024). If we assume a similar process as in 2018, Gavi could reduce costs by about $360 million over a five-year period this way.^[8] The new longlist shown in the table in Appendix C contains eight vaccines, some of which probably would not have been approved even if Gavi had enough funds. An expert suggested that new programs would be most at risk, including Marburg and Ebola pandemic preparedness programs, but that human papillomavirus (HPV) would stay, as it was a priority for Gavi. For the rest, we were not able to find much information on Gavi’s plans for these programs, or data that we could use to calculate its SROI.

Category 2: Cut Down on the Hexavalent and Malaria Vaccines

This category contains vaccines that have been recently approved, namely malaria (RTS,S and R21, Gavi, 2023), the Hexavalent vaccine (ibid.), and a vaccine for COVID-19 (Gavi, 2023). Countries can start submitting proposals for these vaccines now or soon.

• The Hexavalent vaccine is a combination of the previous pentavalent vaccine and the inactivated polio vaccine. It is expected to cost Gavi up to $430 million during 2026–2030, “contingent on financial resources being made available for the next strategic period“ (Gavi, 2023, p. 13). It replaces two other vaccines, so the $430 million costs are higher than we would expect. Perhaps the reduction in costs for other vaccines has not been taken into account here. Moreover, an expert shared in an interview that they did not expect Gavi to cut down on this vaccine, because it is critical for polio eradication. All in all, based on the expert’s input, we thought Hexavalent was not a marginal vaccine. This was with low confidence since we lacked a good understanding of how Gavi weighs polio eradication compared to other programs. The COVID program was much larger, with a projected base expenditure of $9.9 billion for just 2024–2025, but we thought this was most likely earmarked money that Gavi could not cut down on (ibid, p. 12).
• The expected costs for the malaria vaccines were more uncertain; about $1.82 billion over five years (see the data and sources tab of this spreadsheet). With medium confidence, we thought Gavi would not want to postpone this program, because malaria has a high disease burden and the malaria vaccine is high-profile. That take was confirmed in an interview with an expert. However, based on what they said and based on our understanding of Gavi’s decision-making, we thought Gavi might cut down on the RTS,S vaccine, funding mostly the R21 vaccine if it had insufficient funds.^[9] Our main uncertainty was that Gavi might have had prior commitments for buying the RTS,S vaccine. We did not come across information about this, but found that Gavi was planning to invest in both vaccines (Joi, 2023).
• Finally, an expert indicated that it was likely Gavi would cut down on Marburg and Ebola pandemic preparedness programs. The board was asked to commit to a “a US$ 10 million time-limited envelope for a Global Virtual Pooled Inventory (GVPI) for Ebola Sudan and Marburg candidate vaccines” in June 2023 (Gavi, 2023). For the rest, we were not able to find much information on Gavi’s plans for these programs, or data that we could use to calculate its SROI.

Category 3: Slow Down Rollout of the Programs That had Been Paused in 2021 Due to the COVID-19 Pandemic

In light of the COVID-19 pandemic, Gavi paused the rollout of several new vaccines it was planning to fund: diphtheria, tetanus, and pertussis (DTP) boosters; rabies; hepatitis B birth dose; respiratory syncytial virus (RSV); and multivalent meningococcal conjugate vaccine (MMCV). In June 2023, Gavi announced that the rollout of the first three of these vaccines would begin again, and that the MMCV and RSV vaccine had also both been re-approved for rollout when they become available^[10] (Gavi, 2023). The delay in rollout meant that Gavi had not committed to specific country programs so far, making them a candidate for the marginal vaccine in case Gavi has to cut down spending. Gavi could have cut spending by approximately $360 million over a five-year period this way (the initial budget estimated for these vaccines, Gavi, 2018, p. 10). Since the RSV vaccine and MMCV that Gavi intends to support are not currently available yet, Gavi will probably have the least commitment regarding these two, making them the most likely marginal vaccines of this set (though again with low confidence).

