What return hubs are – and what they are not

Return hubs are the latest and most concrete expression of member states’ search for “innovative solutions” to shift elements of asylum and return management outside EU territory. Their appeal rests not only on the prospect of additional returns, but also on a signalling effect: reassurance for domes­tic publics and deterrence of prospective migrants. How realistic and valuable these effects are lies at the heart of the controversy.

Return hubs are intended for individuals who are already subject to an enforceable return decision. That sets them apart from inadmissibility procedures, including the safe-third-country concept, and offshore asylum processing. Under safe-third-country rules an asylum application may be rejected as inadmissible if the applicant is considered to have access to protection elsewhere, while offshore processing arrangements concern the location where initial asylum claims are assessed. Return hubs, by contrast, relate to the post-decision stage, when an individual has already been ordered to leave. The question is how and where the return decision is to be implemented.

In the EU context, return normally implies movement back to a place to which the person concerned is linked through nationality, residence or transit, or to a destination agreed with the individual. The logic of return hubs is different: transfer to an otherwise unrelated state, often in exchange for payments to the receiving state. Such transfers would generally be involuntary.

Three operational models are conceivable, each carrying distinct legal and political risks. First, a return hub may function as a temporary holding centre pending onward return to the country of origin. This relocates custody and case management while documentation and readmission are processed. As such, it merely changes the place where the problem is managed, rather than solving it.

Second, a return hub may become a longer-term destination if the host country offers secure and durable residence. This raises the question whether European states are effectively paying third countries to assume long-term or even permanent respon­sibility for people with whom they have no prior connection.

Third, a hub may be framed as a specific instrument for individuals who have been convicted of serious offences or are considered to pose security risks, and who cannot be returned directly. In this version, the hub is a specialised public security measure rather than a general response to low return rates. It would require strict criteria for in­clusion, individual judicial review of detention or security grounds, and clear rules on sentence enforcement, costs and post-release status. Mere non-cooperation with immigration authorities should not qualify.

European precedents, but no template

No previous European arrangement with a third country fully corresponds to the opera­tional models set out above. The initiatives often cited as precedents in the wider externalisation debate addressed different legal and operational problems. They do not provide a tested template for return hubs, but they illustrate the risks, costs and dependencies that can arise when elements of European asylum and migration policy are shifted beyond EU territory.

The UK-Rwanda scheme that London introduced in April 2022 aimed to outsource entire asylum procedures. The scheme was expensive and generated intense litigation before being politically abandoned; ultimately no one was forcibly transferred. UK government data put the combined cost of the Rwanda partnership and associated implementation of the Illegal Migration Act at £715 million. The UK-Rwanda example shows how offshore arrangements can create significant liabilities and sunk costs without ever becoming operational.

The 2023 Italy-Albania protocol is a bespoke bilateral arrangement under Italian jurisdiction that combines detention with offshore asylum processing. It was designed for accelerated asylum processing of a specific group of irregular migrants and subsequent direct return to their countries of origin. Legal challenges have so far pre­vented full implementation, and the facilities in Albania have been used only for a small number of cases. The agreement demonstrates that multi-purpose facilities for asylum and return can be established outside EU territory under a member state’s authority. But it does not show that such facilities produce faster, cheaper or more durable returns, nor that they can be easily replicated. Albania has underlined that it does not intend to extend the arrangement to other EU member states.

Finally, the 2016 EU-Turkey Statement has been central to the EU’s externalisation strategy over the past decade. Its main impact has been to incentivise Turkey finan­cially to provide protection and basic services to Syrian refugees, while reinforcing its policing of its European borders. It has also been treated as a template for arrangements under which member states may declare certain asylum applications inadmissible. But specifically its return component remained limited in practice and has been suspended since 2020. The deal therefore underlines the difficulties involved in sending large numbers of returnees to a third country, even in a con­text of regional proximity, shared interests and substantial financial support.

The United States: a small, costly and selective precedent

Recent US practice is not a legal template for the EU, but does offer a useful empirical point of reference. Since January 2025, Washington has expanded third-country deportations, even as the legality of these transfers is being challenged in US courts. The comparison is relevant because conditions in the United States should make such transfers easier than in the EU: immigration enforcement and foreign policy are controlled by a single federal executive; US immigration law provides mechanisms for removal to third countries where return to the country of origin is not possible; and the government can combine transport and detention capacity with diplomatic, economic and security leverage. If remote third-country removals were readily scalable, these conditions should favour them. Their modest results therefore make the US experience instructive for the EU’s planned return hub arrangements.

