Myths & Realities about Return Hubs under the Return Regulation
By Dr. Daniel Thym, Universität Konstanz
Ironically, it was Manfred Weber, President of European People’s Party (EPP), who proposed limiting the room for manoeuvre during the negotiations on the Return Directive 2008/115. Two decades ago, Weber served as the EP’s rapporteur; his report rejected the Commission’s proposal to authorise the transfer to any third state, irrespective of ties or transit (here, Amendment 18). This time around, the EP’s majority took a different stance.
A Not Quite so Innovative Solution
Return hubs did not feature in the Return Directive 2008/115, nor have Member States actively pursued the idea in the past. So they are a novelty, albeit one with a backstory. EU institutions have always invested political capital in the expansion of destination countries for return (here, p. 8-9). Readmission agreements with Turkey, Serbia, and Ukraine, amongst others, cover non-nationals who have transited through these countries. Insistence on transit clauses was a stumbling block for the negotiations. Return to transit countries remained a rare occurrence, despite political pressure.
Recurring challenges in negotiating readmission agreements with a transit clause serve as a poignant reminder that it is easy to put return to third states on the ‘wish list’ in Brussels. Conversely, it can be tremendously difficult to identify partner countries that are receptive. The risk of a ‘phantom debate’ is real, with few third states being ready to host a return hub. Moreover, few would expect that more than a dozen or hundred persons could be brought there.
Return hubs deserve our attention despite these limitations. They may be no ‘magical solution’ for the myriad practical and political impediments for an effective return policy. However, they can be an important ‘control signal’ towards public opinion, prospective migrants, and other returnees. In this sense, they might serve as the symbolic hallmark of a holistic approach for specific migratory routes, or subcategories of returnees.
Critical Difference: No ‘Outsourcing’ of Asylum Procedures
In the absence of legislative change, ‘return hubs’ would have been illegal for the most part due to the narrow definition of ‘return’ in Directive 2008/115. The future Return Regulation puts forward a broader definition (here, Art. 4(3)(g)), together with a generic article for the establishment of return hubs (here, Art. 17). This abstract provision leaves numerous practical and legal questions unanswered. Its open-ended character arguably explains some of the confusion in the political debate.
To understand return hubs, we have to read Article 17 in the context of the EU asylum and return legislation. That is because these rules lay out the steps to be followed before the transfer. There are gaps in the legal design of the return hub concept under the Return Regulation, but the rough accusation that it undermines the right to asylum and is a manifest violation of the prohibition of refoulement misapprehends the regulatory framework. Challenges are real, but they are more subtle.
To start with, return hubs should not be confused with safe third country schemes, in the sense of ‘outsourcing’ of asylum procedures. EU asylum legislation facilitates such initiatives as well (here, Art. 59; here, Art. 57, 59), but the logic behind them is different. Safe third country projects, like the UK’s Rwanda plan, concern the transfer of asylum seekers on the basis of a decision that declares the asylum application inadmissible (here, Art. 38(1)(b)). Asylum authorities do not assess whether the applicant is a refugee or not, that is the responsibility of the third state.
By contrast, the Return Regulation requires illegal stay. Return hubs are for people without protection needs, not potential refugees. This critical difference explains why the legal requirements for return hubs differ from the higher standards for safe third countries. Most returnees will have the option to return to their home state on their own, despite the discontinuation of the mandatory period for voluntary return (here, Art. 12(1a), 46). Asylum seekers who shall be sent to a safe third country cannot be ‘rerouted’ towards a return hub either (here, Art. 59(9)).
Return Procedure: Additional Safeguards
Donald Trump’s return policy is subject to fierce criticism—rightfully so. The US administration has sent foreigners to torture prisons in El Salvador, or South Sudan; it has also deported asylum seekers without prior access to a procedure. All these actions would be illegal under EU law. A negative asylum decision does not mean that the person concerned can automatically be removed. Instead, domestic authorities must adopt a return decision designating the countr(ies) of return (here, Art. 7(4)).
This raises several legal issues. First, return hubs are not located in the home state by definition, meaning that the level of safety there has not been assessed during the asylum procedure already (here, Art. 3(5), (6)). Instead, Member States have to verify during the return procedure that the transfer to a return hub complies with the prohibition of refoulement. The Return Regulation says so (here, Art. 5, 7(4b), 14(1)(a), 17(1)), thereby reiterating Article 19(2) CFREU which is directly applicable and must be respected at all times, as the Court has emphasised repeatedly.
