From pledges to budgets: co-financing, control and the conditions for a sovereign health transition under the US bilateral health agreements in Africa
The commitments, budgetary demands and institutional conditions of the bilateral health agreements between the United States and 25 African governments.
On this page
Summary
Twenty-five African governments have agreed to finance a growing share of the health workers and medicines that US programmes previously paid for. In most of these countries the scheduled reduction in US funding exceeds the pledged increase in domestic financing, and the commitments of both parties are non-binding. The transition can strengthen national control over health financing, but only if the pledges are written into medium-term budgets and executed, health workers are absorbed on terms agreed in advance, financing for medicines is protected, community services are contracted, and African governments coordinate their positions on specimen and data terms. The United States can support a durable transition by disbursing on schedule and by publishing what it pays against each agreement.
Introduction
Development assistance for health fell by approximately half between its peak in 2021 and 2025, to an estimated US$39 billion, and US funding is estimated to have declined by 67% in 2025 alone (Apeagyei et al., 2025). The exposure of sub-Saharan Africa to this contraction is considerable, since external sources financed more than one third of health expenditure in 23 of the region’s 46 countries in 2021 (Apeagyei et al., 2024). In September 2025 the United States set out the America First Global Health Strategy, which proposed multi-year bilateral agreements under which recipient governments would co-invest in, and progressively take over, the commodities and frontline health workers previously financed by US programmes (US State Department, 2025). By October 2026, 36 such agreements, most of them in the form of memoranda of understanding (MOUs), had been signed, 25 of them with African governments (US State Department, 2026; Think Global Health, 2026).
The agreements have been interpreted in markedly different ways. The US government presents them as a route to “self-reliance” and to more accountable assistance (US State Department, 2025). The Government of Kenya has described the move to government-to-government delivery as a means of restoring its oversight of resources (Kenya High Court, 2025), and Kenyan health unions supported its agreement (Citizen Digital, 2025). Both positions are consistent with the use of country systems advocated by the Lusaka Agenda and the Accra Reset, although both of these frameworks also call for transitions that are planned and managed (Lusaka Agenda, 2023; Accra Reset, 2026). Critics, among them the leadership of the Africa Centres for Disease Control and Prevention (Africa CDC), civil-society organisations and academic commentators, have directed their attention to the conditions attached to the funding, particularly in relation to specimens and data, and to the pace at which costs are being transferred (Reuters, 2026; Public Citizen, 2025; Bollyky, 2025; Mulumba et al., 2026).
This article addresses three questions: what the signed texts commit each party to; what these commitments require of national budgets; and which institutional conditions are likely to determine whether the transition produces durable and domestically governed health financing. The analysis takes seriously the US argument that recipient governments should bear a larger share of the cost of their own health systems. Its normative reference point is the principle, articulated by the Accra Reset High-Level Panel, that countries “must own the three foundations of their health systems: the decisions, the money, and the data” (Accra Reset, 2026).
Analytical approach
Sovereignty in its legal sense is not in question here. Each government consented to its agreement, and consent is itself an exercise of sovereignty (Krasner, 1999). The concern of this article is with effective sovereignty, understood as a government’s capacity to decide, to finance and to know about its own health system. These capacities need not move together. A government may acquire greater control over financing while ceding discretion over policy or over data, a possibility long anticipated in critiques of donor-defined “country ownership” (Esser, 2014).
Drawing on the literature on transitions from Gavi, Global Fund and PEPFAR support, on the execution of health budgets, and on the political economy of health financing reform (Huffstetler et al., 2022; Ogbuoji et al., 2025; Piatti-Fünfkirchen et al., 2021; Sparkes et al., 2019), the analysis identifies six conditions for a transition that strengthens national control (Figure 1). Two relate to financing: whether the pledges can be funded, and whether they will be appropriated, released and spent. Two relate to decision-making: whether the state can employ the workforce, procure commodities and fund community services, and how the commitments are governed and scrutinised. One relates to data and markets, namely the specimen, data, regulatory and procurement terms that accompany the funding. The sixth, the predictability of US funding, bears on all of the others.
