Testing the Applicability of EU Law Abroad: The Italy–Albania Protocol in the Comeri, Sidilli, and Sedrata Hearings

 

Eleonora Celoria, Post-doctoral researcher, FIERI

Andreina De Leo*, Post-doctoral researcher, Maastricht University

Marcella Ferri, Research Fellow, Florence University

* Funded by the European Union (ERC, SoftEn project, 101165167, PI: Lilian Tsourdi). Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the European Research Council. Neither the European Union nor the granting authority can be held responsible for them.

Photo credit: Jorge Franganillo, via Wikimedia Commons

Introduction

On 23 and 24 March 2026, the Court of Justice of the European Union (CJEU) held two hearings in the ongoing proceedings addressing the compatibility of the Italy-Albania Protocol with EU law and the broader implications of conducting asylum and return procedures extraterritorially (Joined cases C-706/25 Comeri and C-707/25 Sidilli, and C-414/25 Sedrata).

This may appear as a déjà-vu. Less than a year ago, the extraterritorial processing of asylum under the Italy-Albania Protocol had already come before the Court in the well-known Alace and Canpelli cases (discussed on this blog by Zamboni and Favilli & Marin). The focus of those cases was, however, narrower, as it concerned the criteria to designate countries as “safe countries of origin” under the Asylum Procedures Directive (APD) and their judicial review, a classification that directly enabled the use of accelerated offshore asylum procedures in Albania. Following those rulings, the Italian Government adjusted its strategy, implementing a policy based on the relocation to Albania of returnees already detained in Italian pre-removal detention centers under the Return Directive (RD). While awaiting removal, these individuals may lodge an asylum application which, according to the Government, can then be examined in Albania. Against this backdrop, the new preliminary references raise a distinct and more direct question: whether asylum and return procedures can, as such, be carried out in a third country without breaching EU law. In doing so, they place the legality of the Protocol as a whole squarely before the Court.

More specifically, the first preliminary ruling, submitted by the Court of Cassation on 20 June 2025, asks whether Articles 3, 6, 8, 15, and 16 of the Return Directive preclude transferring a migrant subject to a return order to an extraterritorial detention centre, even without a concrete prospect of removal. It also asks whether Article 9(1) of the Asylum Procedures Directive, which requires applicants to remain on a Member State’s territory, precludes keeping asylum seekers who lodge an asylum application after being transferred to a third-country detention centre within that country (see De Leo).

The second preliminary ruling, submitted by the Rome Court of Appeal on 5 November 2025, directly questions Italy’s competence to conclude an international agreement  interfering with an area of exclusive EU competence, as laid down in Articles 4(3) TEU, 3(2) TFEU and 216(1) TFEU (see Montaldo and delli Carri). Alternatively, should the Court find that Italy has such competence, it asks whether the safeguards under the APD and the Reception Conditions Directive (RCD), as well as the Charter, particularly the rights to liberty, an effective remedy, and health, preclude transferring asylum seekers to, and detaining them in, centres outside EU territory.

It is worth recalling that the stated aim of the Protocol is to “reinforce bilateral cooperation in managing migration flows,” allowing Italy to transfer migrants to Albania solely to carry out asylum and return procedures under Italian and EU law (Art. 4(3) Protocol). These procedures remain under Italian jurisdiction and are conducted in accordance with national and EU law “to the extent compatible” (Art. 4(1), Ratification Law 14/2024).

Although raising different questions, both Italian courts focused on the same core issue: whether EU law safeguards can be effectively ensured for asylum seekers and returnees detained in a third country and whether those standards are compromised merely because detention occurs outside the territory of a Member State.

Outline of the post

While awaiting further developments, and notably the Advocate General’s opinions on 23 April (Sedrata) and 11 June (Comeri and Sidilli), this post examines the key issues raised during the hearings and outlines the main arguments on whether asylum and return procedures conducted in a third country, yet under a Member State’s jurisdiction, comply with EU law.

It first considers whether EU law applies to procedures carried out in Albania and whether such procedures may interfere with its application, summarising the positions of the Italian Government and the European Commission. Two aspects are examined: the Dublin system and its founding principle of mutual trust, and whether EU safeguards can be effectively guaranteed in detention centres abroad. Finally, the post clarifies the differences between the Albania model, the so-called “return hubs,” and the notion of a safe third country, as frequently referenced during hearings and in public debate.

EU Law Beyond Territory: Parties’ Arguments on Applicability, Compliance, and Competence in the Italy-Albania Scheme

The Italian government

To justify the compatibility of the Italy-Albania Protocol and its Ratification Law with EU law, the Italian Government advanced two alternative lines of argument. Both were aimed at demonstrating that: first, the Protocol does not risk adversely affecting the internal rules of the Union and therefore does not encroach upon an area of exclusive EU competence; secondly, that the extraterritorial application of the relevant directives in a third country does not undermine their effet utile.

First, the Government argued that the Albanian centres should be considered “representations of Member States” abroad under the APD and RCD, and that EU law applies indirectly through its incorporation via the Ratification Law. On this basis, the Government maintained that the Italy-Albania Protocol is purely technical and logistical: it does not alter the substance of the legal framework on pre-removal detention, asylum procedures or reception conditions, which remain fully applicable because the legislator has chosen to extend EU standards to the situations covered by the Protocol. By virtue of this legal extension, the Government maintained that third-country nationals detained in Albania receive treatment equivalent to that in Italy. It further argued that instruments such as the Dublin Regulation, though not explicitly mentioned in the Protocol, continue to apply because the individuals remain continuously under Italian jurisdiction. Accordingly, the Protocol would neither interfere with the Common European Asylum System (CEAS) nor encroach upon the Union’s external competences.

With respect to the RD, the Government argued that the transfer to Albania does not qualify as a “return” within the meaning of Article 3, but it rather amounts to a temporary relocation to a different pre-removal detention facility, legally assimilated to ones located on Italian territory. Accordingly, it maintained that compliance with the Directive’s objectives and provisions, including Articles 15 and 16 on detention and conditions of detention, are not undermined, as the material conditions would be fully respected. Regarding the APD, the Government contended that Article 9 is not violated when returnees apply for asylum after being transferred to Albania, since they remain under Italian jurisdiction, and thus do not risk refoulement, and continue to benefit from the full application of both the APD and RCD.