While the Migration Policy Institute esti­mates that about 15,000 third-country removals took place in 2025, it is important to note that around 13,000 of those affected were sent to Mexico. Those transfers are more similar to cooperation with a neighbouring or transit state – as under the EU-Turkey Statement – than to a remote return hub. Excluding Mexico, there were roughly 2,000 transfers encompassing a heterogeneous and incompletely documented set of arrangements, including transfers to Panama, Costa Rica and Ghana. Full details are difficult to establish on account of transparency issues.

What is clear, however, is that the US has only a small group of arrangements that resemble the EU return hub concept, where distant states receive individuals with final orders of removal and then hold them, pursue their onward return or permit them to stay. A February 2026 report by Democratic minority staff of the US Senate Foreign Relations Committee found that five governments receiving direct payments had accepted about 300 transfers by January 2026. Around 250 of these occurred within the exceptional El Salvador operation, which was primarily incarceration-driven and was challenged over due process concerns. Other agreements limit the numbers involved and preserve broad discretion for the respective host state. Eswatini reportedly agreed to accept up to 160 deportees and had received 15 by November 2025; Rwanda agreed to consider up to 250 but initially admitted only a small number (seven as of January 2026).

Documented US cases raise question marks over efficiency and effectiveness. A Jamaican national transferred to Eswatini and a Mexican national transferred to South Sudan were both flown onward to their country of origin soon afterwards, at reported transport costs of approximately US$181,000 and US$91,000 respectively. In both cases, it was reported that the respective government had not been informed of the planned third-country repatriation of its citizen and had subsequently indicated that it had not refused to provide documentation. As such examples demonstrate, departure from the sending state is an inadequate measure of success. The relevant indicator is not departures, but the number of verified onward returns to countries of origin or lawful admissions to third countries that would not otherwise have occurred.

In addition, the US cases underline the risks that arise once an individual has been transferred into the custody of the receiving state. The key question is not only whether the transfer can be carried out, but whether detention conditions, access to remedies, protection against onward refoulement and prospects for release or onward return can be verified afterwards. Recent US arrangements appear to provide only weak monitoring and oversight mechanisms, which is especially problematic where receiving states – like Equatorial Guinea and El Sal­vador – have weak institutional transparency, entrenched corruption or documented human-rights concerns. The lesson for the EU is that careful partner selection and enforceable oversight are not secondary safeguards, but central conditions for lawful and sustainable third-country transfers.

Costs reinforce the same conclusion. Remote arrangements require diplomatic concessions, accommodation, detention, escorts, flights, monitoring and follow-up case management. The Senate minority report identified more than US$32 million in direct payments to five governments and at least US$7.2 million in flight costs; in some low-volume arrangements, expenditure exceeded US$1 million per person. Such expenditure may be politically defensible for a narrowly defined group of serious offenders or security threats. But the figures do not support the claim that remote third-country transfers can serve as a general response to low return rates. Claims about deterrence or improved compliance are also difficult to substantiate: third-country returns are embedded in a broader US enforcement agenda, including streng­thened border controls, intensified internal enforcement and expanded detention, making it hard to attribute any wider behav­ioural effects to this instrument alone.

Hence, returns at scale still depend on the core return system – aimed at direct return to countries of origin – rather than remote third-country arrangements. The US case underscores that even where the political, institutional and material conditions for third-country removals are favourable, expanding such agreements beyond established legal safeguards risks violating national and international law. In the case of the EU, these legal constraints are likely to be even more significant, because member states operate under overlapping human rights obligations, including the EU Charter of Fundamental Rights and the European Convention on Human Rights.

From legal opening to political momentum

The European experience and the US precedents point in the same direction: remote third-country transfers can serve a political signalling function, but they remain costly, selective and dependent on fragile bargains. Much of the controversy over return hubs turns on how this trade-off is assessed. For proponents, visible enforcement can reassure domestic electorates, demonstrate state capacity and potentially deter some prospective migrants. Critics, by contrast, question whether such effects are realistic, legitimate or proportionate to the financial costs, diplomatic concessions, legal risks and possible distortion of the wider return system. These competing assessments of the costs and benefits help explain why the mixed record of third-country transfers has not weakened political momentum behind return hubs in Europe. Now that the new Return Regulation provides the legal opening, an increasing number of Member States are seeking to translate the concept into practice.