Conditions for a transition that strengthens national control
Grouped according to the three foundations of health sovereignty identified by the Accra Reset
The money
The decisions
The data (and markets)
The commitments set out in the agreements
The texts follow a common template. The United States undertakes to fund commodities and frontline health workers in 2026 and “a proportion” thereafter, while partner governments undertake to increase their co-investment and to assume responsibility for the workforce and for procurement (US State Department, 2025; Public Citizen, 2026). In the 15 published texts with five-year schedules, US funding in 2030 stands at a median of 47% of its 2026 level. Taken together, scheduled US funding in these texts declines from US$2.7 billion in 2026 to US$1.4 billion in 2030, whereas partner co-investment rises from US$0.5 billion to US$2.2 billion (Figure 2). The scale of the workforce transfer is considerable. Under Kenya’s schedule, the number of US-funded frontline posts falls from 28,668 in 2026 to none in 2028, and Kenya undertakes to finance 13,293 of these posts from 2028, leaving approximately 15,400 without a stated successor arrangement (Kenya–US framework, 2025).
Annual scheduled US funding and partner co-investment, 2026–2030
Amount scheduled for each year, summed across the 15 published agreements with five-year schedules, US$ billion (nominal)
The commitments of the two parties differ in legal character. Every published MOU states that it is “not an international agreement” (Kenya–US framework, 2025; Nigeria–US MOU, 2025), and the State Department has described the agreements to Congress as “formal, non-binding arrangements” (US State Department, 2026). US funding is “subject to the availability of funds” (Uganda–US MOU, 2025), as is any multi-year US commitment, since no federal agency may obligate funds before Congress has appropriated them (31 U.S.C. §1341). The specimen-sharing agreements are of a different character. They are published in the US treaty series and enter into force thirty days after signature, which indicates that the United States regards them as binding (US State Department, 2026; Nigeria–US specimen agreement, 2025). The funding commitments of both parties are therefore political in nature, whereas the obligation to share specimens is legal. The non-binding form also served the interests of partner executives; the Government of Kenya argued before its courts that the framework did not require parliamentary approval (Kenya High Court, 2025).
The rights of adjustment are asymmetric. Every text permits the United States to reduce or discontinue funding, in most cases where co-investment or performance falls short of the agreed schedule. Where a ratio is specified, it ranges from a reduction of two dollars for each dollar of shortfall (Uganda and Côte d’Ivoire), through a reduction of one dollar for each dollar, to a proportional reduction (Nigeria) (Nigeria–US MOU, 2025; Uganda–US MOU, 2025; Public Citizen, 2026). Reductions on a one-for-one basis are broadly comparable with Global Fund practice (Global Fund, 2026), whereas the two-for-one ratios, and the eight texts that permit reductions without any stated formula, go further. The texts also provide for performance payments of US$1–10 per person, and the State Department has informed Congress that funds may be reallocated “upward or downward” (Public Citizen, 2026; US State Department, 2026). Only the Nigerian text permits the partner government to revise its own co-investment in the event that US funding falls short (Nigeria–US MOU, 2025).
The strategy itself contends that less than 40% of US funding reached “frontline supplies and healthcare workers”, that government-to-government delivery entails “significantly lower management costs”, and that recipient governments “only cover on average 30-40% of the costs of the response” (US State Department, 2025). The State Department expects more than 80% of resources under the agreements to reach frontline services (US State Department, 2026). The objective of transition considerably predates the present administration and has been incorporated into appropriations legislation by Congress (GAO, 2010; KFF, 2026; House Appropriations Committee, 2026). The strategy also seeks to “promote American companies and American innovations abroad” (US State Department, 2025), an objective reflected in some of the texts.