Alternatively, the Government argued that EU law could apply directly through a functional interpretation of “territory.” In the absence of a clear EU definition of “border” or “transit zones,” the centres in Albania could be regarded as a fictio iuris, functionally equivalent to transit zones located on Italian soil. On this view, the legal equivalence of procedures in Albania and Italy would justify treating the Albanian centres as falling within the scope of EU law, in the same way as their territorial counterparts.

While the Dublin Regulation was not explicitly mentioned by the referring judges, it emerged as a contentious issue. In particular, the risk of circumventing Dublin criteria for unaccompanied minors (Article 8) and family members (Articles 9–10) was debated during the Sedrata hearing. The Italian Government maintained that the Regulation remains applicable because the Albanian centres qualify as “transit zones” under Dublin III. This, however, raised a fundamental question from the bench: on what legal basis could another Member State be required to accept a Dublin transfer from a non-EU centre, and how would this not interfere with common EU rules? Questions which were left unresolved.

The European Commission

The Commission, by contrast, adopted a strict literal reading of the directives, distinguishing clearly between jurisdiction and territory. While acknowledging that the individuals remain under Italian jurisdiction, it stressed that asylum and reception rules are territorially anchored and do not apply automatically outside Member State territory. Nonetheless, the Commission did not view the Protocol as violating EU law or encroaching on exclusive EU competences, so long as the legislator’s choice to apply EU standards effectively preserves the effet utile of the acquis. Its reasoning followed two separate lines, depending on the instrument in question.

As for the RD, the Commission noted that it does not strictly define its scope in territorial terms but rather sets standards to ensure the effectiveness of returns from EU territory. Since Italy has committed to applying these standards in Albania, the Commission did not see any inherent incompatibility with the temporary transfer of returnees outside the territory pending repatriation. Yet, this conclusion was subject to strict conditions: the Commission emphasised that detention should have been previously judicially authorised, that all guarantees under Articles 15–18 of the Directive are fully respected, and that the transfer to Albania must not constitute the execution of the return decision. In essence, for the Commission the Directive does not require detention to occur on the territory of the Member State: what is rather decisive is that the Member State fully complies with its concrete obligations.

As for the APD and the RCD, the Commission rejected the qualification of the Albanian centres as “transit zones.” It stressed that the notion of territory is autonomous under EU law and cannot be extended to third countries, and that thus a “transit zone” must be located within the Union’s territory. When pressed by the bench on whether an international definition of “transit zone” exists, it indicated that this was ultimately irrelevant, as the notion must be understood as a matter of EU law. Accordingly, the Protocol cannot extend EU territory, and the direct applicability of the asylum acquis is thus excluded.

That said, the Commission did not view the mere extension, via national law, of EU standards to procedures in Albania as automatically undermining the directives’ objectives or interfering with internal EU competences under the ERTA doctrine. Such a violation would only arise if the arrangement risked affecting common EU rules: a risk the Commission considered unproven here. Article 9 APD illustrates this approach: if the asylum application is made in Italy, transfer to Albania is precluded. Conversely, if lodged only after transfer, while the Directive does not apply directly, its purpose, i.e. to prevent removal before assessing refoulement risks, is fulfilled because Italy retains jurisdiction and applies EU standards.

Regarding the Dublin system, the Commission held that the Regulation does not apply outside EU territory, revealing a potential paradox noted by Advocate General Emiliou during Sedrata: if EU law does not formally apply, can it still be circumvented?

More broadly, the question arises whether exclusion from the scope of the Regulation and Directives is merely a consequence of its incidental inapplicability in a third country, or the result of a deliberate choice by Italy, and namely, transferring individuals to Albania. In the latter scenario, how can it be argued that the Protocol and Ratification Law do not effectively allow Italy to evade its obligations under EU law?

As President Lenaerts highlighted, the key concern of the ERTA doctrine is preventing Member States from using external bilateral agreements to regulate matters that fall within EU law, in ways that could undermine its objectives. In other words, it is not enough to claim that EU law does not directly apply in a third country, but what matters is whether the external action could affect the effective internal functioning and uniformity of the EU legal framework. In this respect, the Commission’s position appeared unsatisfactory, as it failed to address the broader systemic implications for the CEAS and the structural risks of circumvention inherent in the scheme.

This brings us to our legal assessment, where we examine the unresolved issues in the approaches of the Italian Government and the Commission regarding the challenges of extraterritorialising asylum procedures and applying EU law in a third country.

Assessing the Potential Interference of Procedures Conducted in a third country with the EU Migration and Asylum Acquis

All in all, both the Italian Government and the European Commission ultimately arrive at a broadly similar conclusion: the Italy-Albania Protocol does not entail a violation of the Union’s exclusive competences, nor does it compromise the effet utile of the relevant directives, insofar as EU standards can be ensured, even in a third country. However, this conclusion leaves several key issues unresolved. First, it does not fully address the implications for the Dublin system. Indeed, the extension of Dublin obligations outside the Union could highly interfere with the CEAS and the principle of mutual trust. Second, the assumption that equivalent standards can be guaranteed extraterritorially remains highly problematic, as the territorial scope of EU asylum law reflects structural constraints necessary to ensure full compliance with procedural and substantive guarantees. These two issues will be addressed separately in the following sections.

The (application of the) Dublin system: a paradigmatic interference with the CEAS

The application of the Dublin system emerged as a highly sensitive issue during the Sedrata hearing. While EURODAC was not discussed, the reasoning for the Dublin Regulation can be extended to it, in light of recital 54 of EURODAC Regulation. The Italian Government and the Commission took sharply divergent positions: the Government qualifies the centres as “transit zones” under Dublin, a view opposed by the Commission, which raises several critical questions.

First, how can the Albanian centres be considered “representations of Member States” under the APD and the RCD, or “transit zones” under the Dublin Regulation? This argument is unpersuasive, as it would imply that the notion of “transit zone” differs across two CEAS instruments, risking to undermine the system’s internal coherence.

