USAID suspends food aid to all of Ethiopia

The food aid was being stolen. The food aid stopped.

A USAID inspection in March 2023 finds American wheat at a commercial flour mill. The trail leads to an industrial-scale diversion operation involving federal officials, Tigrayan officials, ENDF personnel, EDF personnel, and private grain traders. The aid is suspended for five months. Hundreds, possibly thousands, of vulnerable Ethiopians die during the suspension. The trust between Ethiopia and its donors does not recover.

Filed November 2023, after the November resumption of aid

In late February 2023, a USAID inspector working in Tigray Region observed an unexpected sight: a truckload of wheat sacks marked with US Agency for International Development branding, parked at the loading dock of a commercial flour mill in Mekelle. The wheat had been donated by the United States and channelled through the World Food Programme for distribution to vulnerable populations in Tigray; its presence at a commercial mill meant it was being diverted into the commercial market rather than distributed to the food-insecure households for whom it had been donated. The inspector’s report initiated what became, over the following four months, the most consequential food-aid diversion investigation in modern Ethiopian-international development history.

The trail led — according to subsequent reporting by Reuters, Associated Press, and the Washington Post in May and June 2023 — to what US officials would characterise as “industrial-level” theft of food aid. The investigation, conducted jointly by USAID’s Office of Inspector General, the WFP’s internal monitoring functions, and the State Department’s Bureau of African Affairs, identified 63 diversion sites across seven of Ethiopia’s nine federal regions. The diversion involved federal Ethiopian government officials (in particular the federal Disaster Risk Management Commission, which had administered beneficiary lists), Tigrayan regional officials (in some of the post-Pretoria interim administration structures), Ethiopian National Defence Force personnel (ENDF formations had been moving aid in some operational zones), Eritrean Defence Force personnel (in northern Tigray zones where EDF still controlled the territory), private grain traders (who had been purchasing the diverted aid for the commercial market), and intermediate brokers who had been facilitating the transactions. A Tigrayan regional official would later acknowledge to investigators that more than 7,000 tonnes of US-donated wheat had been stolen in Tigray alone over a roughly 12-month period — enough to feed approximately 450,000 people for a month. The total quantity diverted across all seven regions was substantially larger, although final figures have not been published. [Reuters via Martin Plaut, “UN food agency failed to act as U.S. aid was looted in Ethiopia,” 18 October 2024, https://martinplaut.com/2024/10/18/un-food-agency-failed-to-act-as-u-s-aid-was-looted-in-ethiopia/; Washington Post, “USAID cuts food aid supporting millions of Ethiopians amid charges of massive government theft,” 8 June 2023, https://www.washingtonpost.com/world/2023/06/08/ethiopia-usaid-scandal-cutoff-food/]

The mechanism of the diversion was, by the investigators’ reconstruction, simple in concept and complex in execution. The federal Ethiopian government had insisted, from the beginning of the wartime aid operation, on retaining administrative control of beneficiary lists — the rosters of households eligible for monthly food distribution. The federal government had, by Pretoria Agreement provisions and by previous bilateral arrangements with USAID and WFP, presented these lists as authoritative. The lists were, in operational reality, substantially inflated: households that did not exist, households that did exist but did not qualify, households that did qualify but received only fractions of the rations the lists allocated to them, and substantial categories of allocated food that simply did not reach any household at all. The surplus produced by the inflation — the difference between the rations the lists allocated and the rations that were actually distributed to qualifying households — was diverted into the commercial market through the brokers and into the military supply chains through the ENDF and EDF formations operating in the regions.

The WFP staff and other aid organisations had been aware of the inflated-list problem since 2021. WFP country office personnel had communicated the concern to senior management on multiple occasions through 2021 and 2022. The senior management response had been, in operational terms, to maintain a public posture of zero tolerance for diversion while declining to push back operationally against the federal government’s administrative control of beneficiary lists. The political logic was straightforward: WFP and USAID were operating in Tigray during an active armed conflict in which the federal government controlled physical access to the region; aggressive pushback against the federal government’s beneficiary-list practices would, in the worst case, produce the closure of access entirely; the choice, as senior WFP officials framed it internally, was between an imperfect operation with significant diversion and no operation at all. The choice was made, through 2021 and 2022, to operate the imperfect operation. The choice meant, in operational consequence, that food intended for famine-affected Tigrayans was being stolen, and the international community’s most senior managers were aware of the theft and were not stopping it.