The negotiated outcomes varied considerably. Six texts follow the template almost verbatim, seven contain targeted amendments, and those of Kenya, Nigeria and Uganda were substantially renegotiated. Zambia signed its agreement after the specimen- and data-sharing provisions had been removed (Associated Press, 2026), whereas Zimbabwe declined to sign on account of those provisions, and US health funding to Zimbabwe is being brought to an end (The Herald, 2026; Associated Press, 2026). These cases indicate that the terms were open to negotiation, although declining an agreement carried substantial consequences.
Fiscal feasibility
Annual pledges across the 25 signatories amount to US$2.1 billion, equivalent to 15.1% of their combined domestic government health expenditure in 2023. Annual US funding under the agreements amounts to US$3.1 billion, 35% below its level in fiscal year 2023, and this reduction is equivalent to 11.8% of the same expenditure. The aggregate position is, however, heavily influenced by two countries, since Nigeria and Tanzania together account for 45% of all pledges. In the median country, the pledge is equivalent to 9.2% and the reduction in US funding to 19.2% of domestic government health expenditure, and the reduction exceeds the pledge in 16 of the 25 countries (Figure 3).
Annual partner pledge and reduction in US funding, by country
Average of 2026–2030, as a percentage of 2023 domestic government health expenditure
Two qualifications are necessary. First, most US funding was spent outside government budgets, and the 2023 baseline includes technical assistance, management costs and programme areas, such as family planning, that the agreements do not finance. The reduction should therefore be read as a measure of scale rather than as a sum that national budgets must replace. Second, the comparison is sensitive to the way in which each text defines co-investment. The reduction exceeds the pledge in all seven countries whose pledges consist of new budget lines, but in only two of the seven whose pledges are defined as increases in total health expenditure. Across ten alternative specifications, the number of countries in which the reduction exceeds the pledge ranges from 15 to 18.
Annual averages also understate the position at the end of the term. Among the 15 countries with five-year schedules, the median pledge in 2030 reaches 27% of domestic government health expenditure and the median reduction in US funding 31.5% (Figure 4). The US funding that remains scheduled for 2030, a median of 14.5% of domestic government health expenditure and 96% in Malawi, represents a further exposure should it not be renewed. Exchange rates further affect the comparison; when domestic expenditure is expressed in 2025 US dollars, Nigeria’s pledge rises from 27% to 46% of its health expenditure, reflecting the depreciation of the naira (World Bank, 2026).
Partner pledge and reduction in US funding in 2030, by country
Fifteen published agreements, as a percentage of 2023 domestic government health expenditure
These pledges compete with other claims on the same budgets. The International Monetary Fund (IMF) projects Nigeria’s federal interest payments at 53.7% of federal revenue in 2026 (IMF, 2026), interest in Kenya has “consistently accounted for over 30 percent of revenue” (Kenya National Treasury, 2025), and Zambia reached agreement on a new IMF programme on the day after it signed its health agreement (IMF, 2026). None of the IMF documents reviewed refers to the agreements. Global Fund allocations are also declining and are partly conditional on co-financing (Global Fund, 2026; Business Daily, 2026), yet no agreement specifies whether the same domestic expenditure may be counted towards both. By way of comparison, Gavi’s co-financing requirements amount in most cases to less than 2% of government health expenditure (Gavi, 2025), whereas 23 of the 25 pledges exceed that level. Expressed as a share of total government expenditure, the median pledge would require an increase of approximately half a percentage point in the share allocated to health, and full replacement of the reduction in US funding an increase of approximately 1.3 points. Earmarked levies, such as Tanzania’s HIV levy and Zimbabwe’s AIDS levy, demonstrate that additional revenue can be mobilised for this purpose (Down To Earth, 2025; Bhat et al., 2016).