Secondly, if the Dublin system were applied to asylum requests lodged in the Albanian centers, other Member States would be obliged to accept transfers, as the Italian Government argued. From a theoretical perspective, this would directly interfere with the CEAS and undermine the principle of mutual trust on which it rests (Lenaerts, 2017). In fact, the presumption of compliance with EU law standards appears questionable in itself when it is applied to an asylum request lodged outside the territory of a Member State — albeit under its jurisdiction. Yet, save for specific exceptions, that principle prevents other States from verifying whether, in a given case, Italy has actually complied with EU law in processing an asylum application, notably that it has allocated responsibility in accordance with the procedural guarantees laid down in the Dublin Regulation and with the right to an effective remedy against the transfer decision. In short, extending the Regulation to applications lodged in the Italian centres in Albania would run counter to the principle of mutual trust underpinning the Dublin system and, ultimately, the CEAS.

Third, from a practical perspective, applying the Dublin Regulation outside Italy would adversely affect the rights of asylum seekers lodging an application in another Member State and seeking reunification with a family member detained in Albania (Article 10 of the Dublin Regulation). Unless the applicants were also transferred to Albania — which would amount to their de facto detention — they would remain in Italy, thereby rendering effective reunification  impossible. This impossibility, which stems directly from the detention of the family member in Albania rather than in a detention facility on Italian territory, would breach their rights to family life and to the child’s best interests, as laid down in the Dublin Regulation and in Articles 7 and 24 of the Charter.

By contrast, if, as the Commission argues, the Dublin Regulation does not apply, Italy could evade its obligations under Articles 8— 10 of the Regulation and Articles 7 and 24 of the Charter. Once again, such circumvention would be achievable simply through the deliberate extraterritorialisation of asylum procedures. In a nutshell, whether applied or not, the Dublin Regulation illustrates the systemic and structural interference with the CEAS created by the Protocol and Ratification Law.

Ensuring Effective Application of  EU Standards and Rights in a Third Country: Mission Impossibile?

We concur with the Commission that the territorial nature of EU asylum law imposes inherent limits on extending its directives beyond the Union. However, we diverge from the view that voluntarily applying EU standards abroad via national law automatically resolves compatibility concerns. This divergence is anchored in the structural reality that the territorial scope of the directives is not merely formal: it reflects the impossibility of fully guaranteeing procedural and material standards outside the Union. Applying EU law where these guarantees cannot be secured risks undermining the directives’ objectives, the uniformity of EU law, and may raise concerns under the ERTA doctrine (see: De Leo & Celoria, and Montaldo).

As highlighted by the referring judges and the lawyers representing the migrants, a central issue is the gap between the guarantees afforded to migrants detained in Italy and those in the Albanian centres. The Government asserts that procedures in Albania are identical to those in Italy, yet legal equivalence on paper does not ensure effective protection in practice. In our view, the fiction that Albanian centres are legally treated as part of a Member State is insufficient to secure genuine compliance with EU standards on the ground.

The main concerns identified by the referring courts and discussed during the hearings include: the right to liberty under Article 6 of the Charter, notably the requirement for immediate release if detention is unlawful; guarantees related to detention conditions, including access for lawyers, family members, national, international, and non-governmental organisations, and access to healthcare; and the right to a fair trial and effective remedy under Article 47, particularly the right to be “advised, defended, and represented.” As emphasized by the rapporteur judge of Comeri and Sidilli case, these discrepancies risk creating a distinction between “two types of asylum seekers,” or even “two types of detainees,” solely based on where and when an application is lodged.

Both the RD and the RCD require that a person “shall be released immediately” if detention is found unlawful (Article 15(1) RD; Article 9(3) RCD). This requirement is inherently impossible to reconcile with the Italy–Albania scheme. Under Article 6(2) of the Protocol, Italian authorities must “take the necessary measures to ensure the permanence of migrants within the areas, preventing their unauthorised exit into the territory of the Republic of Albania, both during and after the completion of administrative procedures, irrespective of the final outcome.” In practice, release on Albanian territory is impossible: individuals can only be freed once transferred back to Italy. Therefore, even when detention is considered unlawful, the person remains under continuous coercive control throughout the transfer, i.e. from the detention center to the port, during maritime transport, and until release in Italy (typically in Bari or Brindisi): a process that may last from 24 hours to several days, depending on logistical conditions. In this respect, the Italian National Guarantor has recently noted that, if transfer cannot occur the same day, the person remains in a “separate area” of the same detention centre, a practice arguably failing short of the “immediate release” requirement. During this period, individuals are subject to a “coercive measure that deprives [them] of [their] freedom of movement and isolates [them] from the rest of the population” (CJEU, FMS, para. 223). Both the Albanian centres and the transport arrangements display the defining features of detention: a closed, restricted space where movements are limited and monitored and exit is not voluntary (CJEU, FMS, para. 231). The European Court of Human Rights similarly confirms that confinement on ships may constitute de facto detention, even if nominally for the person’s interest (Khlaifia and Others v Italy, paras. 70–71). In sum, the extraterritorial setting alters the legal consequences of detention, making immediate release practically unfeasible and raising serious questions about compliance with EU law.

Extraterritoriality also affects effective access to detention centres. Under national law implementing EU standards (Articles 16(2) and (4) RD; Article 10(4) RCD), a wide range of actors, e.g. parliamentarians, national monitoring bodies, UNHCR representatives, lawyers, family members, religious ministers, and civil society organisations, should be granted access (Articles 21 DPR 394/99; 7 D.Lgs. 142/2015; Ministerial Directive 19 May 2022). By contrast, Article 9(2) of the Protocol limits access to “lawyers, their assistants” and “international organisations and EU agencies,” raising questions about the interplay of the two regimes. Even if Italy seeks to apply its broader framework, as argued during the hearing, effective implementation depends on Albanian authorities’ consent. This is because access requires entry into Albanian territory, which remains under Albanian sovereignty. Article 9(2) subjects access to “applicable EU, national and Albanian law,” while Article 6(1) confirms Albanian responsibility for public order and security outside the centres and during transfers. Consequently, Albania may lawfully restrict access based on domestic entry, residence, or public order rules, limiting in practice the actors able to monitor detention conditions. While the Italian Government argued that this poses no practical concern, citing visa-free entry for holders of Italian passports or residence permits, the effectiveness of EU law guarantees cannot depend on conditions outside the Member State’s control. Making access contingent on the rules or discretion of a third country introduces uncertainty incompatible with the requirement to ensure EU rights in an effective and uniform manner. Moreover, denials imposed by Albanian authorities cannot be effectively challenged before an Italian court, which undermines the right to an effective remedy for those having the right to access the detention centres.

