Time to ring the Bell: Luxembourg’s Light on Pushbacks, Strasbourg’s Shadow on Pullbacks

On Frontex’s Operational Powers, Allocation of Responsibility for Fundamental Rights Violations and Fragmented Justice

 

By Prof. Jean-Yves Carlier and Dr. Eleonora Frasca, 

Université catholique de Louvain (UCLouvain), members of Equipe droits et migrations (EDEM)

 This is a revised version of extracts from the yearly case law column “Droit européen des migrations”, published in French in the Journal de droit européen, no. 3, March 2026.

Photo credit: Luxofluxo, via Wikimedia commons

1.     The Fragmented Architecture of Accountability in EU Migration Control

The judgments of the Court of Justice of the European Union (CJEU) in Hamoudi v. Frontex (C-136/24 P) and WS and Others v. Frontex (C-679/23 P), together with the decision of inadmissibility in S.S. and Others v. Italy by the European Court of Human Rights (ECtHR), can be read as addressing a common structural problem from two different judicial perspectives: how responsibility for fundamental rights violations is allocated in a system of composite, and externalised border controls. Read together, the Luxembourg rulings on Frontex’s non-contractual liability and Strasbourg’s approach to extraterritorial jurisdiction reveal, on the one hand, a tightening of accountability within the EU legal order and, on the other, a persisting fragmentation of protection under the Convention system. Their juxtaposition reveals an emerging asymmetry between pushbacks and pullbacks and highlights the risk of a fragmented landscape of remedies in a field where operational powers are increasingly shared and strategically displaced.

The CJEU confirmed the centrality of access to the territory of the Union in contemporary EU migration policy by subjecting obstacles to external border crossings to strict judicial review. Overturning decisions by the General Court that absolved the Agency from any responsibility for its border operations (T-136/22 and T-600/21), the Court contends that Frontex may incur in non-contractual liability and must comply with fundamental rights obligations when exercising its powers in border control operations. The two rulings provide interpretative clarifications regarding the conditions for engaging Frontex’s non-contractual liability under Article 340(2) TFEU. Anchoring the Agency’s operational role firmly within the constitutional framework of fundamental rights, the Court redefined the division of responsibility between Frontex and Member States in joint operations (Hamoudi v. Frontex) and partially reshaped the concept of causation, clarifying the link between Frontex’s conduct and alleged pushbacks (WS and Others v. Frontex). In contrast, the Strasbourg Court did not approach joint state responsibility in externally coordinated maritime Search-and-Rescue (SAR) operations involving cooperation with Libya (S.S. and Others v. Italy).

The real challenge in the interpretation of EU migration and asylum law no longer lies in the technical refinement of positive norms. Rather, it stems from the operational choices through which the Union and its Member States implement migration control. In particular, maritime operations and cooperation with EU agencies as well as third countries’ national authorities generate complex legal questions precisely because border control activities are organised (and presented) as technical or operational, rather than as exercises of public authority with distinct legal consequences. Two structural features amplify this complexity. First, Frontex’s operational activities are embedded in a hybrid administrative framework that blurs the boundaries between Union and national competences. Second, joint operations disperse decision-making and execution across multiple actors, including third countries’ migration control authorities. In the context of pushbacks, this model of composite governance waters down the attribution of responsibility where illegal coercive practices occur. In the context of pullbacks, the expansion of controls at and beyond the Union’s external borders – through externalisation techniques – further complicates accountability.

2.     Asylum Seekers’ Vulnerability Requires an Adjustment of the Burden of Proof Regarding the Damage Suffered and Caused by Frontex’s Pushbacks

In the CJEU’s own words in Hamoudi v. Frontex, a pushback operation undermines the effectiveness of judicial protection for asylum seekers who have reached the territory of the Union and it “is characterised by the significant vulnerability of the persons subject to it and by the absence of the identification and personalised treatment of those persons by the authorities” (para 88). Following the Advocate General’s Opinion and on the very basis of the Frontex Regulation, the Court clarifies the division of responsibility between Frontex and the Member States: “while Frontex and the national authorities responsible for border management have a shared responsibility […], Frontex is fully responsible and accountable for any decision it takes and for any activity for which it is solely responsible under that regulation” (para 66).

For the time being, the CJEU has adopted a substantive approach to responsibility attribution, refusing to allow operational powers to serve as Frontex’s procedural shield. The Court recalls that Article 97(4) of the same Regulation “provides – like the second paragraph of Article 340 TFEU, to which it gives a concrete expression – that, in the case of non-contractual liability, Frontex is, in accordance with the general principles common to the laws of the Member States, to make good any damage caused by its departments or by its staff in the performance of their duties. Consequently, the case-law of the Court relating to that provision of the TFEU is relevant in the present case” (para 67). Next, and most importantly, after reiterating the obligation to compensate any damage, the Court acknowledges “the need to adapt the burden of proof” in respect of that damage, adapting it to the “specific circumstances” of operations conducted by Frontex, even when carried out jointly with a Member State (paras 86 et seq.).

Unequivocally, the judgment is rooted in the constitutional framework of Article 47 of the Charter, inviting the General Court, where necessary, to make use of the exceptional measures of inquiry permitted under its Rules of Procedure “in order to guarantee [effective judicial] protection […] which is fundamental in the European Union as a Union based on the rule of law” (paras 78 and 80–84). This means that Frontex can no longer rely on the complexity of its operational arrangements nor on the intermediation of Member States to escape effective judicial review of the activities for which it is “fully responsible” (para 66). The Agency cannot claim a “de facto immunity” that would hinder “all legal action by victims of a pushback operation against Frontex” (para 105). The “full respect” for the right to an effective remedy requires an “adjustment of the burden of proof,” which, in particular, must allow applicants to limit themselves to “present prima facie evidence that that operation, in which Frontex participated, occurred and that they were present during it” (para 106). In the present case, this may consist of the applicant’s testimony corroborated by a press article concerning the pushback.