Budget credibility and execution
A pledge becomes credible only when it is incorporated in medium-term expenditure ceilings and is subsequently appropriated, released and spent. Approximately half of the signatories operate July–June or April–March fiscal years, so that each calendar year of an agreement spans two budgets, and the decisive allocations are made when ceilings are set several months before the fiscal year begins. Kenya’s draft 2026 Budget Review and Outlook Paper, which sets indicative ceilings for 2027/28, makes no reference to the agreement and reports that allocations for community health promoters in 2025/26 were not disbursed because of a shortfall in revenue (Kenya National Treasury, 2026). None of the 2026 and 2026/27 budget documents reviewed identifies co-investment as a distinct line. Nigeria’s Coordinating Minister of Health stated that the 2026 budget would capture “almost 346 million dollars” of additional co-financing (Nairametrics, 2026), a figure close to Nigeria’s commitment for the first year; the pledge, however, also extends to “Federal and State budgets”, which the federal government cannot itself appropriate (Nigerian government, 2025). Since budgets may include such expenditure without labelling it, the absence of a distinct line cannot be taken as evidence of default.
Budget execution presents a larger risk. Low- and lower-middle-income countries underspend their health budgets by approximately 13% on average, with non-wage items underspent to the greatest extent (WHO and World Bank, 2025; Piatti-Fünfkirchen et al., 2021), and Nigeria’s federal health ministry underspent its budget by an average of 35% between 2018 and 2021 (Olaleye, IBP, 2023). Audits by the Global Fund have found that co-financing which is not visible in national budgets is difficult to realise (Global Fund OIG, 2024). Where budget programmes are not mapped to the chart of accounts, allocations cannot be traced through to expenditure (Cashin et al., 2017). Pledges defined as increases in total health expenditure are the most difficult to verify, because they may be satisfied by reclassification or by reductions elsewhere in the health budget. Kenya’s estimates for 2026/27, for example, increased total health expenditure while reducing the allocation for immunisation from KSh15.6 billion to KSh6 billion (The Standard, 2026).
Delivery capacity
Salaries paid through the government payroll are the most reliably executed component of health budgets (Piatti-Fünfkirchen et al., 2021). The incorporation of health workers into the payroll, however, requires the verification of personnel, the adoption of schemes of service, the approval of posts, provision within wage ceilings, the alignment of remuneration and arrangements for pensions. In Kenya, where health is a devolved function, a recent precedent was agreed only in the course of a nurses’ strike, under which contract workers employed in national universal health coverage programmes are to be absorbed by county public service boards, with the associated costs reflected in revenue-sharing arrangements for 2027/28 (The Standard, 2026). Only eight of the 47 counties remained within the 35% ceiling on wage expenditure in 2024/25 (Radio Generation, 2025; Kenyans.co.ke, 2025), and Kenya’s agreement does not specify which payroll the absorbed workers will join. Earlier evidence is consistent with this concern. Absorption agreed in advance proved effective under Kenya’s emergency hiring plan (Gross et al., 2010), whereas a transition undertaken without a written plan raised vacancy rates to 25.9%, compared with 8.7% in facilities that retained support (Rodríguez et al., 2021).
Expenditure on medicines and supplies is discretionary, irregular and payable in advance, and it is consequently the first to be reduced when cash is constrained (Piatti-Fünfkirchen et al., 2021; Barroy et al., 2019). Kenya’s agreement transfers procurement and distribution to the national supply agency by the end of 2026 (Kenya–US framework, 2025), an agency for which the Auditor-General reported fill rates in 2024/25 of 79% for donor-programme commodities and 41% for essential medicines (Kenyan Wall Street, 2026). Governments may procure through the Global Fund’s pooled procurement mechanism using US financing (US State Department, 2026), and in July 2026 the Global Fund extended that mechanism to regional platforms and approved pre-financing for governments unable to pay in advance (Global Fund Board, 2026). Because most of these commodities are priced in US dollars, their transfer also shifts exchange-rate risk to national treasuries.
Community organisations delivered a substantial share of HIV services, including services for populations that governments find politically difficult to reach. None of the agreements refers to key populations, although those of Cameroon and Ethiopia commit to pre-exposure prophylaxis for persons who identify themselves as at risk (Public Citizen, 2026). A survey of PEPFAR implementing partners found that 72% of responding organisations serving key populations had closed those services (aidsmap, 2026). Social contracting, under which governments finance non-governmental organisations to deliver services, is an established mechanism for maintaining such services, but it has rarely been applied to them (Harm Reduction International, 2022), and where the populations concerned are criminalised it also exposes public officials to legal risk.