Similarly, under Article 10(4) of RCD, lawyers, family members, and civil society organisations have the right to communicate with and visit detainees, and any limitations must not make access impossible or excessively difficult. In the Albanian centres, however, restrictions appear structural and generalised rather than exceptional, making them difficult to reconcile with Articles 7 and 47 of the Charter and the proportionality requirement under Article 52(1). The Government’s argument that similar limitations exist in Italy is unpersuasive: in Albania, restrictions are the norm, access is more complex, time-consuming, and costly, and a recent monitoring report by the Italian National Guarantor for the Rights of Persons Deprived of Liberty notes that some detainees were not informed of their transfer, further hindering visits. Crucially, as noted above, the ultimate decision rests with Albanian authorities, meaning that limitations stemming from a third State’s sovereignty cannot be effectively challenged or remedied in Italy, thereby weakening the effective enjoyment of EU-law-derived rights.

Additionally, discrepancies in safeguards for lawyers are particularly pronounced. In Italy, appointed lawyers can access detention centres without prior authorisation and often be appointed in person after initial telephone contact. In Albania, this is not possible, nor foreseen by the Protocol or the Ratification Law, which allows travel at public expense only “when remote connection is not feasible” during the detention validation hearing (Article 4(5), Law 14/2024). This effectively precludes in-person meetings before or after the hearing to prepare a defence or appeal a negative decision. Given the current five-day deadline to challenge a validation decision, detainees without immediate lawyer access face heightened risk of ineffective representation. In-person meetings are often essential to identify vulnerabilities or health conditions that may render detention unlawful, tasks difficult to perform remotely, particularly when access depends on the private centre manager’s discretion and mobile phone use is restricted. Physical distance, limited contacts, and communication restrictions materially weaken lawyer-client interaction, undermining the right to effective assistance and representation under Article 47 of the Charter. Remote hearings further exacerbate these issues. By analogy with the Court’s reasoning in FP and Others (C‑760/22), videoconference participation is not inherently incompatible with a fair and public hearing, but its appropriateness depends on the individual being able to follow proceedings, be heard without technical obstacles, and communicate effectively and confidentially with their lawyer. In Italy, lawyers attending remote validation hearings can choose to be in the courtroom with the judge or at the centre with their client: a flexibility that is simply unavailable for detainees in Albania, further undermining effective representation.

Finally, significant disparities exist regarding the right to health, guaranteed under Article 16(3) RD and Article 17(2) RCD. Detainees in Albania face substantial limitations in medical services, particularly psychiatric and addiction care. Unlike in Italy, where detainees are integrated into the National Health System, no equivalent framework operates in Albania, and on-site medical teams cannot provide comparable treatment. These deficiencies have already led to serious consequences, including self-harm and suicide attempts (see, report, here). A recent report by the National Guarantor for the Rights of Persons Deprived of their Liberty found that, of 26 detainees for whom updated medical assessments were requested, 25 were deemed unfit for detention and released, highlighting the centres’ inadequacy to ensure proper care. While Article 4(8) of the Protocol obliges Albanian authorities to provide treatment in emergencies beyond Italian capacity on site, healthcare in a third country is by definition different from that provided by a Member State’s National Health System.

Taken together, the disparities analysed above make detention in centres located in a third country significantly more burdensome than detention in Italy, casting doubts as for its compliance with the principle of proportionality. There is no justification for imposing a harsher coercive measure than would be applied in Italy, particularly as the transfer to Albania does not demonstrably enhance the effectiveness of return procedures under the RD. On the contrary, a report by the Italian National Guarantor for the Rights of Persons Deprived of their Liberty shows that only 56 of 192 individuals transferred to Albania were ultimately returned to their country of origin, representing a return rate of approximately 29%, which is lower than the general return rate.

Conclusion

In conclusion, while the Italian Government frames the issue in terms of formal equivalence between procedures carried out in Albania and those in Italy, and the European Commission emphasizes that the voluntary application of EU standards suffices to ensure compliance, our assessment shows that this approach does not adequately address the structural limitations inherent in the de facto extraterritorial application of EU migration and asylum law.

While awaiting the Court’s decision, we maintain that the strict territorial scope of EU law in this area is not merely formal but reflects a substantive requirement: it is meant to operate within the territory of Member States, where its guarantees can be fully implemented and monitored. Remaining under the jurisdiction of a Member State alone is insufficient to ensure compliance, as EU law cannot be effectively applied outside the Union without undermining its purpose, coherence, and the procedural and material protections that are inherently tied to the territorial context in which they are guaranteed.

Two key issues arise. First, if the Dublin Regulation applies to applications lodged in Albania, the Protocol and Ratification Law would undermine mutual trust and the functioning of the Dublin system, since other Member States cannot verify compliance with EU law outside the Union. Even if it does not apply, the deliberate delocalization of asylum procedures to a third country allows circumvention of EU obligations. Second, procedural and material safeguards ensured in Italy cannot be fully replicated in a third country.

Thus, in our view, relocating individuals to Albania undermines the effet utile of the CEAS, creating tangible discrepancies in treatment that formal equivalence cannot remedy. This, in turn, risks breaching the principle of loyal cooperation and jeopardizes both the uniform application of EU law and the trust-based cooperation that underpins the EU legal order.