Category 4: If Gavi Had to Cut Down Spending Even More, It Could Be Stricter on New Proposals by Countries for Other Supported Vaccines

We thought these cuts would mainly impact Gavi’s spending on Typhoid and OCV vaccines, since these were the two vaccines with the highest number of countries that are eligible but had not submitted a program proposal yet (Gavi, 2023, p. 86). The expected expenditures for Typhoid for 2021–2025 are $302 million (Gavi, 2021, p. 14). If we assume similar costs in the next replenishment cycle, and assume half of that is due to new proposals, this gives Gavi a saving opportunity of $151 million for Typhoid. Similarly, Gavi could cut roughly $142 million of spending on OCV by not approving new proposals during the new replenishment cycle.^[11] We were uncertain about the percentage of spending that Gavi could cut this way, in particular around the amount of spending it had already committed to, but we were reasonably confident that this was in the right ballpark.^[12]

In 2023, Our Best Guess SROI for Marginal Funding to Gavi Targeting the 2025 Replenishment Vaccines was Between 200–1400x

(This section has been reorganized to be readable to non-CG audiences.)

We express cost-effectiveness as a Social Return on Investment (SROI): impact per dollar spent, or (CG value generated) / (US$ of expenditure). We work with CG units of impact, defined as the value of giving $1 to someone earning $50,000 per year in the United States (approximately the median national income).^[13]

We use the numbers Gavi itself publishes to calculate the cost per DALY averted (in CG units) for each category of marginal vaccine. The underlying calculations and sources are in our cost-effectiveness model. We begin by considering the assumptions and procedures applied across all categories before moving on to the SROI calculation details related to each category of vaccine.

For the categories where Gavi publishes investment-case data, we took the cost-weighted average SROI across the vaccines in the category, weighted by the cost of each vaccine program (using the midpoint of Gavi’s upper and lower cost bounds), and calculated the SROI as if Gavi cut those programs fully, or proportionally to spend.

Assumptions Applied Across all Categories

• Haircut: We apply a 20% haircut to the lives-saved numbers Gavi reports. This figure is a low-confidence rough guess, based on an expert suggesting others had used a haircut of around 50% relative to academic researchers (to account for lower coverage and slower rollout than academic modeling typically assumes); our view that Gavi understands the situation on the ground better than academic researchers and likely produces more realistic numbers; and the fact that Gavi assumes that without it none of the vaccines it supports would have been administered, so some haircut is justified.
• DALYs per life saved: We treat a life saved as 51 disability-adjusted life years (DALYs) averted for under-5s and 32 DALYs averted for over-5s (CG DALYs), and value 1 DALY at $100,000 (CG conversion).
• Under-5 share: We assume 62% of lives saved are children under 5 (see Li et al., 2021, p. 403), which finds that in Gavi countries, 62 million of 100 million deaths averted across 10 specific vaccines are under-5s).
• Lives only, not disability: We count DALYs only for lives saved, not for averted disability. Including disability would make the calculated returns higher, but would require strong additional assumptions.
• Government counterfactual: We assume the money governments spend would otherwise be spent at an SROI of 40x (roughly a cash transfer to an average citizen in a low- or middle-income country). We capture this by netting out the DALYs governments could otherwise have averted with that money (at $100,000 per DALY) from the DALYs Gavi reports.
• Cost per DALY averted: We divide all of Gavi’s costs by the adjusted number of DALYs averted. We do not also count government costs, to avoid double-counting (since we have already subtracted the DALYs those funds could counterfactually have averted).
• Upper and lower bounds: Where Gavi reports lower and upper bounds, we carry them through, producing upper and lower bounds for the SROIs. These should not be read as confidence intervals, since we have not vetted how Gavi derived the bounds.

Table 2: SROI for marginal vaccines. Calculations and sources for numbers in this spreadsheet.

The headline range of ~200 to 1400x reflects the categories Gavi is most likely to treat as marginal (Categories 1 and 2). Category 3’s much higher SROIs are excluded because we think Gavi is unlikely to cut these established programs, and Category 4 was not calculated.

We thought, with medium confidence, that Gavi would attempt to maximize cost-effectiveness when deciding which vaccine program to fund (Gavi indicates cost-effectiveness is around 30% of its prioritization criteria; Gavi, 2024). That makes it more likely Gavi chooses a more cost-effective vaccine within each category, which lowers the likely marginal value of a CG dollar; we therefore tentatively suggest the SROI sits close to the lower estimates above, where these are available.