Second, return decisions are subject to legal remedies before a court, including where the location of the return hub is designated at a later stage (here, Art. 12(2), (2a), 26(1)). Unfortunately, EU institutions support misunderstanding when they say that ‘this Regulation does not provide for an automatic suspensive effect of the appeal’ (here, Art. 28(1))—a statement that seems to contradict the requirements for an effective remedy put forward by the ECtHR and the CJEU. This is not the case.
According to settled case law, domestic courts must be able to issue a suspension order (here, paras 46-57; here, paras 56-58; here, para 58). The Return Regulation provides for such an option as matter of principle (here, Art. 28(1)). It satisfies the requirements under human rights law, provided domestic courts apply its contents in a reliable manner. Concerns about practical deficits in some Member States are justified, but this does not render EU legislation illegal. Note, however, that corollary aspects limiting the grounds for court intervention ex officio (here, Art. 26(3), 28(1a)) might have to be set aside in light of Article 47(1) CFREU and pertinent case law. Judges will also have to ascertain that transfers do not take place during the period for lodging an appeal and up until their decision.
Return beyond Asylum: Pitfalls at National Level
The interplay of asylum and return proceedings gives rise to additional practical drawbacks. The effectiveness of the prohibition of refoulement depends on the procedural framework in place. Not very problematic are situations where the asylum authorities adopt the return decision together with the asylum decision (here, Art. 37). Asylum authorities have the expertise to assess the situation in various third states. They might have to update country information regarding the location of a return hub, but to do so is standard practice whenever a new situation arises.
More complex are scenarios where the location of the return hub is designated as a country of return later-on (here, Art. 12(2a)). This will often be done by the immigration authorities which may have a limited experience in assessing the principle of refoulement in some Member States, thus harbouring the potential of mistakes (here, p. 49-54). Against this backdrop, it was a wise decision to give Member States flexibility to entrust this task on the asylum authorities (here, Art. 12 (3a)). This scenarios is equally valid for foreigners who did not apply for asylum and receive a return decision for other reasons, for instance criminal records or overstaying a visa.
Finally, the prohibition of refoulement under the Charter and the ECHR is wider than the meaning of international protection under the Asylum Qualification Regulation 2024/1347 (here, p. 398-405). It can cover three scenarios, in particular, in which an asylum application will normally be rejected: ‘medical cases’ where the applicant faces immense suffering and hardship (here, paras 60-72), ‘family ties’ when spouses or minor children have a right to reside, and scenarios of ‘exclusion’ when an asylum seeker has committed war crimes abroad or very serious crimes in the host state (here, Art. 12, 14, 19).
Many Member States have established various forms of ‘complementary protection’ for medical cases and similar scenarios, meaning that beneficiaries obtain a residence permit. When this is not the case, there must be an assessment on a case-by-case basis whether the transfer to the return hub would be justified. Again, the legal situation for the non-refoulement assessment may be straightforward, but it harbours the potential for mistakes.
Uncertainties Regarding Practical Implementation
Article 17 Return Regulation adds little to the procedure to be followed before the transfer. Its focal point is are bilateral relations between Member States (or the EU) and the third state in which the return hub will be located. The contents are decidedly open-ended and give Member States much leeway, thereby nurturing a pervasive sense of uncertainty about the return hub concept.
In essence, two options are available: Member States may regard the ‘return hub’ as a provisional destination for onward movement to the home state. The term ‘hub’ suggests such a set-up. The Return Regulation does not specify the length of the stay in this respect, whether it be weeks or months if not years. Alternatively, the third state can be the final destination for longer sojourn, unless the returnee decides voluntarily to go elsewhere. We may speak of ‘return centres’ in this case. Article 17 covers both scenarios, thus entrusting the Member States to hammer out the details during the diplomatic negotiations with the partner country.
Numerous questions follow from this abstract categorisation: Are we speaking about ‘camps’, possibly involving detention, or decentralised accommodation? Who will provide the accommodation and run potential centres: Member States, the third state, IOM and UNHCR, or private service providers? Will there be assisted voluntary return packages—or forced—removal to home states and if so by whom? Can money from the EU budget be used to support Member State initiatives? Can Frontex possibly play a role? Will returnees obtain a residence permit or have labour market access in case they stay in the third state for longer periods?