A Final Note: Why Albania Is Different from Other Externalisation Models

Throughout the hearings, “return hubs” were frequently cited as a possible analogy to justify the legality of the Albania model. However, as repeatedly emphasized by all parties, the two schemes are fundamentally different from a legal perspective. The Commission’s proposed reform of the notion of a ‘return decision,’ which provides the legal basis for return hubs, presupposes a formal removal decision to a country other than the state of origin with which an agreement exists. In that context, the transfer constitutes an actual return under EU law (discussed, in this blog, here). By contrast, the Albania model does not involve a removal but a temporary relocation of the detention stage of the return procedure under the full jurisdiction of the Member State, prior to any formal removal to the individual’s country of origin or habitual residence. Consequently, it cannot be assimilated to the return hub framework.

Similarly, the Albania model cannot be framed within the reformed notion of a safe third country without a connection requirement (discussed, in this blog, here), as no inadmissibility decision is taken in Italy based on Albania’s acceptance of the asylum application. Instead, the procedure merely relocates the processing of the asylum claim to Albania while keeping the substantive application under Italian jurisdiction.

This distinction highlights that equating the Albania model with return hubs or safe third country schemes overlook its unique legal and procedural features. It underscores the complexity of assessing its compliance with EU law and why the pending preliminary rulings are essential to clarify the boundaries and practical limits of this peculiar migration management scheme.

The European Parliament’s role in shaping the EU rules on return: from safeguards to securitisation

 

 

Dr. Izabella Majcher, Independent Consultant

Photo credit: Alamy stock photo

 

The current revision of the EU rules on return has progressed rapidly since the European Commission published its proposal for a Return Regulation on 11 March 2025 (discussed here and here). The Council of the European Union adopted its “general approach” to the proposal on 8 December 2025 (discussed here). In addition to numerous critical opinions from civil society organisations (see, for instance, here, here and here), on 26 January 2026, in an unprecedented step, 16 UN experts addressed a joint letter to both the Commission and the Council expressing serious concerns about the proposal. The measures attracting particular criticism included weakened procedural safeguards; expanded immigration detention; the de-prioritisation of “voluntary” departure; increased cooperation obligations, penalties and restrictions; risks of racial and religious profiling; limitations on access to socio-economic rights; insufficient protection of persons in vulnerable situations; and the introduction of “return hubs”. To date, neither institution has responded to the UN experts.

On the side of the European Parliament, within the Committee on Civil Liberties, Justice and Home Affairs (LIBE), the rapporteurship for this file was assigned to the Renew political group. The Rapporteur presented his draft report on 30 October 2025, which attracted over 2,200 amendments from political groups. As intra-LIBE negotiations on these amendments proved challenging, shortly before the planned vote, the largest political group, the European People’s Party (EPP), tabled alternative compromise amendments with the support of right-wing and far-right groups (European Conservatives and Reformists, Patriots for Europe and Europe of Sovereign Nations). Holding a majority within LIBE, these groups secured the report's adoption on 9 March 2026. Despite mobilisation by civil society and observations issued by the Council of Europe Commissioner for Human Rights, the Parliament adopted the report in plenary on 26 March 2026, with 389 votes in favour, 209 against, and 32 abstentions.

This blog post provides a brief discussion of the European Parliament’s position on selected key measures proposed by the Commission and may be read alongside the previous analysis.

 

Detection measures: As a rare instance in which the Parliament prevents the application of coercive measures, it removed the detection measures proposed by the Commission in Art. 6. These measures, aimed at identifying persons staying in an irregular situation for the purpose of their removal, had been criticised as likely to lead to racial profiling.

Return decision and safeguards (Art. 7, 12 and 26–28): The Parliament expanded the possibility, already envisaged in the Commission’s draft, of not indicating the country of return in the return decision. It provides that, where the country of return cannot be determined at the time of issuing the return decision, the decision may provisionally designate none, one or several countries of return. It further clarifies that the country of return may be determined either in the return decision itself or subsequently, in the removal order. Moreover, where a country of removal has not previously been specified in the return decision, or where it differs from the country designated therein, the Member State is required only to notify the person in accordance with national law. In such cases, the issuance of a new return decision or removal order is not required. This approach may prevent individuals from effectively appealing removal to a specific country, as notifications falling short of a formal decision or act may not be subject to review. It therefore risks undermining the principle of non-refoulement, the right to an effective remedy, and the principle of good administration.

The absence of an indication of the country of return in the return decision is highly problematic from a non-refoulement perspective, as it suggests that an individual assessment of the risk of refoulement may not be carried out at that stage of the procedure. Indeed, the Parliament’s report requires Member States to determine the country of return only “prior to carrying out the removal.” In addition, it weakens the risk assessment envisaged in the Commission’s proposal by transforming what appears to be an obligatory assessment into one that is triggered only when the person raises the risk or when the authorities otherwise become aware of it.

These concerns are compounded by a further weakening of remedies. The Parliament’s report explicitly states that an appeal does not have an automatic suspensive effect, while merely preserving the possibility for national law to provide for such an effect.

Mutual recognition of return decisions (Art. 9): The Parliament largely retains the Commission’s provisions on mutual recognition. It adds, however, that not only return decisions but also removal orders may be enforced by another Member State. This raises practical challenges, as only the main elements of the return decision are to be included in the standardised form (European Return Order) made available through the Schengen Information System. Crucially, given the expanded possibility of not indicating the country of return in the return decision, the enforcing Member State may in practice need to carry out a new procedure to issue a return decision. The Parliament further clarifies that Member States are not required to adopt administrative decisions or acts for the purposes of recognition, which may further limit the individual’s ability to challenge enforcement measures.

Voluntary departure (Art. 13): Compared to the Commission’s proposal, the Parliament further deprioritises “voluntary” departure. While the Commission’s draft implicitly allows for a period of zero days for voluntary departure, the Parliament’s report explicitly provides for the possibility of imposing immediate “voluntary” departure. It also makes the granting of a longer period contingent upon a request by the person concerned and removes the non-exhaustive list of circumstances justifying such an extension, including family links and children’s schooling. The Parliament appears to overlook that voluntary departure is not only more favourable to the individual, and consistent with the principle of proportionality, but also serves the interests of the State.

Safeguards pending removal (Art. 14): The Parliament removes the basic rights to which persons are, in principle, entitled during the period of postponement. This list includes the fulfilment of basic needs, respect for family unity, emergency health care and essential treatment of illness, access to the basic education system, and the protection of the special needs of vulnerable persons. These safeguards are laid down in the current Directive (Art. 14(1)), and the Commission’s proposal slightly expanded this list. They also reflect human rights obligations, including the rights to education, health, and basic subsistence.