SROI Calculation Process and Assumptions by Category

Category 1: Postpone Rolling Out Any Newly Approved Vaccines on the VIS 2024 Longlist

We did not know which vaccines from the longlist Gavi would approve, did not have Gavi investment-case data for any of them (not yet available), and did not try to estimate from other sources. Our approach was therefore to assume an SROI similar to the VIS 2018 shortlist vaccines, but taking the 50% least cost-effective of that shortlist (meningococcal, RSV, OCV), for two reasons:

• Diminishing marginal returns: If these vaccines were very cost-effective, we would expect them to have been considered in the 2018 longlist, unless they were not available yet.
• Low burden of disease: This raises the cost of reaching a large number of people and lowers cost-effectiveness. We have not looked into the burden data for these specific vaccines; we hold this with moderate confidence, based on our own intuition and an expert’s comment that these vaccines do not have a high burden of disease. COVID might have a higher burden of disease, but will most likely have earmarked funding, as described in the previous section.

We then applied the cost-weighted method described above across these three vaccines. We have low certainty that this gives the right SROI; finding better estimates was out of scope, but could be pursued in follow-up.

Category 2: Cut Down on Recently Approved Vaccines

The marginal vaccines in this category are malaria, Ebola, and Marburg. We focused on malaria because Gavi had not published the relevant data for the other programs, and because Gavi could spend a large amount on the malaria vaccine, making it the most relevant for CG to gain some insight into.

We have modeled the SROI of the RTS,S malaria vaccine. Gavi has not published all the numbers we need, so unless stated otherwise, we base our assumptions on the Malaria vaccine market shaping roadmap (Gavi, 2023):

• $9.30 procurement cost per dose of RTS,S.
• Countries pay on average $1.51 procurement cost per dose (based on our own calculations, assuming a different funding trajectory for countries in each Gavi eligibility phase—see the model). This is a rough estimate of the price countries would pay.
• For operational and recurring delivery costs we assume the same per-dose cost as DTP boosters, given the large number of doses administered: $0.63 for countries and $0.08 for Gavi (Gavi, 2018, p. 22-23).
• We assume 30 million doses of RTS,S are administered with Gavi support each year, saving 24,000 lives yearly, 76% of them children under 5.

All in all, for RTS,S, Gavi would spend about $3.5 billion and countries about $1.0 billion to save 360,000 lives over the 2021–2035 period.

Category 3: Slow Down Rollout of the Programs Paused in 2021 due to the COVID-19 Pandemic

Gavi publishes investment cases with data for these five vaccines (hep B, DTP, meningococcal, rabies, RSV), so we can calculate the SROI using the assumptions above. We apply the cost-weighted method described above across these vaccines.

Category 4: Being Stricter on New Proposals by Countries for Other Supported Vaccines

We have not calculated the SROI for these vaccines, since we think it is unlikely that Gavi would cut down on them.

Comparison to the CG Bar

CG’s Global Health & Wellbeing (GHW) team identifies highly cost-effective opportunities as those with returns above 2,000 using their DALY valuations. Under our initial analysis, the categories Gavi is most likely to treat as marginal sit below it:

At the same time, more pressing funding shortages could make investment more attractive than this comparison suggests: a deeper shortfall both strengthens the counterfactual (these programs really would go unfunded without support) and pushes higher-SROI programs toward the margin, so the relevant marginal vaccine in a severe shortfall may look closer to Category 3 than to Categories 1 and 2.

Appendix A: Recent Global Fund and CEPI Funding Shortfalls of Limited Generalizability to Current Gavi Situation

[Medium-low confidence: This is an initial take based on a quick first look, as a more detailed look was deprioritized as a result of our reduced team capacity, but could be considered for follow-up work.]

• According to Ravelo (2022), CEPI met less than half of its funding goal in March 2022 replenishment (slightly over $1.5 billion against its $3.5 billion ask), though expected “additional funding commitments.”
• Reasons for the funding gap (Ravelo, 2022):
  • Canada and Saudi Arabia (which previously contributed) did not pledge anything.
  • CEPI and Global Fund replenishments took place in the same year (Global Fund half a year after CEPI.
  • “CEPI also lacks many of the community and civil society voices that compose Global Fund’s constituencies and are instrumental in engaging with and raising awareness about the fund’s works and fundraising. In addition, making the case to raise funds to prepare for what’s yet to come — which is what CEPI is doing — poses difficulties.”
  • “CEPI occupies a unique place in the global health infrastructure and raising funds for the [research and development] that we invest in presents a different set of challenges than raising funds for other global health deliverables, such as the procurement of medicines or vaccines. In these cases the donor knows exactly what they are purchasing and what the outcome will be. That proposition with R&D is very different,” Hatchett said.