Governance and accountability
The ministries that signed the agreements varied, with finance or planning ministers signing for some governments and health or foreign ministers for others (Public Citizen, 2026; Government of Tanzania, 2026). Where the signatory does not control the budget, implementation of the pledge depends on a further negotiation within government. No text requires parliamentary approval. Legislatures will vote on the relevant appropriations each year, but they are not asked to approve the multi-year commitment or its adjustment clauses, which a finance ministry would ordinarily treat as fiscal risks. In Kenya, a petitioner challenged the agreement on grounds that included treaty-making procedure, public participation and privacy (Kenya High Court, 2025). The High Court suspended the agreement in December 2025, the Court of Appeal subsequently set that order aside pending appeal, citing the risk to patients, and as of 27 October 2026 the merits of the case remained undecided (Capital FM, 2026).
The accountability created by the agreements runs principally to the State Department and to Congress, although implementation plans were not shared with Congress before implementation began (Devex, 2026). Earlier experience suggests that texts alone provide limited assurance. By July 2012, 19 countries had signed PEPFAR Partnership Frameworks, but only four had concluded implementation plans (Institute of Medicine, 2013), and those plans were later found not to have provided “a way of holding stakeholders accountable” (HFG Project, 2017).
Terms of exchange
The non-financial terms of the agreements vary more widely than their financial provisions (Table 1). The specimen-sharing agreements require partners to begin sharing within five to eight days of a US request, remain in force for five to ten years, and preserve the right of the United States to use shared material after termination (US State Department, 2026; Nigeria–US specimen agreement, 2025). The United States justifies these provisions by reference to the need to characterise pathogens in the earliest days of an outbreak (US State Department, 2025). The principal benefit offered in the leaked model agreement is priority access to resulting countermeasures “immediately behind the U.S. Government’s domestic need” (US State Department (leaked draft), 2025), although Nigeria’s signed agreement omits that ranking (Nigeria–US specimen agreement, 2025). The variants accepted by the United States, which range from prior national approval in Mozambique to renewal conditional on a joint audit in the Democratic Republic of the Congo, indicate that rapid access and national control are not incompatible.
These provisions also interact with multilateral rules. WHO member states are negotiating the pathogen access and benefit-sharing (PABS) annex to the Pandemic Agreement, under which participating manufacturers would make 20% of real-time pandemic production available to WHO, with a deadline of May 2027 (World Health Assembly, 2025; World Health Assembly, 2026). Each specimen-sharing agreement provides that participation in a multilateral arrangement “shall not prejudice” the bilateral agreement (Nigeria–US specimen agreement, 2025). Since the United States has withdrawn from WHO (US HHS, 2026), the bilateral agreement would govern relations between the United States and a signatory bound by both instruments (Vienna Convention, 1969), whereas African member states have maintained that PABS materials should flow exclusively through the WHO system (Health Policy Watch, 2025).
The regulatory and commercial provisions also differ across texts. Reliance on the decisions of stringent regulatory authorities is accepted good practice, and the strategy itself supports the African Medicines Agency (US State Department, 2025). The most far-reaching provisions, however, require legal recognition of approvals by the US Food and Drug Administration by a fixed date. Preferences for US suppliers are difficult to reconcile with procurement at the lowest available price, whereas the exemption of US assistance from taxation reflects a long-standing requirement of US law.