Monitoring of removal (Art. 15): The detailed provisions on return monitoring included in the Commission’s proposal have been significantly curtailed. Notably, the Parliament removed key elements relating to the powers and capacities of the independent monitoring mechanism, including the requirement to ensure appropriate means, autonomy in action, and a mandate to communicate substantiated allegations of failures to respect fundamental rights to the competent national authorities.

Detention (Art. 29–35): The Parliament’s report retains the Commission proposal’s deletion of the current requirement that detention may be imposed only where no sufficient but less coercive measures can be applied in a specific case (Art. 15(1)). It also expands the grounds for detention to include non-compliance with cooperation duties. Given this extensive list of grounds, detention risks becoming a regular measure, in violation of the principles of necessity and proportionality inherent in the right to liberty. The maximum detention period of two years (12 months + 12 months) in a given Member State is maintained, representing a significant increase from the current 18-month limit (Art. 15(5)-(6)). In contrast to the Commission’s draft, the review of detention would no longer need to be concluded within 15 days. On a positive note, the Parliament introduced, in the preamble (§32), a provision stating that children, as a rule, should not be detained.

As regards detention conditions, the Parliament retains the limitations proposed by the Commission and, in some respects, further amplifies them. Compared to the current directive (Art. 16), the proposed rules significantly lower standards governing detention conditions and treatment. Detention may take place not only in specialised facilities but also in designated sections of other facilities, and where prisons are used, migrants are to be kept separated from ordinary prisoners only where possible. Temporary measures may be adopted to expand capacity in exceptional return situations. Access to open-air spaces, introduced by the Commission, may be temporarily restricted where necessary and proportionate to ensure the proper functioning of detention facilities. Contact with the outside world will be subject to security and administrative considerations. Families without children will not be entitled to separate accommodation ensuring privacy.

Entry ban (Art. 10–11): The Parliament further broadens the already expanded scope of entry bans proposed by the Commission. It removes the maximum duration of 10 years for entry bans (compared to the current maximum of 5 years in cases not involving security threats under Art. 11(2)) and also eliminates any maximum period for persons considered to pose security risks. In addition, the Parliament renders less explicit the circumstances in which an entry ban may be withdrawn, suspended, or shortened.

It also removes the safeguards proposed by the Commission for entry bans imposed when a person’s irregular stay is detected at exit, including the requirement for justification based on the specific circumstances of the individual case, respect for the principle of proportionality, and the rights of defence. Furthermore, the Parliament’s report introduces additional situations in which entry bans may be imposed in the context of departure, such as where a person leaves the territory without a return decision having been issued, or departs before such a decision is adopted. Entry bans imposed on persons seeking to leave the territory without completing a return procedure are counterproductive, disproportionate, and disregard the requirement of an individual assessment.

Return hubs (Art. 4(3) and 17): The Parliament’s report provides that the agreement or arrangement forming the basis for return hubs may be concluded not only by a Member State but also by the Union. It also weakens notification obligations towards the Commission and removes families with children from the category of persons exempted from removal to return hubs, leaving only unaccompanied children exempted. Crucially, the list of countries of return, which includes countries hosting return hubs at its end, clarifies that the order of the list does not determine the sequence in which those countries may be applied. This suggests that a person may be removed to a return hub even before any attempt is made to return them to their country of origin, raising concerns as to compliance with the principle of good faith and increasing the risk of human rights violations associated with transfers to third countries.

Alternatives to return (Art. 7(9)): The Parliament retains the approach in the Commission’s proposal whereby the current possibility under Art. 6(4) of not issuing a return decision is removed from the list of situations in which such a decision is not adopted. Under the Parliament’s proposal, decisions by Member States to grant an autonomous residence permit, long-stay visa, or other authorisation granting a right to stay on compassionate, humanitarian, or other grounds may only result in the withdrawal or suspension of a return decision, rather than its non-issuance.

References to national law: Finally, the Parliament defers to national law in defining a range of measures, which sits uneasily with the nature of a regulation, as opposed to a directive, and may give rise to disproportionate and arbitrary outcomes. The measures whose applicability or scope may be determined at the national level include removal (Art. 12(1)), cooperation duties (Art. 21(2)), penalties for non-compliance with such duties (Art. 22(1)), detention (Art. 29(3)), and restrictive measures (so-called alternatives to detention) (Art. 31(2)). The notion of “security risks”, which triggers the application of multiple coercive measures, also encompasses threats defined under national law.

 

Concluding thoughts

The European Parliament’s report on the proposed Return Regulation reflects a markedly coercive approach towards persons in an irregular situation, raising serious concerns as to its compatibility with fundamental rights. These include the prohibition of refoulement, arbitrary detention, torture and ill-treatment, the right to due process and procedural safeguards, the rights of the child, as well as socio-economic rights, alongside broader principles of proportionality and human dignity. This approach stands in stark contrast to the Parliament’s December 2020 resolution, which reflected safeguards developed by UN human rights mechanisms and may be regarded as a counterbalance to the Commission’s 2017 recommendation on the implementation of the Return Directive. In that resolution, the Parliament emphasised that voluntary return should be prioritised over forced return; unaccompanied children should only be returned where this is demonstrably in their best interests; children should never be detained for immigration purposes; and detention must remain a measure of last resort. It also expressed concern over the widespread automatic imposition of entry bans and broad criteria for assessing the risk of absconding, and encouraged greater use of autonomous residence permits.

The current European Parliament does not appear to provide an effective counterweight to executive proposals that raise concerns from both a human rights and effectiveness perspective. While the Commission’s proposal for the Return Regulation had already been widely criticised as disproportionate and problematic from a fundamental rights standpoint, the Parliament largely retains its most contentious elements and, in many respects, further reinforces their coercive nature. In practice, the Parliament’s position is closely aligned with that of the Council, itself regarded as particularly restrictive. As a result, trilogue negotiations are expected to proceed swiftly, mirroring the accelerated adoption of recent instruments such as the regulations on safe third countries and safe countries of origin.

