A new study finds that the sudden termination of USAID funding in early 2025 corresponded with an increase in violent events in African locations that had previously received U.S. assistance. Published in Science, the paper — led by University of Chicago data scientist Austin Wright and colleagues — uses detailed aid-disbursement maps and event-level conflict data to estimate how withdrawing aid affected instability on the ground.
Aid’s complex relationship with violence
The authors begin by noting a central tension in research on foreign assistance: aid can both reduce and intensify conflict. By creating jobs, providing services and supporting livelihoods, aid can lower incentives to join armed groups or take part in unrest. But aid-funded assets — clinics, food distributions, infrastructure and the services they enable — can also become valuable items to control, and thus a source of contestation.
Method and data
To test the consequences of an abrupt aid cutoff, the researchers mapped USAID disbursements at the state or provincial level before the agency’s shutdown. They then matched those maps to a high-frequency database of violent events — the Armed Conflict Location and Event Data (ACLED) project — tracking armed clashes, protests, riots and violence against civilians in the ten months before and after the early-2025 disruption. The timing of incidents in ACLED allowed the team to examine short-term changes in violence that coincided with the disappearance of funding.
Findings and mechanisms
The study reports that locations that had received more USAID assistance experienced larger increases in violent activity after the aid evaporated than places that had received little or no assistance. Wright and his coauthors argue the effect operated through several channels: the sudden loss of wages and programs (health clinics, food distributions, public-sector contracts) made many people economically vulnerable, increasing the pool of recruits for armed groups and raising incentives to engage in protest or predation. At the same time, the assets and positions worth controlling did not disappear overnight, creating conditions for clashes and looting.
The observed uptick in violence included armed-group combat, demonstrations that turned violent, and deliberate attacks on civilians. Wright summarizes the dynamic as a collapse of livelihoods combined with the persistence of things to fight over — a combination that, at least short-term, produced “chaos and violence.”
Illustrative example: Kakuma
The paper points to protests at Kakuma refugee camp in northwest Kenya in July 2025 as an example of the pattern. About 300,000 refugees who relied on food assistance and other services cut by the aid pause demonstrated in response; distributions were reduced, residents threw rocks and set fires, and one person was killed. Wright says such incidents are the sort the analysis captures.
Official response
NPR sought comment from the U.S. State Department. Spokesperson Tommy Pigott argued the study misreads developments in Africa, saying the administration had made “unprecedented progress towards the advancement of peace on the continent” and that recent policy changes focused U.S. assistance on “efficiency, effectiveness, and partnership.”
Where the effect was smaller
The researchers also found heterogeneity in outcomes. Areas with stronger institutional constraints on executives — where legislatures, courts or other checks limit leaders’ unilateral power — were less likely to see a spike in violence after the funding cut. Countries that moved to plug gaps with domestic spending provided counterexamples; the study cites Nigeria’s $200 million supplementary health allocation and South Africa’s additional HIV funding as instances where governments stepped in to blunt the shock.
Peer assessment
Independent researchers say the topic is difficult to study but that the results are persuasive. Andy Solow, a statistician at the Woods Hole Oceanographic Institution who was not involved in the work, noted the complexity of conflict dynamics and raised technical questions such as spatial contagion — the idea that violence can spread from place to place and thus create interdependence among events. Still, he told NPR he found the paper’s conclusions convincing and unlikely to be overturned by those concerns.
Implications
The authors warn that the short-term surge in violence matters for the medium and long term, because recent conflict is a strong predictor of future conflict: once violence escalates it tends to be self-reinforcing. That implies that even if aid were restored, the damage done during the period of elevated violence would not necessarily reverse quickly. The study frames the abrupt dismantling of a major humanitarian program as having “enormous consequences on the ground,” undermining livelihoods and fueling instability in ways that can persist beyond the initial shock.