Austria’s May 2026 mobility agreement with Uzbekistan illustrates the emerging search for partners. It provides for Uzbekistan to serve as a transit route for removals of third-country nationals, in particular Afghans. Although this agreement does not officially establish a return hub, Uzbekistan has reportedly been discussed as a potential host state, alongside Rwanda, Armenia and Mauritania. Persistent uncertainty over likely partner countries contrasts with growing political momentum within the EU. In recent months, a core group of member states – Austria, Denmark, Germany, Greece and the Netherlands – has been pushing for pilot return hubs by the end of 2026. Other EU member states have also expressed their support, calling for funding from the EU budget and the involvement of international organisations. France and Spain have rejected the concept, though elections in 2027 may alter their positions. After the Ceuta crisis, several European leaders and the EPP leadership used the inci­dent to renew calls for third-country return hubs, even though the incident constituted primarily a border-management and immediate readmission challenge rather than a return-hub scenario. At the time of writing, Rwanda has reportedly entered more formal negotiations with the countries belonging to the core group of EU member states, but concrete agreements are yet to emerge.

Even if this momentum leads to the estab­lishment of a pilot scheme, this would not change the basic scale problem. Any return hub operationalised by a European govern­ment or minilateral coalition would have to be understood as a residual instrument. Member states issue several hundred thousand return decisions each year, while the US experience suggests that a single hub would probably handle only a few hundred cases a year. It therefore cannot be expected to meaningfully increase the EU’s overall return rate.

Key risks associated with future return hubs

A first cross-cutting risk is diminished scrutiny after transfer. In ordinary return procedures, key legal risks are assessed before removal, including non-refoulement, access to remedies and protection against arbitrary detention or inhuman treatment. In a return hub arrangement, by contrast, some of the most serious risks may arise only after the person has left EU territory: detention conditions, access to legal representation and remedies, onward transfer, or refoulement. Because removal is legally complete upon arrival, political and administrative attention may shift away from the person concerned precisely when safeguards become harder to verify.

Another risk is weak additionality. A hub adds value only if it produces lawful and durable outcomes for people whose return could not be achieved through existing channels. It may also have indirect value if a credible prospect of transfer encourages direct return, but this effect would need to be demonstrated, not assumed. If the coun­try of origin is willing to document and readmit the person, third-country transfer merely adds cost and complexity, as the US experience suggests. If onward return remains impossible, the hub becomes less a temporary return instrument than a form of externally financed long-term admission or containment. Yet prospective host states are least likely to accept those whose return is hardest to resolve, including people whose identity cannot be reliably established.

A third risk is diversion. Once a facility exists, administrators may see incentives to populate it. Fixed costs, negotiated capacity and political sponsorship could encourage transfers to return hubs even in cases where direct return to country of origin might have been achieved through documentation, consular engagement, assisted voluntary return or reintegration support.

A fourth risk is policy entrapment. Return hub agreements require diplomatic concessions, infrastructure, funding, security arrangements and political capital. Once these costs have been incurred, poor results may lead not to closure, but to attempts to rescue the project through higher payments, weaker safeguards or diplomatic or political concessions. The EU’s own experience demonstrates the vulnerability that arises when operational control is delegated. Turkey’s 2020 border opening illustrates how migration pressure can become a bar­gaining chip. Similarly, recent US experience shows how a third-country removal agreement can be leveraged by partner countries: after accepting US deportees, South Sudan reportedly requested sanctions relief, visa concessions and US support for the prosecution of an opposition leader.

A fifth risk lies in a potential inversion of priorities. The core return system depends on determination of nationality, documentation, consular cooperation, direct return, readmission diplomacy, case management, voluntary return support and reintegration. These instruments possess less political visi­bility than a facility abroad. If return hubs divert attention and resources away from the core return system, they may end up weakening the system they are supposed to strengthen. Reintegration is especially vulnerable because its effects are difficult to measure and attribute. Yet this should not be mistaken for limited relevance. Reintegration support cannot by itself prevent future movement, but it can strengthen sta­bility and social cohesion in receiving communities as well as the willingness of countries of origin to cooperate on readmission.

Conditions for a limited pilot

Because pilots may differ by region and model, and may be negotiated bilaterally, common minimum standards are essential. The new Return Regulation requires an agreement or arrangement with the host state, but not a legally binding treaty. Participating member states should go beyond this minimum and commit to binding conditions on function, target group, legal responsibility, monitoring, evaluation and financing.