Variation in selected provisions across the published texts
The 16 published memoranda of understanding and the related specimen- and data-sharing agreements
| Provision | Range | Examples |
|---|---|---|
| US reduction if co-investment falls short | From two dollars per dollar of shortfall to proportional reduction; eight texts permit reductions without a formula | Uganda and Côte d’Ivoire 2:1; Kenya, Ethiopia, Malawi, Mozambique 1:1; Nigeria proportional |
| Steps before a reduction | From none to joint remediation and phased reduction | Uganda: reductions a “last resort” after remediation; Cameroon: 180-day mitigation plan; Eswatini, Ethiopia: good-faith consultation |
| Reciprocity and incentives | Partner reciprocity in one text; performance payments in most | Nigeria may revise its pledge if US funding falls short; Botswana may withdraw if data are misused; US$1–10 per person incentives |
| Specimen sharing | Provided for in ten texts; absent in six | Kenya: testing under Kenyan law; Mozambique: national ethics approval; Nigeria: case-by-case committee; Democratic Republic of the Congo: renewal after joint audit |
| Data governance | Data agreements of up to 25 years; national ownership clauses in some | Uganda “data owner”; Cameroon “full rights over national data”; Kenya: ownership under Kenyan law |
| Reliance on US FDA approvals | From changing national law by end-2026 to reference only | Rwanda, Liberia: legal change; Kenya: routine products, after checking sameness; Cameroon: national authority retained |
| Commercial provisions | Absent in most; present in some | Uganda: purchase from US firms that manufacture locally; Rwanda: market access, “proof-of-concept market”; Kenya: support for American companies |
| Clauses required by US law | Present in nearly all | Tax exemption of US assistance; audit for compliance with the Helms Amendment |
| Parliament and audit | Mentioned in few texts | Uganda: parliamentary appropriation and joint audits; Sierra Leone: disclosure to Parliament |
Predictability of US funding
The appropriation for fiscal year 2026 provided US$9.42 billion for global health programmes and maintained five-year availability for PEPFAR funds (Congressional Research Service, 2026; KFF, 2026). For fiscal year 2027, the administration requested US$5.12 billion, while the bill passed by the House of Representatives would provide US$8.88 billion, available for three years; the Senate had not acted at the time of writing (Congressional Research Service, 2026; KFF, 2026). Fiscal year 2027 began under a continuing resolution to 11 December 2026 that precludes new programmes (Congressional Research Service, 2026). The Government Accountability Office found that the State Department, which assumed responsibility for the remaining programmes of the US Agency for International Development, “lacks complete and accurate data about foreign assistance awards, grants, and staffing” (GAO, 2026). Disbursement under the agreements could not begin on 1 October 2026 as planned, and a further bridging arrangement was expected (Bush Institute, 2026); no public confirmation of the first disbursements had been identified by 27 October. The first awards made directly to governments, for Kenya, Malawi and Uganda, are milestone-based and equivalent to between 7.5% and 28.5% of those countries’ funding for 2026 (Bass, 2026; Devex, 2026), and earlier milestone payments in Senegal were made three to four months after verification (MSI, 2018). The predictability of US disbursement is therefore as much a condition of success as partner co-investment.
Discussion and implications
The findings bear on three distinct questions. The first is whether governments can afford their pledges, which depends on revenue and expenditure choices that this analysis does not assess, although for the most aid-dependent signatories the arithmetic is unlikely to balance without phasing, bridging finance or a narrower range of transferred functions. The second is whether the pledges offset the reduction in US funding, and in most countries they do not, although national budgets need not replace off-budget overheads. The third is whether the pledges will be translated into services, which depends principally on institutions and is the question on which decisions in the next budget cycles will have the greatest bearing.
For African governments, the evidence points to a sequence of measures. Before the next budget ceilings are set, which for July–June budgets will occur by February 2027, finance ministries could co-sign costed transition plans, define co-investment in local currency, publish a single schedule that reconciles commitments under the agreements with those to the Global Fund and the IMF, and disclose adjustment clauses as fiscal risks. Within the budget, co-investment would be assigned a distinct code in the chart of accounts, and funds for medicines would be held in protected accounts with early releases and authority to carry balances forward, as Uganda’s Poverty Action Fund once protected priority expenditure (Foster and Mijumbi, 2002). The absorption of health workers should be agreed for each cadre before US funding for that cadre ends, specifying the employing level of government and providing for wages at least one budget cycle in advance (Gross et al., 2010; Zakumumpa et al., 2021). Where health is a devolved function, transfers to sub-national governments require multi-year frameworks and enforceable rules on release, as South Africa’s HIV conditional grant illustrates (South African National Treasury, 2026). Social contracting requires both budget lines and a legal basis (Gotsadze et al., 2019; Harm Reduction International, 2022), and supreme audit institutions could review co-investment within nine months of the end of each fiscal year.