Appendix B: Case Study of Previous Gavi Replenishment Within a Small Time Window of Global Fund

(This section relies on spreadsheets that are available upon request.)

We looked into Gavi’s first replenishment conference that, theoretically, should have overlapped with the Global Fund. However, Gavi’s first replenishment conference came in 2011 (Gavi, 2026), which did not overlap with the Global Fund’s third replenishment cycle Pledging Conference (held in New York, 4-5th Oct 2010). The Gavi Alliance held a “Call for Action and Resources” in New York on the 6th of Oct. (Gavi, 2019), which later led to Gavi’s first replenishment conference in June 2011 (Gavi, 2026). Given the timing and location, the overlap in stakeholders, and the involvement of the UN Director-General in both, we can safely assume this was an intentional and coordinated effort.

Based on Gavi’s reported funding data, there was no meaningful shortfall on Gavi’s end that year. In the 2010 replenishment, the numbers were as follows (amount pledged / amount contributed):

• Governments: $5947 million / $5531 million
  • Majority of the gap came from the International Finance Facility for Immunisation (IFFIm), which had pledged $1192mil but delivered $600 mil. In general, this is consistent with the IFFIm being the largest source of funding discrepancies. This has been true in either direction, e.g., when the IFFIm contributed to overfunding relative to the pledged amount in the first decade, as well as underfunding in the subsequent replenishments.
• Direct contributions: $3980 mil / $3984 mil

Reasons to update positively:

• Previous evidence of coordination and successfully reaching funding targets (on Gavi’s end, at least)
• Both are during periods that are just preceding the 15-year UN development goal cycles (2010 for the 2015 MDGs [Millennium Development Goals], 2025 for the 2030 SDGs [Sustainable Development Goals})
• This will be the first replenishment, after which the impact of COVAX will be known (claims 2.7 million lives saved; Gavi, 2023)^[16]
  • Based on $12.4 billion USD (Gavi, 2022) invested in COVAX, this comes to a $/life saved figure of ~$4600.^[17]
  • Based on the claim of 2 billion vaccine doses delivered (Gavi, 2023, p. 8), this is approximately 0.00135 lives saved per vaccine, or 741 vaccines per life saved.

Reasons to update negatively:

• This will be the first post-COVID replenishment, so funding needs might be higher due to programs or other progress delayed due to COVID.
  • Additionally, COVID may represent a funding need that is not reflective of “typical” funding needs, i.e., countries may be more willing to donate during a time of crisis.
  • Lastly, this was Gavi’s first-ever replenishment conference, so outcomes here may not generalize to 2025.

Appendix C: Vaccines in Several Stages of Gavi’s Approval Process.

The table below gives an overview of the vaccines Gavi currently supports, as well as the most recent longlist of vaccines it is considering to support. Sources: Gavi, 2024; Gavi, 2023, p. 86; Gavi, 2023. Note that the vaccines listed differ slightly across the Gavi website. These differences are partly because of new additions, and partly because there are several products and several names for specific diseases.

Appendix D: Using Global Fund Funding Trends as a Proxy for Gavi Trends

While looking into the Global Fund’s funding trends was out of scope initially, it was suggested as a potential proxy measure for Gavi in our conversation with CG. The Global Fund operates on 3-year replenishment cycles (The Global Fund, 2025), most recently in 2022, compared to Gavi’s five-year cycles. We have used this approach—on the basis that the Global Fund’s cycles might more accurately track recent funding trends—to gain a perspective on Gavi’s projections above:

• We have summarized major donors’ pledges and contributions to the Global Fund between 2000 and 2025 here: Global Fund pledge/contributions
• We crudely applied the proportion of pledges contributed to the Global Fund to Gavi’s estimate for expenditure approved (6.3 billion) + expenditure projected ($4.6–$5.2 billion) for 2026–2030 ($10.9–11.5 billion of expenditure total)
  • We generated upper and lower estimates based on the contribution proportion returning to the pre-COVID figure of ~94%, and continuing to decrease from the 2020–2022 level of 87% to ~80%, respectively.
• This gave us an estimate of Gavi’s qualifying resources using the global funds funding trend as a proxy in the range of $8.7–10.8 billion
  • This would give us a total funding shortfall of $0.7 to 2.2 billion for the 2026–2030 period if we exclude the COVAX funding overshoot.^[22]
  • This is comparable to Gavi’s internal estimates of $2.1–2.7 billion, or $0.73–1.33 billion.

While we have low confidence in this crude calculation and the oversimplifications of using the Global Fund as a proxy for Gavi funding trends (more below), it might suggest that Gavi is expecting a funding shortfall in line with recent funding trends of the Global Fund. We had a brief look at whether past Global Fund replenishments seemed to be a proxy for how GaviAVI did. Eyeballing the top-line of Gavi and Global Fund data on pledges/contributions over replenishment periods, Gavi seems to have done slightly better than the Global Fund and particularly better recently (2020–2025) period (100% vs. 87%, Gavi vs. Global Fund). However, an important caveat is that these are pledged vs. contributed funds, not replenishment target vs. pledged/contributed, the latter being a better indicator of underfunding of Gavi’s plans

We feel uncertain about the usefulness of Global Fund numbers as a proxy for Gavi. Some factors that drive this uncertainty:

• • Due to time constraints, we used the Global Fund’s “Pledges and Contributions Report,” which did not give us yearly breakdowns (The Global Fund, 2026). In particular, this makes it hard to disambiguate funding trends during COVID years.
  • The Global Fund has 3-year replenishment cycles instead of every five years (The Global Fund, 2025).
  • GF has received higher funding figures from the US (both pledges and actual contributions compared to Gavi). This figure is ~4x higher for the years 2017–2019,^[23] and ~30% higher for the years 2020–2022.^[24]
  • The Gates Foundation seems to contribute more (~$300+ million/year in the last seven years), and makes up a larger proportion (9–19%) of Gavi funding compared to the Global Fund ($150—$250 million/year in the last decade, and <5% respectively).

Appendix E: A Note on CG Units

We work with CG units of impact, defined as the utility of giving $1 to someone earning $50,000 per year in the United States (approximately the median national income). This serves as a baseline to which all income gains, regardless of where they occur globally, are compared.

To make universal comparisons, CG relies on a logarithmic utility function. This model acknowledges that the wellbeing derived from additional income is not linear: the more money one has, the less additional wellbeing is gained from each extra dollar. Under CG’s conversion formula, giving $1 to a person making $100 annually is therefore significantly more impactful than giving it to someone making $50,000. This model has two specific mathematical consequences:

1. An additional dollar is proportionally more valuable to a lower-income individual. For instance, a $1 increase for a person earning $500 per year represents a 0.2% raise (1/500 = 0.002). For a person earning $50,000 per year, that same dollar represents only a 0.002% raise (1/50,000 = 0.00002). Therefore, the impact of that single dollar is 100 times greater for the lower-income individual. (0.002/0.00002 = 100).
2. A fixed percentage increase in income generates the same value regardless of the initial income. Because the model values proportional gains rather than absolute dollar amounts, a 10% raise generates the same utility for a person making $100 as it does for one making $500. When converted back into CG units (baseline dollars), this value is constant. The calculation for both individuals is: 50,000 * ln(1 + 0.10) = ~4,765 CG units.

Finally, CG benchmarks DALY-denominated health gains against income. Specifically, each DALY averted is valued at 100,000 CG units. This allows for direct comparison between health and economic interventions. For example, if an intervention averts one DALY for a cost of $1,000, it delivers 100 CG units of impact for every dollar spent (100,000 impact / 1,000 cost). By using these inputs, we can consolidate deaths, DALYs, and economic effects into a single metric.