The country director Claude Jibidar privately acknowledged the diversion problem in internal WFP communications through 2022. Subsequent reporting based on leaked WFP internal communications has documented the specific terms in which Jibidar and his senior staff discussed the trade-offs. The internal acknowledgment did not produce operational change. The public posture remained zero tolerance; the operational reality remained substantial tolerance; the result was that the diversion continued at scale through the period covered by the subsequent USAID investigation.

The USAID announcement of the food-aid suspension came on 8 June 2023. The suspension was, in its initial scope, limited to Tigray Region. The WFP followed with a parallel suspension on 12 June. As the investigation expanded through June and into July, the suspension expanded with it: by late July 2023, the suspension covered USAID food aid to all of Ethiopia, including the drought-affected Somali Region, the conflict-displaced populations in Amhara Region, the Sudanese refugees in Gambella Region (where approximately 300,000 refugees from Sudan’s RSF–SAF war had begun arriving from May 2023 onwards), the Eritrean refugees in remaining camps, and the broader food-insecure populations across the country. The total population dependent on USAID-WFP food aid in Ethiopia at the time of the suspension was approximately 20 million people. [The New Humanitarian, “USAID/WFP food aid freeze leads to suffering and deaths in Tigray,” 23 June 2023, https://www.thenewhumanitarian.org/news/2023/06/22/usaidwfp-food-aid-freeze-leads-suffering-and-deaths-tigray]

The mortality consequences of the suspension began to surface within weeks. The Tigrayan regional interim Disaster Risk Management Commission, working with limited resources and capacity, documented 595 hunger deaths in just three of Tigray’s seven zones during the period from June through August 2023 — and noted that the figure was almost certainly an undercount because the Commission’s monitoring infrastructure did not reach the rural areas where most of the deaths were occurring. Associated Press reporting through July, August, and September 2023 documented “hundreds, possibly thousands” of hunger deaths across the country during the five-month aid suspension. The Tigrayan diaspora press, working through contacts with families in Tigray, produced higher estimates; the federal government disputed all figures.

The mortality figures during the five-month suspension are, in mid-2026, contested in their specifics but uncontested in their basic shape. People died of hunger in Tigray during the suspension. People died of hunger in Somali Region during the suspension (where the 2022–2023 drought had produced famine conditions that USAID food aid had been mitigating). People died of hunger in Amhara Region during the suspension (where the August 2023 Fano insurgency was producing additional displacement). People died of hunger in Gambella refugee camps during the suspension. The number who died is uncertain. That deaths occurred — that the suspension produced incremental mortality beyond what would have occurred without it — is not in serious dispute. [PBS NewsHour, “U.S. will resume food aid to millions in Ethiopia after monthslong pause,” 15 November 2023, https://www.pbs.org/newshour/world/u-s-will-resume-food-aid-to-millions-in-ethiopia-after-monthslong-pause-over-massive-corruption]

The resumption of aid, when it came on 15 November 2023, was structured around new anti-diversion measures that USAID and WFP had negotiated with the federal government over the preceding months. The new measures included: biometric registration of beneficiary households (replacing the federal-government-administered lists with internationally-administered registries); third-party distribution (in which non-Ethiopian-government contractors handled the physical distribution of food at the household level); real-time third-party monitoring (in which independent monitors observed distribution and reported in operational time to USAID and WFP); removal of Ethiopian government officials from the supply chain at the warehouse-and-distribution level; and reduction of the role of Ethiopian government beneficiary-list compilation in favour of internationally-administered systems. The federal government’s acceptance of these measures, after months of negotiation, represented a significant retreat from its previous administrative posture; the acceptance was, by US officials’ subsequent characterisations, not voluntary but the price of resumed assistance.

The new measures held, in modified form, through 2024, 2025, and into 2026. Subsequent USAID and WFP audits have not identified diversion at the scale of the 2021–2023 period. The accountability mechanism, in operational terms, has succeeded in its narrow purpose. The mechanism has not, in the broader political terms, succeeded in restoring the trust between Ethiopia and its donors that the diversion scandal had broken.

What the diversion scandal demonstrated, beyond the specific facts of the diversion itself, was three things about the operational relationship between the Ethiopian federal government and the international development community that has structured post-war Ethiopian politics.

