The Fiction of Non-Entry Meets the Fiction of Remaining: AG Emiliou in Sedrata

 

 

Andreina De Leo*, Post-doctoral researcher, Maastricht 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: Fred Romero, via Wikimedia commons

 

Introduction

 

On 23 April 2026, Advocate General Nicholas Emiliou delivered his Opinion in Sedrata, one of two pending preliminary references concerning the Italy–Albania Protocol. The Opinion marks a significant development in the ongoing proceedings before the Court of Justice of the European Union, addressing for the first time the question of whether asylum and return procedures may be conducted in a third country while remaining fully under the jurisdiction of a Member State.

 

While affirming that EU law continues to apply wherever a Member State exercises jurisdiction, and that the relocation of procedures outside EU territory does not, in itself, remove them from the scope of the acquis, the Opinion does not exclude, in principle, the possibility of delocalising asylum and return procedures to a third country, provided that all guarantees under EU law are fully respected.

 

This post examines the key elements of the Opinion and highlights the tensions that arise from its underlying logic.

 

EU Law Beyond Territory: Applicability and Conditional Compatibility in the AG Opinion

 

A central feature of the Opinion is its unequivocal rejection of a strictly territorial understanding of EU migration law. The AG finds that both the Return Directive and the Asylum Procedures Directive apply directly to the situations at issue, not merely by virtue of their incorporation into national law, but as a matter of EU law itself. In his view, the transfer of individuals to centers located in a third country cannot have the effect of rendering those directives inapplicable or of prejudicing the application of common standards, insofar as the procedures clearly fall within the material scope of EU law (see, in particular, paras. 34 and 48).

 

More broadly, the Opinion endorses a jurisdiction-based understanding of the scope of EU law, echoing principles familiar from international law. Where a Member State exercises authority over individuals, it remains bound to ensure compliance with all applicable EU rules, irrespective of the geographical location in which those activities are carried out. In this respect, the Opinion directly contradicts the position advanced by the European Commission during the hearing, which emphasized the territorial anchoring of the asylum acquis and denied its direct applicability outside the Union. Instead, the Advocate General affirms that EU law does not “switch off” when procedures are relocated beyond EU borders (para. 47).

 

Having established that EU law applies, the Opinion adopts a relatively concise approach to compatibility. As regards the Return Directive, the AG considers that nothing in its provisions expressly precludes the operation of detention facilities outside the territory of a Member State. In his view, the Directive regulates the grounds, duration, and conditions of detention, but remains silent on the geographical location of such facilities. From this silence, he infers that EU law does not, in principle, prohibit transfers to centers situated in a third country. The Opinion recognizes that such a scenario was likely not envisaged by the EU legislature. On this basis, and in the absence of explicit regulation, Member States retain a margin of discretion as to the organization of detention, including its possible extraterritorial dimension (paras 58-59, and 64).

 

At the same time, the Opinion draws a crucial conceptual distinction: the transfer to Albania cannot be qualified as a “return” or a “removal” within the meaning of the Directive. It neither brings the return procedure to an end nor constitutes the enforcement of a return decision. This clarification is significant, as it confirms that the Italy–Albania model is conceptually and legally distinct from the idea of “return hubs” currently discussed in the context of the proposed EU Return Regulation. In that framework, the transfer of a third-country national to a third State, other than the country of origin or habitual residence, would amount to the execution of a return decision directed towards that State, a possibility which is not permitted under the existing EU law. By contrast, the transfer to Albania is characterized as an intermediate step within an ongoing procedure: individuals remain fully under the authority and jurisdiction of the Member State, and the return decision is neither executed nor exhausted (paras 60-62).

