Eswatini, Uganda, Rwanda, Ghana: The New Third-Country Architecture
- Citation
- Various: Uganda Asylum Cooperative Agreement (signed July 29, 2025); Eswatini agreement ($5.1M USD; up to 160 deportees); Rwanda agreement (Aug. 2025; ~$7.5M USD; up to 250 deportees); Ghana agreement (14 deportees as of late 2025)
- Court
- U.S. district courts; foreign tribunals including Uganda Law Society proceeding
- Statute
- INA § 241(b)(2)(E); foreign domestic law
- Holding
- Multiple challenges pending in U.S. and foreign forums; no consolidated U.S. ruling on the legality of the bilateral architecture as such; Uganda Law Society 2026 filing seeks “bespoke reliefs” against the cooperative agreement.
The bilateral architecture the Trump administration assembled in 2025 to support third-country removals is broader than the South Sudan and El Salvador episodes that generated the most U.S. attention. By the spring of 2026, agreements were in place or under implementation with at least four African states — Eswatini, Rwanda, Uganda, and Ghana — in addition to the previously-existing arrangements with various Central American countries.
The agreements
Each agreement has its own structure, but the general pattern is recognizable. The receiving country agrees to accept a specified number of deportees from the United States. The United States provides funding, in amounts ranging from approximately $5.1 million (Eswatini) to approximately $7.5 million (Rwanda). Implementation occurs through executive-to-executive arrangements rather than treaty processes, and so without parliamentary review on either side. Deportee categories typically exclude those with criminal records and unaccompanied minors; preferences are often expressed for nationals of regional or African countries.
Operational implementation has varied. Nineteen people had been transferred to Eswatini by April 2026 across three flights (July 15, 2025; October 6, 2025; March 11, 2026). Deportees from Ghana’s arrangement totaled fourteen by late 2025. Uganda received its first transfers in April 2026.
The U.S. litigation
On the U.S. side, the principal challenges have proceeded through the D.V.D. class action and individual habeas petitions. The procedural-protections analysis discussed elsewhere on this site applies equally regardless of receiving country: notice, language access, attorney notification, CAT screening. The substantive APA challenges to the underlying directive have been litigated principally in the District of Massachusetts.
What U.S. courts have not yet squarely addressed is the legality of the bilateral architecture as such — whether the executive’s ability to enter into and implement these agreements without statutory authorization or treaty process is itself constitutionally suspect. The closest the federal courts have come is the procedural-protections framework Judge Murphy developed in D.V.D., which operates as a constraint on implementation rather than a challenge to the underlying foreign-relations conduct.
“The Government’s arrangements with foreign sovereigns cannot displace the constitutional protections owed to the persons it seeks to remove.”
The foreign-court challenges
The most novel litigation has come not from U.S. courts but from foreign ones. In April 2026, the Uganda Law Society filed in Uganda’s High Court seeking what its pleadings termed “bespoke reliefs designed to arrest this patent international illegality.” The Costa Rican Constitutional Court had earlier — in June 2025 — ordered the release of Asian migrants the United States had transferred to Costa Rica via a separate bilateral arrangement, on the ground that detention without information about immigration status or counsel access violated Costa Rican constitutional protections. These foreign-tribunal habeas-equivalent rulings are now an unfamiliar but increasingly important feature of the third-country removal landscape.
Significance
The third-country architecture matters for habeas practice in two distinct ways. First, it expands the universe of removal destinations the government can credibly invoke when arguing under Zadvydas that removal is significantly likely in the reasonably foreseeable future. The countervailing argument is that arrangements without operational track records do not satisfy Zadvydas’s particularization requirement, but that argument requires litigation in each individual case.
Second, the architecture creates a category of removed persons whose subsequent treatment is largely outside U.S. judicial reach. The constructive-custody theory developed by Chief Judge Boasberg in the AEA context may prove transferable, but it has not yet been tested as to the African-state arrangements. The next twelve to twenty-four months of habeas litigation will likely test the limits of that theory.
Filed under Third-Country Removal. Published Through April 2026.