1. The same as lower estimate due to point estimates only being available for Gavi data for vaccines used to calculate this SROI, see here. ↑
2. This is largely driven by Hepatitis B which has a very high SROI. The value for the four other vaccines is 939 (lower) and 3,280 (upper). ↑
3. We are happy to share our calculations on request. ↑
4. A table containing an overview of the vaccines Gavi currently supports, as well as the most recent longlist of vaccines it is considering to support can be found in Appendix C. ↑
5. It is not made clear whether it would prefer to fund proposals that score high or low on these metrics ↑
6. If the funding gap is more than the costs for rolling out one vaccine (very roughly $60 million based on VIS 2018 $360 million total expected investment over a five-year period/six vaccines were approved Gavi 2018, p. 10), Gavi would have to cut down on several vaccines. An expert mentioned that which of these options Gavi chooses could depend on the size of its funding gap (interview notes). ↑
7. By this we mean that Gavi would first cut down spending on vaccines in the first category, and if that did not cover the gap, it would cut down on vaccines in the first and second category, etc. ↑
8. In the 2018 Vaccine Investment Strategy, five out of nine endemic disease prevention vaccines on the longlist were approved, for a total expected investment of approximately $360 million over a five year period (Gavi, 2018, p. 6-10). ↑
9. The RTS,S vaccine is the first malaria vaccine that has passed medical trials, and WHO recommended in October 2021 to start using this vaccine in Sub-Saharan Africa (Gavi, 2023). In April 2023, the R21 vaccine was first approved by Ghana, with the final results of clinical tests expected soon. Gavi wrote on its website: “the idea is not to replace RTS,S but to be complementary – Gavi has already approved funding for a malaria vaccine programme and is ready to support rollout of R21 alongside RTS,S.” (Joi, 2023). An expert shared in an interview that the R21 vaccine is expected to be about three times cheaper than the RTS,S vaccine, with a similar effectiveness. He also said that it would be available in larger quantities (interview notes). ↑
10. Gavi writes somewhat cryptically: “In addition multivalent meningococcal vaccine and RSV vaccine were also both approved in 2018, but are only reaching the stage where vaccines are available and approved for rollout.” (Gavi, 2023). ↑
11. The projected spending from 2020–2035 is $849 million. We assume a third of that is spending for the next replenishment cycle, and that Gavi has already committed to projects for half of that, resulting in $849 million / 6 = $142 million. ↑
12. An expert said that she does not think it is likely that Gavi would cut down on HPV, since it can save a lot of lives (interview notes). ↑
13. See Appendix E. ↑
14. The same as lower estimate due to point estimates only being available for Gavi data for vaccines used to calculate this SROI, see here. ↑
15. This is largely driven by hepatitis B which has a very high SROI. The value for the four other vaccines is 939 (lower) and 3,280 (upper) ↑
16. We could not find the source that discusses the methodology of this estimate. Sources just point to “independent modeling” by ICL (Gavi, 2023; Gavi, 2023, p. 8), though one source (Cullinan, 2023) claims it is based on Watson et al. (2022). ↑
17. Investment into COVAX / lives saved: 12,445.7 / 2.7 = 4609.5 (Gavi, 2022) ↑
18. This is a combination of the inactivated polio vaccine (IPV) and the pentavalent vaccine that protects against “diphtheria, tetanus, pertussis (whooping cough), hepatitis B and Haemophilus influenzae type b”(Gavi, 2023) ↑
19. RSV and MMCV “could soon see approved products on the market,” according to Gavi (2023). ↑
20. We have not researched why COVID-19 is on this list while Gavi is already supporting COVID vaccine programs, so we think this is probably about a new COVID product. ↑
21. The term “stockpile” refers to the fact that the vaccines are not administered in routine programs, but are available to prevent a potential outbreak from spreading further. ↑
22. Upper estimate = 94% of $11.5 billion expenditure total contributed to Gavi = 0.94*$11.5bn=$10.8bn ($11.5bn-$10.8bn= $0.7bn shortfall)Lower estimate= 80% of $10.9 billion expenditure total contributed to Gavi = 0.8*$10.9bn=$8.7bn ($8.7bn-11.5= $2.2bn shortfall) ↑
23. 2017–2019 contributions to GF/2017–2019 contributions to Gavi: $3700 / $(275+290+290) = 4.33 (this figure is $4.83 for pledges), details available upon request. ↑
24. $6475 / ($290+$4020+$580) = $1.32 (this figure is 1.67 for pledges); details available upon request. ↑