 

A similar reasoning is adopted in relation to the Asylum Procedures Directive and the Reception ConditionsDirective. The AG considers that Article 9(1) of the Procedures Directive – which enshrines the right to remain on the territory of Member States pending the first instance asylum decision – does not, in principle, preclude the transfer of asylum seekers to a third country, provided that applicants are able to exercise their procedural rights effectively, including appearing before the competent authorities and benefiting from the guarantees laid down in EU law. While acknowledging that a literal reading might suggest that applicants must be brought back to the territory of the Member State once they lodge an application, the AG rejects such an interpretation when the provision is read in context and in light of its purpose. In his view, the notion of “remaining in the Member State” must be understood by reference to the definition in Article 2(p) of the Directive, which already includes border areas and transit zones, thus pointing towards a functional rather than strictly territorial understanding (paras 93-97).

 

On this basis, and supported by a functional and systemic reading, he argues that Member States cannot be regarded as precluded, as a matter of principle, from organizing asylum procedures in facilities located outside their territory, provided that those facilities fall under their jurisdiction and that EU guarantees are fully ensured in practice. The core objective of Article 9(1) is, in his view, to prevent removal before a final decision on the application has been taken and to safeguard the effectiveness of the right to seek asylum, rather than to guarantee physical presence on the territory of the Member State in all circumstances. In situations where those guarantees are maintained and no risk of refoulement arises, the relocation of procedures does not, in his view, undermine the Directive’s purpose. Accordingly, Emiliou concludes that Article 9(1) does not, in principle, preclude national legislation such as that at issue in the main proceedings (paras. 102–103).

 

The Functional Turn: “Magically turning non-EU territory into EU territory via some mystical legal alchemy” (Peers, here)

 

As argued elsewhere (see here, with Celoria, and here with Celoria and Ferri), I concur that EU law does not apply as a matter of voluntary choice, but follows from the fact that the procedures fall within its material scope, thereby ensuring the need to preserve the primacy and uniform interpretation of EU law, and to prevent unilateral circumvention of the acquis. Likewise, the transfer to Albania does not amount to a “return” or a “removal” within the meaning of the Return Directive, nor can it be assimilated to the concept of “return hubs”, as it does not constitute the enforcement of a return decision. It is also plausible that the silence of the EU legislature reflects the fact that such configurations were not contemplated at the time of adoption. However, beyond these points, the reasoning of the AG appears unconvincing, and is affected by internal inconsistencies.

 

The Opinion relies on a broad and functional interpretation of spatial concepts within EU law. The AG adopts an expansive reading of notions such as “territory,” “border,” and “transit zones,” suggesting that their meaning should be determined in light of function rather than physical location. This approach makes it possible to treat facilities located outside the Union as functionally equivalent to border or transit zones situated within Member State territory.

 

This is not entirely new: for instance, in the recent Danané (discussed, in this blog here), the Court, on the basis of an Opinion by the same AG, accepted that procedures traditionally associated with entry control may operate through a legal fiction of non-entry and be carried out in facilities located inland where national law designates them as “border” or “transit” settings. In that sense, the underlying logic of functional interpretation of geographical concepts is not unprecedented.

 

However, those cases remain anchored, in practice, within the geographical space of the Union, understood in a legally fragmented but still territorially contained manner. What is distinctive in the present context is the further conceptual step that is being contemplated: the extension of this legal fiction beyond the geographical borders of the Union itself. This marks a qualitative shift, as the “border” is no longer merely redefined within the territory of the Member State, but effectively displaced outside it.

 

This development raises significant concerns. By effectively allowing Member States to shape the spatial reach of EU procedures, it risks extending derogatory regimes beyond their traditionally circumscribed scope and weakening the safeguards attached to them. Moreover, the Opinion itself appears to expose tensions regarding the internal coherence and normative limits of this reasoning.

As mentioned, the interpretative ambiguity stemming from the absence of an explicit prohibition on relocating EU asylum and return procedures outside the Union is addressed through an expansive reading of “territory” and of “border” or “transit zones,” extending these concepts to extraterritorial settings. In other words, in the absence of a clear treaty or legislative definition clarifying that “territory” refers to the geographical territory of the Union rather than what a Member State unilaterally designates as such, the question arises as to whether Member States retain any discretion to determine the spatial scope of EU asylum law in this field.

