▣ DATA BRIEF · GLOBAL HEALTH AID
☀ Key Stats
◆ Health-sector foreign aid fell by an estimated $4.2 billion, or 13%, in 2025, according to a model of donor cuts and bilateral aid flows.
◆ Emergency-response aid declined by an estimated $5.5 billion, or 21%.
◆ Aid for government and civil society recorded the largest estimated dollar loss at $10.2 billion, a decline of 29%.
◆ Overall, an estimated 130 aid-recipient countries lost a combined $25.9 billion in country-to-country aid flows.
◆ Preliminary OECD data show total official development assistance fell 23.1% in real terms in 2025, the largest annual contraction on record.
◆ Cancelling 18% of annual external debt-service payments could theoretically replace aid losses for roughly half of the 105 countries included in the paper’s detailed financing analysis.
◆ Even cancelling 100% of annual debt service would not replace the estimated losses in 18 countries.
◆ The Central African Republic, Malawi and Zambia are identified as cases where neither full debt-service relief nor the study’s maximum additional remittance capacity would be sufficient on its own.
Continue reading ↓
Foreign aid for health fell by an estimated $4.2 billion, or 13%, in 2025 as major donor countries sharply reduced international assistance, according to a new research paper.
Emergency-response funding fell by another $5.5 billion, while the researchers estimate that 130 recipient countries lost aid overall.
The sector figures are estimates rather than final observed 2025 country-level aid totals.
Researchers Andrea Vismara, Rafael Prieto-Curiel and Rosie Hayward calculated them by applying donor-specific 2025 funding cuts to the bilateral aid relationships recorded in 2024.
◎ HEADLINE NUMBER
$4.2 billion
Estimated reduction in foreign aid directed to health in 2025 — equivalent to a 13% decline.
Health was one of the sectors hit hardest
The model finds large reductions across several types of development assistance.
−$4.2B
Health
−13%
−$5.5B
Emergency response
−21%
−$10.2B
Government & civil society
−29%
The only broad aid category in the study to record a large increase was general budget support, which rose by an estimated $12.7 billion, or 28.7%, largely because of increased financing directed toward Ukraine.
That redistribution means the overall contraction did not fall evenly across countries or across the purposes aid was intended to finance.
OECD data point to further pressure on health funding
The paper’s estimate sits within a broader decline in international assistance.
Preliminary OECD figures show official development assistance fell 23.1% in real terms in 2025, the largest annual contraction on record.
The five largest providers accounted for 95.7% of the overall decline, with US assistance falling 56.9% from the previous year.
Separate OECD projections for health aid suggest the pressure could continue.
The OECD estimates bilateral health aid could be 29% to 46% lower in 2026 than in 2024, a reduction of roughly $5 billion to $8 billion.
Important distinction: The OECD’s 2026 figures are separate projections. They are not part of the paper’s estimated $4.2 billion health-aid reduction for 2025.
Aid losses reached 130 countries
Across all sectors, the researchers estimate that 130 aid-recipient countries experienced some reduction in financing.
The combined estimated loss in country-to-country aid flows was $25.9 billion.
About 40% of affected countries were in Sub-Saharan Africa, followed by 17% in Latin America and the Caribbean and 15% in East and Central Asia.
India and Syria each lost close to $1 billion in the model, while Jordan, Ethiopia, Turkey and the Democratic Republic of Congo lost an estimated $750 million to $936 million each.
Syria’s estimated loss equalled more than 5% of national income
The impact was much larger when measured relative to the size of some recipient economies.
5.25%
Syria
estimated GNI loss
4.6%
South Sudan
estimated GNI loss
4.2%
Somalia
estimated GNI loss
These countries are also among those where replacing lost aid through conventional financing mechanisms is particularly difficult.
That connects with earlier StatsJournalist coverage of the unusually limited fiscal buffers in fragile and conflict-affected economies.
↳ CAN OTHER FINANCE REPLACE THE LOSS?
Debt relief could help many countries — but not all
The researchers tested whether reduced external debt-service payments could create enough fiscal space to equal each country’s estimated aid loss.
A universal reduction equal to 18% of annual debt-service payments would be enough to fully compensate roughly half of the countries in the detailed analysis.
↔ DEBT-RELIEF SCENARIO
18%
Universal debt-service relief at this level could theoretically replace the aid loss for about half of analysed countries.
But 18 countries had estimated aid losses larger than their entire annual external debt-service payments.