For the United States, several measures would be consistent both with its stated commitment to accountability and with US law. These include the publication of implementation plans and of annual disbursements against each schedule, the inclusion of adjustment clauses operating in both directions, as in the Nigerian text, which commit no additional US funds, the publication of an eligibility framework for direct funding of governments, and the retention of multi-year availability of funds.
At the regional level, African governments could use the PABS negotiations to clarify the relationship between the annex and existing bilateral agreements. In the interim, Africa CDC and the African Union could publish minimum standards for agreements already concluded, covering national approval of each request, onward sharing, benefit sharing, data ownership and use after termination, which would serve as a floor for existing agreements rather than as a template for new ones.
More generally, the credibility of non-binding commitments rests on domestic institutions that raise the cost of reneging, notably appropriation, audit and parliamentary scrutiny (Lipson, 1991; GAO, 2010). It may further be hypothesised that the dependence that makes these agreements necessary also limits a government’s capacity to withdraw from them, and that each increase in domestic financing of core commodities strengthens a government’s position in subsequent reviews of the financial terms, though not of the specimen terms, which survive termination.
Decision points, 2026–2030
- Nov 2026 – Feb 2027Budget ceilings for 2027/28 set in July–June budget countries; 2027 budgets in calendar-year countriesIndicator of a credible transition: Pledges in the ceilings; penalty clauses disclosed as fiscal risks; a budget code created
- 11 Dec 2026US stopgap funding expiresIndicator of a credible transition: A full-year appropriation near 2026 levels, with multi-year availability
- Early 2027First disbursements and direct awards to governmentsIndicator of a credible transition: Payments reported against schedules; share of funding paid to governments
- May 2027World Health Assembly deadline for the PABS annexIndicator of a credible transition: Bilateral specimen terms consistent with the annex
- 2027/28 budgetsCo-investment steps up steeply in most textsIndicator of a credible transition: Coded appropriations; quarterly releases; medicine orders placed and paid
- 2027 – 2028Health-worker absorption (Kenya by 30 June 2028)Indicator of a credible transition: Posts approved and paid before US funding ends; no service gaps
- 2028Mid-term reviewsIndicator of a credible transition: Joint review using execution and audit data from both sides
- 2030Most agreements endIndicator of a credible transition: Domestic financing sustained in spending data; residual US funding renewed or tapered
Conclusion
The agreements transfer a growing share of the cost of HIV, tuberculosis, malaria and related services to African budgets. In the terms of the Accra Reset, they may strengthen sovereignty over financing, although the pace of transfer sits uneasily with the panel’s caution against abrupt withdrawal (Accra Reset, 2026). They contribute less to sovereignty over decisions, and their implications for sovereignty over data depend on terms that vary considerably from one text to another. Whether that potential is realised will depend on budgets that are appropriated and executed, on absorption arrangements agreed in advance, on protected financing for medicines, on the contracting of community services, on coordinated regional terms, and on a US partner that disburses predictably.
Note on data and methods. Documents and data are current to 27 October 2026. The analysis covers the 16 MOU texts published through US Case Act reporting or freedom-of-information litigation and compiled by Public Citizen (Public Citizen, 2026), the specimen-sharing agreements published in the US treaty series (US State Department, 2026), and official releases for the nine signatories without a public text. Pledges and US funding are annualised and expressed as shares of 2023 domestic general government health expenditure, as reported in the WHO Global Health Expenditure Database (WHO, 2025); the US baseline is each country’s fiscal year 2023 Global Health Programs allocation (US State Department and USAID, 2025). All amounts are in nominal US dollars. Ten sensitivity analyses test the definition of pledges, alternative US baselines (allocations, obligations and disbursements), the annualisation period and exchange rates. The principal limitations are that 20 of the 36 texts remain unpublished, that pledge definitions differ across texts, that some official amounts are stated as bounds, that budget evidence is national rather than sub-national, and that several statements are known only through press reports. A full-length version, with country data, analysis code and clause coding, is available on request.