First, the federal government’s administrative capacity for managing wartime humanitarian operations was, by 2023, characterised by structural corruption that the international community had been aware of and had nonetheless continued to operate through. The diversion was not, in any meaningful sense, a surprise to the international community in June 2023. The WFP internal communications had documented the problem since 2021; USAID country-office reporting had documented it through 2022. The decision to continue operating despite the documented diversion had been a senior-management decision in both Washington and Rome (WFP headquarters). The decision was rationalised on the grounds that the alternative — losing access to Ethiopia entirely — would produce worse humanitarian outcomes than continued operation under conditions of corruption. The rationale was probably correct in operational terms, in the limited sense that ending access in 2021 or 2022 would have produced more deaths than the diversion was producing. The rationale was also a structural admission that the international community’s relationship with the Abiy government had become a relationship in which the international community knowingly funded substantial corruption because the alternative was worse.

Second, the federal government’s calculation of the political costs of the diversion was substantially mistaken. The federal government had, through 2021 and 2022, calculated that the international community’s need for Ethiopian government cooperation in wartime humanitarian operations gave the federal government leverage over how those operations were conducted; that the international community would tolerate diversion at industrial scale because the alternative was worse; that the federal officials, ENDF personnel, and aligned actors who were profiting from the diversion could continue to profit without consequence. The calculation was correct through 2021 and 2022 and substantially incorrect by June 2023. The USAID and WFP decision to suspend aid in June 2023, despite the obvious humanitarian costs of suspension, demonstrated that the international community’s tolerance had a limit and that the limit had been exceeded. The cost of exceeding the limit was paid, in immediate terms, by the vulnerable Ethiopian populations whose food aid was interrupted; it was also paid, in longer-term terms, by the federal government’s diplomatic and developmental relationships with the donor states whose trust had been violated.

Third, the diversion scandal marked the operational point at which the Western donor community’s broader political posture toward the Abiy government began to harden. The European Union had suspended €88 million in budgetary support to Ethiopia in December 2020 over wartime conduct concerns. The United States had imposed restrictions on military assistance through 2021 and 2022. The diversion scandal of 2023 was the moment at which the broader development-assistance relationship — much larger than the military and budgetary-support components, encompassing health, education, agriculture, infrastructure, and macroeconomic support — entered a period of structural review. By 2024, the World Bank and IMF programmes that would eventually produce the July 2024 birr float had been negotiated under operational conditions that incorporated significant skepticism about Ethiopian government administrative capacity and integrity. The Western donor community’s posture by 2024, 2025, and into 2026 has been one of conditional engagement — willing to support Ethiopian development on specific terms with significant verification requirements, but no longer willing to operate on the basis of federal-government administrative declarations alone.

The diversion scandal had a fourth, less remarked, consequence that has shaped the broader Ethiopian humanitarian landscape through 2024 and into 2026. The new anti-diversion mechanisms — biometric registration, third-party distribution, third-party monitoring — produced an operational architecture in which international assistance to vulnerable Ethiopian populations could continue substantially independent of federal-government administrative control. The architecture was expensive (the additional verification costs added approximately 15–20% to the unit cost of distributed aid) and was less politically responsive (the federal government’s input into beneficiary selection was significantly reduced), but it was operationally functional in conditions of low trust. The architecture has, in mid-2026, become the operational template for international assistance not only in Ethiopia but in several other African contexts where donor-government trust has broken down. The Ethiopian diversion scandal of 2023 has, in this sense, contributed to the broader transformation of international humanitarian operations in conditions of weak governance.

The mortality consequences of the five-month suspension remain, in mid-2026, the central ethical question of the diversion scandal. The international community’s choice in June 2023 — to suspend aid in response to documented diversion, knowing that the suspension would produce additional civilian deaths in the short term — was a choice with ethical content that has not been adequately addressed in subsequent international discussion. The alternative choice — to continue aid despite documented diversion, knowing that the diversion was funding actors implicated in atrocity conduct — was also a choice with ethical content. The framing of the choice in the international press of June 2023 was substantially in terms of accountability for the diversion; the framing produced political pressure for suspension that was not adequately balanced against the predictable mortality consequences of the suspension itself. The decision to suspend was made; the deaths followed; the resumption produced an operational system less subject to diversion but more expensive and more politically constrained.

The fundamental ethical question — whether the international community was, on balance, right to suspend aid in June 2023 — does not admit a confident answer. The deaths from the suspension were real and were not adequately predicted by the decision-makers. The continued diversion would have been real and was being funded with money the donor publics had not approved for such uses. The two ethical costs do not commensurate easily. The decision was made; the consequences accrued; the operational architecture that emerged is, in mid-2026, the architecture that exists.