 

EU institutions, including the Commission during the hearing, have consistently rejected such an approach, clarifying that asylum claims cannot be lodged or processed outside the Union, as this would require an extraterritorial application of EU law considered neither legally feasible nor politically acceptable. At most, Member States may transfer individuals to third countries in compliance with the principle of non-refoulement, but without retaining responsibility for the examination of asylum applications after disembarkation. The European Parliament’s Legal Service has adopted a similar position, recognizing that jurisdiction may, in certain circumstances, extend extraterritorially, while stressing that this does not transform such locations into EU territory for the purposes of asylum law. The institutional position thus converges on a key distinction: while jurisdiction may extend beyond borders in limited situations, the notion of “territory” under EU asylum law remains autonomous and confined to the geographical territory of the Union (for more info, see De Leo & Celoria, p. 604).

 

The AG introduces, however, a preliminary contextual remark noting that the agreement underlying the Italian legislation concerns Albania, a geographically close State to Italy, separated only by the Adriatic Sea, which is also an ECHR Contracting Party and an EU candidate country engaged in accession negotiations. These elements are treated as relevant factors in assessing the lawfulness of the scheme under EU law (paras 68-69). This is where a significant internal incoherence emerges.

 

If “territory” under EU asylum and migration law is not an autonomous legal concept, meaning that Member States are free to define its spatial reach at their discretion, then the assessment of compatibility with EU law in abstracto cannot, without contradiction, be made to depend on contextual variables such as geographical proximity or an ex ante assumption of compliance with fundamental rights. The attempt to anchor the analysis in Albania’s proximity and its presumed rights compliance therefore sits uneasily with the jurisdictional reading endorsed elsewhere in the Opinion. It effectively introduces an ad hoc relational criterion into what is otherwise presented as a non-autonomous legal concept under EU law, thereby blurring the distinction between a principled determination of the spatial scope of EU migration and asylum law and a pragmatic, case-by-case assessment of legality. Either the asylum and migration acquis is territorially bound to the Union’s geographical space, including only its legally constructed exceptions such as borders and transit zones, or it is not territorially constrained at all. In the latter scenario, the question would no longer concern abstract spatial limits but the concrete application of EU guarantees in individual cases, leaving national courts to assess compliance with EU law obligations without any role for EU law in further delimiting the spatial reach of “territory” itself.

 

The result is a conceptual inconsistency: what is framed as a question of legal scope, i.e. whether “territory” under EU law is autonomous or subject to Member State discretion, is made dependent on operative factors based on contextual and geopolitical considerations. These considerations, however, have no clear legal basis in determining the applicability of the acquis and end up suggesting, implicitly, a degree of autonomy in the concept of territory that the initial premise simultaneously denies.

 

This tension is further reinforced by the Advocate General’s acknowledgment that, while compatibility in abstracto may be preserved, the extraterritorial location of detention facilities inevitably generates practical and structural difficulties in ensuring full compliance with EU guarantees, thereby highlighting how relocation outside the Union introduces an inherent and additional layer of complexity in the effective protection of rights that would not arise within the territorial space of the Union.

 

More concretely, these difficulties are linked to structural factors such as geographical distance, the need to cross an international border to access the centers, and dependence on both third-country authorities and those operating within the centers for the enjoyment of key guarantees, including family visits and effective communication with legal counsel. The AG emphasizes that this does not, as such, imply a violation of EU law, but rather that Member States may be required to adopt specific organizational and logistical measures to ensure compliance with EU standards (para 80).