Even cancelling 100% of that debt service would therefore be insufficient.
Syria and Somalia would require relief worth about 30 times their annual debt-service payments to match their estimated losses, while Afghanistan would require roughly eight times its annual payments.
The problem in these cases is that there is relatively little conventional debt service available to cancel.
Higher remittances face a different constraint
Remittances are much larger than foreign aid in aggregate, but they follow migration networks rather than the geography of aid need.
Countries affected by aid cuts received about $632.2 billion in remittances in 2024.
The researchers calculate an additional theoretical capacity by assuming every working-age migrant who is not already remitting begins sending money.
Mobilising 10% of that estimated unused capacity would compensate roughly half of the analysed countries.
Yet even full mobilisation would not replace the estimated loss in 16 countries, all of them in Africa.
Health-dependent countries can face limited remittance capacity
The researchers identify a group of 25 “aid-remittance constrained” countries, all in Sub-Saharan Africa except Jordan.
For nine countries in this group, most foreign aid was directed toward health and education programs.
The group’s median estimated aid loss was equivalent to 1% of GNI, while replacing that loss through additional remittances would require 226.7% of the modeled unused remittance capacity.
In other words, the typical country in this group would require more than twice the study’s theoretical available increase in remittances to replace its lost aid through that channel alone.
Debt relief may be more feasible for some of these countries because they remain more integrated into international lending markets.
Three countries fell beyond both modeled financial backstops
The Central African Republic, Malawi and Zambia stand out in the analysis.
For all three, neither cancelling all annual external debt service nor fully mobilising the paper’s estimated additional remittance capacity would individually replace the aid loss.
That does not mean no financing option exists.
It means the two mechanisms specifically tested by the researchers are insufficient even under their maximum scenarios.
✦ Why it matters ✦
A reduction in health aid is more than a change in the overall volume of development finance because much of that funding is directed toward specific programs and services.
The paper itself does not estimate how the $4.2 billion reduction will affect deaths, disease rates, medicine availability or individual health systems.
It does, however, show that replacing the missing financing is not simply a matter of finding an equally large pool of money somewhere else in the global economy.
Debt-service relief can potentially free government resources, but governments ultimately determine how those savings are spent.
Remittances work differently because they are private transfers to households and cannot automatically replace externally funded health programs, public clinics or government services.
The result is an especially difficult position for countries that lose substantial aid while having limited fiscal capacity and weak access to alternative external finance.
With separate OECD projections pointing toward further reductions in health assistance in 2026, those financing constraints could become increasingly important.
ⓘ How to read the findings
The paper was posted to arXiv in August 2026 and should be treated as preliminary research. A peer-reviewed journal publication of this study has not been verified.
The estimated $4.2 billion health-aid loss, $5.5 billion emergency-response loss and $10.2 billion government-and-civil-society loss are produced by the authors’ model. They are not final observed 2025 recipient-level sector totals.
The model starts with bilateral aid relationships reported for 2024 and applies each donor’s 2025 non-Ukraine reduction proportionally across its other recipients because detailed final 2025 donor-recipient data are not yet available.
The observed OECD result is the 23.1% real decline in overall official development assistance. The paper’s country and sector estimates are downstream calculations based partly on those donor-level changes.
The detailed debt, remittance and vulnerability analysis covers 105 countries with populations above 1 million and sufficient financial data, not all 130 countries with estimated aid losses.
The 18% debt-service relief scenario assumes the same proportional reduction across annual external debt-service payments. It does not predict that creditors will agree to provide such relief.
Debt-service relief also does not guarantee that an equivalent amount would be spent on the same health, education or humanitarian programs previously financed through aid.
The 10% remittance scenario refers to 10% of modeled unused capacity, not a 10% increase in current remittance flows.
The paper estimates that capacity using a theoretical ceiling in which every working-age international migrant sends remittances and assumes participating migrants transfer 18% of monthly per-capita income in their destination country.
The research does not estimate deaths, illnesses, hospital closures, treatment interruptions or other health outcomes caused by the 2025 health-aid reduction. Claims about those outcomes would require separate evidence.
The four vulnerability groups are produced using k-means clustering of modeled aid losses, debt-service capacity and remittance capacity. They are analytical categories rather than official international classifications.
The analysis operates at the country level and cannot show where within a country individual aid projects or remittance recipients are located, an important limitation acknowledged by the authors.