This marks a striking departure from the legislative dynamics during the 2005–2008 legislative process leading to the adoption of the Return Directive (discussed here, here and here). At that time, acting under the then newly established co-decision procedure, which placed it on equal footing with the Council, the European Parliament was widely regarded as a key defender of fundamental rights. It significantly strengthened safeguards in the draft Directive proposed by the Commission and, during the subsequent negotiations with the Council, played a decisive role in preventing the adoption of some of the Council’s more restrictive proposals. The institutional dynamics were famously characterised as a triad of the “Good” (Parliament), the “Bad” (Council), and the “Ugly” (Commission). Nearly two decades later, however, this configuration appears to have shifted, with all three actors now converging towards the “Bad.”

 

Detention by Delay? A Structural Paradox in EU Migration Law

 

 

Miguel Alconero Bravo, Predoctoral Research Fellow (Formación de Personal Investigador – FPI) at the University of Valladolid*

*Part of the Research Project “Proceso Penal y Espacio de Libertad, Seguridad y Justicia: Garantías, Cooperación Transfronteriza y Digitalización” (Ref. PID2023 – 152074NB – I00). 

Photo credit: berthgmn via wikimedia commons

 

Introduction

 

Predicating the detention of a migrant on circumstances entirely beyond his or her control seems prima facie unjust. Indeed, even a cursory reading of the provisions of EU law empowering Member States to extend the detention of a foreign national due to delays in obtaining the necessary documentation from third countries unmistakably exposes the inherent disproportionality of such a measure.

 

Ordering detention v. Extending detention

 

It must first be emphasised that both Article 15(6) of the Return Directive and Article 32(3) of the Commission proposal to overhaul EU return procedures enable Member States solely to extend a period of deprivation of liberty that has already been lawfully ordered.

 

Accordingly, where the conditions required under EU law for the detention of a third-country national are not satisfied in a given factual scenario, it is legally immaterial whether any of the grounds permitting a State to prolong that measure are present. If a person could not lawfully have been deprived of liberty from the outset, there is, fundamentally, no valid detention susceptible of being extended.

 

In this regard, Article 15(5) of the Return Directive states that “detention shall be maintained for as long a period as the conditions laid down in paragraph 1 are fulfilled and it is necessary to ensure successful removal”, whereas Article 32(1) of the proposal for a Return Regulation provides that “detention shall be maintained for as short a period as possible and for as long as the conditions laid down in Article 29 are fulfilled […]”.

 

Reference must therefore be made to Article 15(1) of the Return Directive, which adopts a numerus apertus approach. This provision refers in particular to two situations that may lead to the detention of a foreign national pending removal: the existence of a risk of absconding, and the possibility of identifying conduct through which the migrant seeks to avoid or hamper the preparation of return or the removal process.

 

Similarly, Article 29(3) of the proposed Regulation sets out, in exhaustive terms, the only situations in which the requesting State may order the detention of an irregularly staying third-country national: (a) where there is a risk of absconding determined in accordance with Article 30; (b) where the third-country national avoids or hampers the preparation of the return or the removal process; (c) where the third-country national poses security risks in accordance with Article 16; (d) in order to determine or verify his or her identity or nationality; (e) in the event of non-compliance with the measures ordered pursuant to Article 31.

 

Thus, the vast majority of the grounds for detention laid down in the Return Directive and in the proposed Regulation relate to scenarios in which the third-country national concerned has sought to frustrate the enforcement of return or the removal process. In essence, deprivation of liberty under both instruments is generally linked to some form of non-cooperation by the person concerned.

 

In this context, it must be stressed that construing the foreign national's lack of cooperation as a ground for extending detention – as Article 15(6)(a) of the Return Directive and Article 32(3) of the proposed Regulation do – makes it easier to postulate that the ground originally justifying detention still subsists. In other words, such a construction allows the authorities to establish the presence of almost any of the grounds set out in Article 15(1) of the return Directive and in Article 29(3) of the proposed Regulation for the purpose of ordering the detention of a migrant.

 

As a matter of fact, when a State invokes the “lack of cooperation by the third-country national concerned” in order to prolong detention, it is essentially relying on a concept whose indeterminacy allows national authorities to subsume within it many of the very grounds on which they could initially have relied to deprive that person of liberty.

 

In sum, extending the detention of a foreign national as a consequence of his or her lack of cooperation will, in most cases, make it possible to ascertain that the grounds for detention set out in EU law continue to be met.

 

Nonetheless, an exception must be made in respect of Articles 29(3)(c) and 29(3)(d) of the proposed Regulation –namely, where the person concerned is a third-country national deemed to pose a security risk, and where detention is ordered for the purpose of determining or verifying the foreign national’s identity or nationality.

 

In that respect, it should be noted that detention ordered on the grounds that a migrant poses a risk to public order, public security or national security bears, strictly speaking, no relation to the degree of cooperation he or she may have been willing to display during the return procedure. Irrespective of the individual’s willingness to cooperate, once a foreign national has been deprived of liberty on the basis that he or she is regarded as a threat to a State’s public order, public security or national security, detention will persist for as long as that classification remains in force.

 

That being so, authorities cannot invoke a migrant’s lack of cooperation to extend detention where the initial deprivation of liberty was based on security grounds or on the need to verify identity, as those circumstances do not, as such, relate to the individual’s conduct during the return procedure (unless a genuine and subsequent lack of cooperation actually materialises). Nevertheless, it is equally true that, in our view, Brussels errs in treating said situations as grounds for detention in the first place.

 

Against this backdrop, the pivotal question is as follows: under which of the grounds set out in either Article 15(1) of the Return Directive or Article 29(3) of the proposed Regulation could a Member State’s intention to prolong the detention of a third-country national be classified when that intention is based, purely and simply, on the likelihood that the procedure will take longer owing to delays in obtaining the required documentation from third countries?

 

As far as the Return Directive is concerned, the answer may lie in the numerus apertus structure of Article 15(1). Article 29(3) of the proposed Regulation, however, does not, under any circumstances, permit a third-country national to be deprived of liberty solely on the basis of delays in obtaining documents from third countries.