Article 17 remains silent on these issues, in striking contrast to the guarantees for asylum applicants in safe third countries (here, Art. 57). This silence can be approached from a political and a legal perspective. Politically, it appears self-evident that Member States should be held accountable for the conditions that prevail in return hubs. The idea behind Article 17 is to introduce an—albeit abstract—regulatory framework for their operation. In contrast, obligations under human rights law are less clear.
After the Transfer: the Return Regulation Stops Applying
A critical juncture for the legal analysis is the definition of country of return. It shall be the third state where the return hub is located (here, Art. 4(3)(g)). From a legal perspective, this definition means that return process is completed once the individual arrives at the designated return hub. That is a crucial difference between the Return Regulation and the Italian centre in Albania, which serves as an intermediate step in an ongoing return procedure—an option that is permitted under the Return Directive 2008/115 according to two Advocates General in the cases of Sedrata and Comeri & Sidilli.
Conversely, the Return Regulation declares the return process to be finished upon the transfer. As a result, any additional steps, notably voluntary return or forced removal to home states, are not subject to the procedural and substantive safeguards in the Return Regulation. Instead, the third state, or international actors, apply the domestic laws of the third state, as well as the treaty or arrangement under Article 17 Return Regulation. The opinion of AG Medina about mandatory respect for the safeguards in the Return Directive in Albania presupposes an ongoing return procedure (here, paras 60-62, 75-90). It cannot be projected upon return hubs.
In practice, the completion of the return process means that the legislative provisions in the EU asylum and return legislation on detention, legal remedies, cooperation in return, subsequent asylum applications, etc do not apply. Human rights take over as the pertinent benchmark. In that respect, we have to distinguish between the obligations incumbent on the Member States before the transfer and the query as to whether they can be held responsible for the treatment in the third state.
Mandatory Respect for Non-Refoulement
Settled case law on the prohibition of refoulement requires Member States to verify as to whether the person concerned would face a ‘real risk’ of torture, degrading, or inhuman treatment upon return to a specific country, include the risk of ‘chain-refoulement’ to another third state (here, paras 77-105). To do so is an everyday practice during asylum and return procedures. Domestic authorities and courts will employ these standards to assess whether the location of the return hub is sufficiently safe.
Crucially, this assessment happens before the transfer, in accordance with the procedural safeguards described earlier. The transfer to a return hub is not any different from regular asylum and return procedures, except for the focus on a country whose nationality the returnee does not have. Nevertheless, there are a several legal uncertainties the return hub concept gives rise to. Three comments offer food for thought for experts and courts as they try to pinpoint the relevant benchmarks.
First, Member States will apply the wide European understanding of non-refoulement, including persecution by non-state actors and situations of general violence, irrespective of whether they are mandatory under the Refugee Convention. There will be no ‘rebate’ for return hubs. Having said this, the outer edges of the case law remain ambiguous. With regard to socioeconomic living conditions, the stricter standards for ‘medical cases’ cover return scenarios, although some will try to argue that the elevated level of protection for asylum seekers under the Dublin system should apply (here, p. 23-27; here, p. 352-353). These higher standards foresee access to housing and other social benefits, which are not usually a required for return.
Second, judgments on non-refoulement have centred on mistreatment by – destination countries, including miserable detention conditions. Conversely, the issue whether detention as such can give rise to a prohibition of refoulement has not been the subject of extensive litigation. We know for certain that Member States need not guarantee full respect for Article 5 ECHR, which, rather, will only prevent a transfer when there is a ‘flagrant denial’ of the right to liberty (here, paras 231-233; here, paras 93-96). The CJEU came to the same conclusion by invoking the ‘essence’ of human rights as the relevant benchmark. The contours of the lower standard for the ‘flagrant denial’, or ‘essence’, remain indeterminate. The uncertainty is amplified by the predominance of extradition, not return, in the existing case law.
Third, the final text excludes unaccompanied minors, not, however, families with young children (here, Art. 17(4)). Article 24 CFREU applies nonetheless whenever Member States adopt a return decision. As a consequence, families with children cannot usually be transferred to a return hub, in accordance with innovative CJEU rulings. This is evident whenever the return hub serves as an intermediate destination for eventual return to the home state at some point in the future (here, paras 43-54). Insistence on an option that is unlikely to be implemented was not a politically smart move, at the price of losing the support of many liberal MEPs.