Country data: pledges and US funding for the 25 signatories
| Country | Pledge type | Pledge, US$m a year | US funding under agreement, US$m a year | US allocation FY2023, US$m | Pledge, % of GGHE-D | Reduction in US funding, % of GGHE-D | Difference, points |
|---|---|---|---|---|---|---|---|
| Angola | Official release only | 10 | 14 | 52 | 0.8 | 3.0 | +2.2 |
| Botswana | Includes existing spending | 127 | 33 | 52 | 13.9 | 2.0 | -11.8 |
| Burkina Faso | Official release only | 21 | 29 | 68 | 4.1 | 7.4 | +3.2 |
| Burundi | New budget lines | 5 | 26 | 59 | 9.2 | 57.4 | +48.3 |
| Cameroon | Increase in total spending | 90 | 80 | 120 | 27.3 | 12.1 | -15.2 |
| Côte d’Ivoire | Increase in total spending | 90 | 97 | 158 | 8.1 | 5.4 | -2.7 |
| DR Congo | Official release only | 60 | 180 | 257 | 12.9 | 16.6 | +3.6 |
| Eswatini | New budget lines | 7 | 39 | 68 | 4.7 | 19.2 | +14.5 |
| Ethiopia | Increase in total spending | 90 | 203 | 224 | 9.2 | 2.1 | -7.1 |
| Guinea | Official release only | 10 | 18 | 42 | 5.5 | 13.0 | +7.6 |
| Kenya | Increase in total spending | 170 | 326 | 392 | 8.1 | 3.1 | -5.0 |
| Lesotho | Includes existing spending | 26 | 46 | 67 | 20.2 | 15.9 | -4.2 |
| Liberia | New budget lines | 10 | 25 | 51 | 17.8 | 44.9 | +27.1 |
| Madagascar | New budget lines | 8 | 27 | 69 | 5.3 | 26.9 | +21.6 |
| Malawi | Increase in total spending | 63 | 149 | 252 | 56.2 | 92.0 | +35.8 |
| Mozambique | New budget lines | 14 | 366 | 493 | 3.0 | 27.0 | +24.0 |
| Niger | Official release only | 14 | 21 | 54 | 8.8 | 19.9 | +11.1 |
| Nigeria | Increase in total spending | 595 | 416 | 563 | 27.3 | 6.7 | -20.5 |
| Rwanda | New budget lines | 14 | 32 | 108 | 4.1 | 22.2 | +18.1 |
| Senegal | Official release only | 5 | 13 | 77 | 1.5 | 17.6 | +16.2 |
| Sierra Leone | New budget lines | 9 | 26 | 49 | 17.6 | 45.4 | +27.8 |
| South Sudan | Official release only | 7 | 49 | 73 | 18.5 | 67.9 | +49.4 |
| Tanzania | Official release only | 360 | 260 | 523 | 47.1 | 34.4 | -12.6 |
| Uganda | Increase in total spending | 115 | 344 | 473 | 24.1 | 27.0 | +2.9 |
| Zambia | Official release only | 195 | 304 | 440 | 29.0 | 20.2 | -8.8 |
Annual averages over the funded years (five; three for Botswana and South Sudan), in nominal US dollars. GGHE-D is domestic general government health expenditure in 2023 (WHO). Malawi’s pledge sums the annual increases in its text; Mozambique’s counts named budget lines only; those of Lesotho and Botswana include existing expenditure; some official amounts are stated as bounds.
Systems and Sovereignty examines the financing, institutions, and evidence behind health policy in Africa. Published by EPIC Health Systems.
Systems and Sovereignty by email
New issues, sent when they are published.