 

The limits of this reasoning become particularly visible in the AG’s own discussion of the requirement of “immediate” release under the Return Directive. In assessing the meaning of this guarantee, the Opinion accepts that the term “immediately” cannot be interpreted in an overly rigid or literal manner, but nevertheless acknowledges that national authorities must be able to organize, within a short timeframe, the practical modalities necessary to give effect to a judicial order ordering release, including transport by air or sea where required (para. 82). Even accepting this flexible interpretation, however, the Albanian model reveals the structural impossibility of equivalence introduced by extraterritorialisation. In a purely domestic setting, a finding that detention is unlawful triggers immediate release within the legal and physical space of the Member State, without any further material impediment to the restoration of liberty. By contrast, in the Albanian setting, even where a judicial authority orders release, the individual remains subject to continued physical constraint until the completion of a cross-border transfer back to Italy, thereby introducing an additional and legally significant phase of deprivation of liberty which is absent in a territorial context and stems solely from the extraterritorial location of the detention facility. This example illustrates, more broadly, that the issue is not confined to a single instance of unequal treatment. Rather, it reveals a pattern of structural frictions affecting a range of guarantees, i.e. access to legal assistance, the possibility of family visits, and the effectiveness of judicial protection, which, taken together, seriously undermine the claim of functional equivalence.

 

Seen in this light, the reasoning concerning the conditions for compatibility appears particularly unconvincing. The difficulty is not simply that individual guarantees may, in specific cases, be ineffectively implemented. It is that the extraterritorial design itself generates structural constraints that systematically hinder their effective exercise. These are not incidental or remediable deficiencies, but structural features of a system that operates outside the territorial and institutional framework in which the EU acquis was conceived. The issue, therefore, is not whether compliance can be demonstrated case by case, but whether a model that introduces additional layers of dependency, delay, and fragmentation can, as such, meet the standard of effectiveness and uniform applicability required by EU law, irrespective of any ad hoc organizational measures adopted by Member States. As such, this interpretation departs not only from a literal reading of the relevant provisions, but also from their teleological interpretation, which is precisely to ensure the effective and practical enjoyment of the rights they enshrine.

 

Furthermore, accepting the relocation of procedures outside the Union leaves unresolved a set of central issues concerning the Dublin system and mutual trust (a dimension addressed in detail in our previous analysis here), which the Opinion largely sidelines, merely noting that its applicability follows from that of the Asylum Procedures Directive and that its provisions must be interpreted coherently with it (para. 49). These questions therefore remain insufficiently developed in the reasoning, plausibly because they were not directly raised by the referring court, but are likely to be clarified in pending related proceedings (Joined cases C-706/25 Comeri and C-707/25 Sidilli), which more broadly concern whether the Protocol may adversely affect the proper functioning of internal EU rules, in potential breach of the Union’s exclusive competence in the light of the ERTA doctrine.

 

Conclusion

 

Overall, the Opinion of Advocate General Emiliou in Sedrata adopts an ambivalent approach to the legality of the de facto extraterritorial application of EU asylum and migration law.

 

On the one hand, it clearly affirms that EU law continues to apply wherever Member States exercise jurisdiction, thereby ruling out any attempt to circumvent obligations through geographical relocation. At the same time, it preserves the compatibility of the Protocol with EU law through a functional redefinition of the notion of “territory”. On the other hand, the delocalization of asylum and return procedures to third countries is accepted only under strict conditions: full compliance with EU standards, coupled with an uncertain reliance on contextual factors such as geographical proximity and an ex ante assumption of adequate fundamental rights protection based on formal international commitments. Yet the AG himself acknowledges that ensuring such compliance is significantly more demanding in an extraterritorial setting, while also leaving open a degree of Member State discretion in shaping the spatial understanding of “territory” and “border”.