 

The conclusion follows inexorably. The wording of Article 32(3) of the proposed Regulation allowing deprivation of liberty to be prolonged “where the return procedure is likely to last longer owing to […] delays in obtaining the necessary documentation from third countries” ought to be removed. If EU law does not recognise the possibility of detaining a foreign national on that ground, it is a fortiori manifestly contrary to the principle of proportionality that such delays should then be relied upon to justify prolonging the duration of a custodial measure.

 

A structural contradiction

 

It must be acknowledged that both Article 32(3) of the proposed Return Regulation and Article 15(6)(b) of the Return Directive pursue an entirely legitimate aim: namely, to prevent, through the use of all available means, a situation in which enforcement of the return decision becomes impossible as a result of a temporary delay in receiving the necessary documentation from the requested State.

 

That said, it is crucial to recall that, in order for the detention of a third-country national to be prolonged, there must first exist a custodial measure resting on a valid legal basis. The analysis must therefore begin by determining whether in the factual circumstances of the case the substantive conditions laid down in EU law for depriving a migrant of liberty are actually satisfied.

 

Against this background, neither Article 15(1) of the Return Directive nor Article 29(3) of the proposed Regulation treats a mere delay in the completion of the administrative or diplomatic formalities necessary to obtain the required documentation from third countries as a circumstance capable of justifying an initial deprivation of liberty.

 

This gives rise to a stark legal contradiction: if EU law precludes the initial detention of a third-country national on the basis of administrative delays in obtaining the necessary documentation from third countries, why does Brussels permit its Member States to extend detention on that very ground? Although such delays cannot justify the initial deprivation of liberty, they may nonetheless serve to prolong the detention of a migrant.

 

The result is a structural paradox whereby a custodial measure that should come to an end from the moment the substantive conditions underpinning detention are no longer met is nevertheless lawfully prolonged by the requesting State under the provisions governing extension.

 

That tension becomes all the more striking when viewed in the light of other instruments of EU migration law. In particular, Article 11(1) of Directive 2024/1346 provides that “delays in administrative procedures that cannot be attributed to the applicant shall not justify a continuation of detention”.

 

For all these reasons, if the EU wished to preserve this ground for prolonging detention in the proposed Regulation, its express incorporation into Article 29(3)  as a ground, not just for extending, but also for ordering detention would be necessary.

 

As regards Directive 2008/115, the only interpretative solution would be to conclude that, by merely identifying the risk of absconding and the possibility that the individual may avoid or hamper the preparation of return as the two grounds that, in particular, entitle a State to deprive a migrant of liberty, that instrument implicitly leaves room for the authorities to rely on other situations as independent grounds for detention. One such additional scenario would be, precisely, delays in obtaining the necessary documentation from third countries.

 

However, both of those alternatives must be categorically rejected. Otherwise, the ultima ratio nature traditionally associated with the detention of an irregularly staying third-country national would be severely compromised.

 

In any event, the release of the individual concerned does not prevent the authorities from resorting to less restrictive measures designed to secure the enforcement of return pending receipt of the necessary documentation.

 

In this regard, Article 31 of the proposed Regulation explicitly provides for various alternatives to detention, which must be “proportionate to the level of the risk of absconding assessed in accordance with Article 30”. Furthermore, Article 23 of said proposal provides for a series of options which, by restricting the individual’s geographical mobility, are capable of ensuring his or her availability during the removal process.

 

Mahdi: When can detention be extended?

 

The Court of Justice of the European Union (CJEU) clarified in Mahdi the scope of the assessment that national courts must undertake before authorising any extension of a detention period on the basis of Article 15(6) of the Return Directive

 

In order to rule on the questions referred by the Administrativen sad Sofia-grad, the CJEU began by identifying the defining features of Article 15 of the Return Directive. It emphasised, in particular, that this provision is both “unconditional” and “sufficiently precise”, qualities which explain why its implementation by the Member States requires “no other specific elements”.

 

The Court then turned to the provisions whose interpretation had been requested by the referring court, namely Articles 15(3) and 15(6) of the Return Directive.

 

Concerning Article 15(3), the Court underlined the clarity with which its wording indicates that any detention measure prolonged over time must be subject to the “supervision” of a judicial authority. Yet the CJEU also acknowledged that the provision does not specify the precise nature of that examination, thereby making it necessary to “recall the rules deriving from Article 15 […]” in order to complete the analysis required by the question raised by the Bulgarian court.

 

It was at that stage that the Court brought Article 15(6) of Directive 2008/115 into the discussion.

 

Significantly, however, the CJEU did not confine itself to identifying the consequences flowing from the substantive conditions set out in that provision. Instead, it examined Article 15(6) in the light of Article 15(4) of the same Directive.

 

In so doing, the Court of Justice of the European Union introduced two decisive elements into the judicial review of the deprivation of liberty of an irregularly staying third-country national.

 

First, it held that, “at the time of the national court’s review of the lawfulness of

detention”, there must be “a real prospect that the removal can be carried out successfully”, having regard to the time limits laid down in Articles 15(5) and 15(6) of Directive 2008/115.

 

Secondly, the Court established that the authority determining whether the individual’s detention should be extended or whether he or she should be released must re-examine the substantive conditions laid down in Article 15(1) of the Return Directive. Crucially, this entails verifying that the grounds “which have formed the basis for the initial decision to detain the third-country national concerned” continue to subsist.

 

This point is of central importance to the present discussion, since Article 15(1) of the Return Directive makes no reference whatsoever to the possibility of detaining a foreign national solely on account of delays by third countries in issuing the necessary documentation. It follows that reliance on such a ground does not relieve the competent authorities of the burden of demonstrating, first, that there remains a “reasonable prospect of removal” and, secondly, that the substantive grounds which justified the initial deprivation of liberty continue to subsist.

 

The disappearance of either the reasonable prospect of removal or the substantive grounds that initially justified the detention of the third-country national necessarily renders that deprivation of liberty unlawful, as Article 15(4) of the Return Directive makes clear. That conclusion is not altered by the mere fact that the authorities continue to await the necessary documentation from other States.