Application of the Charter to the Operation of the Centres
A decisive legal battleground will be the territorial scope of human rights obligations. While it is beyond doubt that they cover the adoption of the return decision, it is unclear whether they cover the operation of return hubs if they are—unlike the Italian centres in Albania—operated by the partner country or international actors (here, Nos 84-91). The ECtHR will probably hold that Member States do not have jurisdiction under Article 1 ECHR, given that it has recently rejected a functional understanding of state responsibilities (here, paras 78-83, 91-97, 107). In the Chișinău Declaration, State Parties exert pressure on the Strasbourg Court not to submit return hubs to the ECHR by means of changing course.
The Charter will apply, in accordance with Article 51(1), whenever we conclude that Member States ‘only when they are implementing Union law’. In this respect, the silence of the Return Regulation on the legal and practical parameters for the operation of return hubs might prove decisive. In the absence of legislative prescriptions, it will be an uphill struggle to maintain that the Charter, as the proverbial ‘shadow’ of an obligation under EU law, will be applicable (here, paras 34-35; here, para 26). Frontex activities might support a different outcome—a prospect that might disincentivise governments from having recourse to the agency to start with.
Against this backdrop, the legal consequences of the mandatory respect for ‘international human rights standards and principles’ remains uncertain (here, Art. 17(1)). It is not apparent whether the provision governs the choice of partner countries and/or the operation of return hubs. In any case, it does not specify what the pertinent ‘standards and principles’ shall be, presumably the UN Covenants. The Return Regulation does not specific the requisite degree of safety either. Isolated human rights violations happen in every country, so the standard seems to be about an abstract safety assessment. EU institutions do not explain what they had in mind, nor is it certain whether individuals can invoke the provision to object the transfer in situation where the prohibition of refoulement would not be violated.
Memorandum of Understanding or International Treaty
Respect for the prohibition of refoulement centres on the forward-looking prognosis whether there is a real risk of inhuman treatment after the transfer. By contrast, the adoption of the return decision and its enforcement are not contingent upon the conclusion of a legally binding treaty or non-binding arrangement with the destination country. Germany can repatriate a Nigerian without the need to sign a readmission agreement or memorandum of understanding. Some sort of bilateral cooperation may be warranted, so as to render interstate cooperation effective, but it is not a legal prerequisite.
As a result, the obligation to conclude a bilateral treaty or arrangement, in Article 17(1), goes beyond human rights law. Diplomatic assurances would only be mandatory if there was a real risk of refoulement without them; they can contribute to dispelling any doubts regarding the legality of the transfer. The ECtHR has laid down a collection of factors national courts shall take into account when assessing the credibility of diplomatic assurances; binding legal force is only one element amongst many others (here, paras 187-189; here, paras 129-131). A monitoring mechanism can be another element supporting credibility (here, Art. 17(2a)(d)). It may be politically desirable even if it is not a legal prerequisite for the transfer.
The Return Regulation stipulates that diplomatic negotiations will have to determine ‘the consequences to be drawn in case of violations of the agreement or arrangement or significant change adversely impacting the situation of the third country’ (here, Art. 17(2)(f), (2a)(f)). They will also have to say what happens when returnees do not return to their home state, either voluntarily or by force (here, Art. 17(2a)(a)). Member States have much leeway in this respect. It is not even apparent whether the ‘consequences to be drawn’ are about the future-oriented suspension of bilateral cooperation, or possibly about the readmission to the EU of those third country nationals who had been transferred to the return hub previously.
A final comment pertains to the added value of a legally binding treaty. The prevailing assumption is that individuals have less rights under soft law instruments. This a last myth our contribution puts to rest. Judges in Luxembourg have established a series of criteria that must be met in order for individuals to invoke an international treaty: its nature and broad logic must support direct effect and the provision concerned has to be precise and unconditional (here, para 45). Most readmission agreements are not directly applicable in this sense, in the absence of extended guarantees for returnees. Their contents focus on practical aspects, such as contact points or the organisation of return. The same might turn out to be true for the ‘respective obligations and responsibilities of the parties’ in future return hub agreement (here, Art. 17(2)(a)-(c), (2a)(b)).
This reminds us of our initial message: an isolated inspection of the open-ended prescriptions for return hubs in Article 17 Return Regulation paints an incomplete picture. Anyone trying to grasp the pertinent legal framework and salient practical challenges has to read this abstract proviso in light of the broader EU asylum and return legislation.
The initial version of this blogpost refers to the numbering of the compromise text for the Return Regulation. Article numbers will be updated for publication in the Official Journal.