The diversion scandal’s deeper political legacy is the moment of broken trust it represents between the Abiy government and the international development community whose support has been, since the early days of his premiership, one of the operational pillars of his political project. The trust was broken in 2023 and has not, in mid-2026, been fully restored. The federal government’s subsequent fiscal program — the December 2024 Eurobond default, the IMF-conditional reform agenda, the privatisations under negotiation through 2025 and 2026 — has been conducted in a development environment of substantially reduced political latitude. The international community’s willingness to extend further trust to the Abiy government has been calibrated against the diversion scandal’s documented record. The political consequences of the scandal have, in this sense, outlasted its immediate humanitarian consequences and become structural features of the development-assistance landscape in which the Ethiopian government now operates.

The 7,000 tonnes of wheat that disappeared from the supply chain in Tigray in 2022 produced, in the cumulative political effect of their disappearance, the structural collapse of the political assumption that had organised the Abiy government’s relationship with its donors through the first four years of his premiership: that the donors would, in the end, tolerate what the federal government chose to do because the alternative was operationally worse. The assumption was, by mid-2023, no longer operative. The federal government has, since then, been operating in a development environment in which the donor states’ tolerance is conditional and the conditions are enforced.

The hungry Tigrayans, Somalis, Amharas, and refugees who died during the five-month suspension paid the most concentrated portion of the cost of this transformation. The federal government’s senior officials, ENDF officers, and aligned actors who had been profiting from the diversion paid a smaller, more distributed cost in subsequent political consequence. The international community’s senior managers who had been operating the imperfect humanitarian programme since 2021 paid no formal personal cost but were, in the operational record, the actors whose tolerance had enabled the diversion’s scale.

No one was prosecuted. No senior federal official has been charged. No WFP or USAID senior official has been disciplined. No Tigrayan regional official whose role in the diversion was documented has been held to account. The accountability gap that the diversion scandal opened has not been closed. The hunger deaths during the suspension have been counted in approximate ranges. The wheat that disappeared has not been recovered. The donor trust has not been restored. The operational architecture for distributing aid has changed; the political architecture for trusting the federal government to handle aid responsibly has, in 2026, not been rebuilt.

The diversion scandal of 2023 was, in the most precise sense, the moment at which the international community discovered what its operational tolerance had been funding. The discovery did not undo what had been done. The discovery did, however, change the terms on which the international community engages with the Abiy government going forward. The change is the scandal’s most durable legacy.

---

Sources for Article 15

1. Reuters via Martin Plaut, “UN food agency failed to act as U.S. aid was looted in Ethiopia,” 18 October 2024 — https://martinplaut.com/2024/10/18/un-food-agency-failed-to-act-as-u-s-aid-was-looted-in-ethiopia/
2. Washington Post, “USAID cuts food aid supporting millions of Ethiopians amid charges of massive government theft,” 8 June 2023 — https://www.washingtonpost.com/world/2023/06/08/ethiopia-usaid-scandal-cutoff-food/
3. The New Humanitarian, “USAID/WFP food aid freeze leads to suffering and deaths in Tigray,” 23 June 2023 — https://www.thenewhumanitarian.org/news/2023/06/22/usaidwfp-food-aid-freeze-leads-suffering-and-deaths-tigray
4. PBS NewsHour, “U.S. will resume food aid to millions in Ethiopia after monthslong pause,” 15 November 2023 — https://www.pbs.org/newshour/world/u-s-will-resume-food-aid-to-millions-in-ethiopia-after-monthslong-pause-over-massive-corruption
5. Associated Press reports on Ethiopia food aid suspension, June–October 2023
6. World Food Programme Ethiopia situation reports, 2023
7. USAID Office of Inspector General reports on Ethiopia food assistance, 2023–2024
8. UN OCHA Ethiopia Humanitarian Updates, 2023
9. Tigray Interim Disaster Risk Management Commission reports on hunger mortality, 2023
10. International Crisis Group, “Time to Get Real with the Tigray Peace Deal,” 19 January 2023 — https://www.crisisgroup.org/africa/horn-africa/ethiopia/time-get-real-tigray-peace-deal
11. Heinrich-Böll-Stiftung, “The Pretoria Agreement for Tigray: One Year After,” 1 November 2023 — https://hoa.boell.org/en/2023/11/01/pretoria-agreement-tigray-one-year-after

---

End of expanded article for Dispatch 15. One article, approximately 2,800 words.