Taken together, this gives rise to a structural tension and an internal incoherence that ultimately weakens the persuasiveness of the reasoning. From this perspective, coherence and effectiveness can in my view only be preserved by recognising that “territory”, for the purposes of EU asylum and return law, is an autonomous concept confined to the geographical territory of the Union, including its legally constructed exceptions such as border areas and transit zones. This implies, in practical terms, that both the examination of asylum applications under ordinary and accelerated border procedures and the detention phase under the Return Directive, prior to the enforcement of removal to the country of origin or to a third different country (should the proposed Return Regulation be adopted), must take place within the Union’s geographical space, irrespective of any unilateral qualification by Member States.

This view is grounded in the structural logic of the asylum and return acquis prior to removal, whose safeguards are designed to operate within the Union’s territorial framework. It is within this framework that supervision, enforceability, and effective protection are meant to be ensured. Accordingly, such guarantees cannot be transposed outside the Union through a functional extension of jurisdiction without altering their operation and effectiveness, even where formal complianc appears to be ensured.

Ultimately, it is now for the Court of Justice to decide whether this logic is to prevail, or should it be rewritten.

One Legal Fiction After Another: The Court of Justice judgment on the asylum border procedure in Joined Cases C‑50/24 to C‑56/24 (Danané)

 

 

Dr. Vasiliki Apatzidou, Associate Tutor at the University of London

 

Photo credit: Nicolas Economou, via Wikimedia Commons

 

Introduction

 

The Court’s judgment in the Danané case largely confirms the core argument made by Advocate General Emiliou in his Opinion: that the border procedure in EU asylum law (which provides for a fast-track assessment of asylum claims, with the implied legal fiction that the applicant has not entered the territory) is not exclusively attached to the physical border. In my earlier analysis of the Opinion, I argued that the Advocate General had moved towards a functional rather than pragmatic understanding of the ‘border’. According to him, procedures traditionally associated with entry control may not only take place in border facilities, but also in inland facilities if they are designated as such by national law. The Court of Justice has now broadly endorsed that logic by holding that EU law does not prohibit Member States from detaining asylum applicants during asylum border procedures in centres that are not geographically located at the border, and that such detention may continue in the same place after the expiry of the four-week period envisaged for the border procedure provided that a new legal basis for detention exists. In that sense, the original thesis still stands. However, the judgment also tries to contain the consequences of this argument through a focus on safeguards.

Summary of the Judgment

 

The joined cases arose from applications for international protection lodged at the Belgian border by third-country nationals who were refused entry and detained in the Caricole Transit Centre, a facility located inside Belgian territory but treated under national law as a place ‘at the border’ for the purposes of the border procedure. In Danané, the Court confirms that Article 43 of the recast asylum procedures Directive (APD), which sets out the current version of the border procedure, does not require border procedures to be implemented only in facilities physically located at the external border, since a border procedure may also be conducted in an inland detention centre that national law treats as a place ‘at the border’.

At the same time, once the four-week period in Article 43(2) recast APD expires, the case no longer remains within the border procedure and must continue under the regular procedure, even though the applicant may still be kept in the same facility if continued detention is separately justified under the detention grounds laid down in Directive 2013/33 (the reception conditions Directive, or RCD). The Court therefore accepts a dual classification of the same place of detention. However, this may happen only on condition that the applicant is informed of the change in legal status – it should be clear that he or she is now allowed to enter the territory – and that continued detention complies with the requirements of necessity, proportionality, individual assessment, and judicial review. It also makes clear that investigative steps already carried out during the border phase may still be relied on in the subsequent procedure, and that the examination of the application may be prioritized. Significantly, detention after the four-week limit cannot become automatic or systematic for all those who are channelled from the border to the regular procedure.

 

One Legal Fiction After Another

 

What makes Danané case especially important is that the Court relocates the decisive legal limit on the use of border procedures from geography to proper legal justification and safeguards. The judgment accepts that the same detention facility may first function as a place ‘at the border’ for the purposes of Article 43 recast APD and then, once the four-week period has expired, as a place ‘in the territory’, without any necessary change in the applicant’s physical surroundings. Instead, the Court tries to preserve a merely legal distinction between the two phases. Once the four-week period in Article 43 recast APD expires, the application no longer falls within the border procedure and his or her application will be examined under the other provisions of Directive 2013/32, while any continued detention may take place in the same facility, but must comply with the safeguards and requirements envisaged in Directive 2013/33. The Court also makes clear that detention cannot simply continue under the old border-procedure logic, that the applicant must be informed of the change in his or her legal situation, and that, where appropriate, the person must receive the document certifying their status as an asylum-seeker referred to in Article 6 of Directive 2013/33 or equivalent evidence, although in practice detained applicants often do not receive prompt documentation giving practical effect to this safeguard.

Yet the real problem is that this legal transition may leave the applicant’s position almost untouched in practice. If the same person remains in the same closed facility and under the same deprivation of liberty, the transition from border procedure to regular procedure risks becoming a legal fiction itself. That is where the judgment deserves a more critical reading. Formal entry into the territory may mean very little if it is not accompanied by adequate safeguards. The reception conditions Directive links applicant status to documentation, reception conditions, education for minors, and access to the labour market under the conditions laid down in the directive, yet the judgment says relatively little about how immediate and effective those rights must be in practice where the person remains detained. If an applicant still lacks prompt documentation, cannot effectively enjoy reception rights, and remains unable to exercise rights that normally follow from being an asylum seeker who has entered the territory and whose application is processed inside the territory under the regular procedure, then the shift from border to regular procedure serves primarily the asylum authorities, which in practice gain more time to decide while preserving continuity of control.

For that reason, continued detention after the expiry of the time limit foreseen in Article 43 recast APD must be treated with particular rigour. In this regard, it is very important that the Court itself turns from geography to safeguards, stressing that such detention must satisfy the ordinary standards of the recast RCD: necessity, proportionality, individual assessment, and one of the exhaustively listed grounds in Article 8 of the RCD, together with judicial review under its Article 9. If those requirements are not applied strictly, the transition endorsed in Danané risks turning the strict four-week time limit (extended to 12 weeks according to the Asylum Procedures Regulation, part of the EU Asylum Pact, which applies to asylum applications starting on 12 June) of the border procedure into a merely formal threshold.

This case matters beyond the Belgian transit centre and beyond the immediate interpretation of Article 43 recast APD. As my earlier analysis already suggested, the judgment fits a broader trajectory in EU asylum law in which the border is becoming less a fixed territorial line and more a shifting legal frontier that can be reproduced inside the territory. Read alongside the Pact texts, especially the Asylum Procedures Regulation, which more openly accommodates designated inland locations and expands the operational importance of border procedures, Danané looks less like an isolated ruling and more like a bridge toward a more internalised model of border procedures, in which procedures traditionally confined to the border and justified as exceptional may increasingly be reproduced inside the territory. For that very reason, safeguards become more important than ever, as the more flexible the geography of the border, the stricter the requirements governing detention, procedural guarantees, and the legal consequences of the expiry of time limits must be.

Conclusion

The deeper concern, then, is not merely that the Court has de-geographised the border. It is that, once the border is treated primarily as a legal status rather than a territorial threshold, ‘non-entry’ itself risks becoming a legal fiction capable of sustaining restrictive forms of asylum processing well inside national territory. The Court’s answer is that this is acceptable so long as the relevant guarantees, both in the asylum procedure and the detention safeguards, are preserved.

Whether that answer is convincing will depend entirely on practice. If detention after four weeks is truly exceptional and tightly reviewed, the distinction between border procedure and regular procedure may still make sense. But if applicants remain in materially unchanged detention while the determining authority simply acquires more time to decide, Danané will stand as a case of one legal fiction after another. Especially under the Pact, where screening and border procedures are further mainstreamed, the central question is whether safeguards are strong enough to prevent exception from becoming the ordinary mode of asylum governance.

 

 

 

 

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.