The General Court ought to have granted the requests for measures of inquiry and hearings to actively seek the evidence held by Frontex, for example by ordering the Agency to produce documents in its possession. The Court notes that, on the one hand, Frontex is “likely to possess information that is relevant for the purpose of proving the existence of pushbacks” (para 96) and, on the other hand, that that failure to cooperate by Frontex “justif[ied] the General Court’s involvement” (para 148). The standard of proof must necessarily be relaxed, considering the informational asymmetry inherent in situations of pushbacks at the borders. Operational decisions adopted by Frontex must be traceable, and the Agency’s practices – long presented as purely technical – must be acknowledged in their full legal significance (on Frontex’s growing power not matching its fundamental rights responsibility, see G. Raimondo).

3.     On the Reasonableness of Asylum Seekers’ Choices in the Assessment of the Causal Link

The Court further develops the analysis of causation in light of the asylum seekers’ conduct in WS and Others v Frontex (Joint return operation). In its 2023 judgment, the General Court had taken into account numerous factual elements relating to the applicants’ conduct in order to dismiss their claim for damages. However, these elements concerned events subsequent to the refusal of entry into Greek territory and the return to Turkey, such as their departure from Turkey and their settlement in Iraq. According to the General Court, such decisions were autonomous choices, the risks and costs of which the applicants had knowingly assumed. They were, in a sense, rational choices comparable to those made by economic operators in other cases concerning the Union’s non-contractual liability. The General Court, therefore, concluded that the damage resulting from such choices could not be attributed to Frontex’s conduct, in the absence of a sufficiently direct causal link with the Agency’s actions.

At that time, we already believed that this line of reasoning was highly questionable (see our analysis in the Journal de droit européen). It characterised the subsequent conduct of the Syrian asylum seekers as “autonomous choices” and effectively neutralised the prior legal and factual constraints inherent in the asylum context. Access to the territory of the EU is not a strategic option. Without access to the territory, there can be no access to the asylum procedure. Where Frontex, through the actions of its agents – whether alone or in cooperation with those of a Member State –, unlawfully prevents access to the EU territory, those actions constitute a decisive cause of the damage consisting in the impossibility of accessing the refugee status determination procedure. Even if shared with the Member States, there is a responsibility of the Union that cannot be ignored.

Similarly, the CJEU censured the General Court’s flawed causal reasoning. The Court observed that, while an “entirely rational decision-making may be expected of economic operators experienced in the management of risks involved in the exercise of their usual activities […], such rational behaviour cannot be elevated to the rank of a criterion of general application, in particular when natural persons are concerned” (para 155). The applicants’ decision, “although not the only possible response, may be regarded as a reasonable response having regard to all the circumstances characterising that situation” (para 157, emphasis added). Such choices are not capable of breaking any sufficiently direct causal link between the conduct complained of and the alleged damage without first assessing in concreto their reasonableness “in the light of all the circumstances characterising the context in which they were made” (paras 161 and 197).

The Court thus clarifies the method for examining the causal link between alleged violations of fundamental rights committed by Frontex and the damage suffered by asylum seekers following their expulsion from Union territory. To that end, the Court provides a clear reconstruction of the obligations incumbent upon Frontex regarding the protection of fundamental rights, particularly in the context of joint return operations. These obligations include verifying the existence of “written and enforceable return decisions […] for all persons whom a Member State intends to include in such operations” (para 107), in order to ensure compliance with the principle of non-refoulement. Admittedly, the Court specifies that this obligation of verification does not automatically entail the existence of a causal link, the assessment of which “must be undertaken taking into account of all the relevant facts […] and the legal assessments required” (para 112).

Furthermore, given the joint nature of such operations, the fact that Frontex provides “technical and operational” support to Member States does not mean that any alleged violation of the asylum seekers’ fundamental rights would result exclusively from the Member State’s conduct (in this case, Greece), thereby excluding the possibility of engaging Frontex’s liability. In doing so, the Court requires a well-articulated reading of liability arising from the exercise – even in hybrid form – of the Union’s public authority, subject to full and effective judicial review.

The lesson is clear: litigation concerning Frontex cannot constitute exceptional litigation. The more autonomous capacities the Agency possesses, the more it might be held legally accountable for their exercise. Clearly defining the contours of responsibility thus becomes a central issue of the EU administrative and constitutional law. The Court’s judgment reflects a firmer recognition of Frontex’s own obligations regarding fundamental rights protection and a more open approach to causation in joint operations. It stands in clear opposition to the restrictive interpretations of the conditions for engaging the Agency’s non-contractual liability adopted by the General Court in 2023.

4.     On Allocation of Responsibilities for Fundamental Rights Violations and Competences? 

One can only endorse these two Frontex rulings, which hold the EU agency accountable for its actions. That is not to say that assessing their consequences will be straightforward. As the cases have been referred to the General Court following the annulment of its original decisions, the concrete analysis of causation, damage and compensation for the harm suffered remains open. However, it cannot be denied that the Court’s reasoning leads to a certain shared responsibility between the Member States and the EU agency. Could this division of responsibility result in joint and several liability (in solidum) of all the actors for the entirety of the damage? The question is not definitively settled. Advocate General Tamara Ćapeta devoted a fairly extensive analysis to this issue in her Opinion in WS and Others. Limiting herself on this point to examining causation, she nevertheless suggested that “in situations in which both Frontex and Member States share obligations in joint return operations, Frontex can be held liable for damage caused by the breach of such obligations, even if a Member State can be liable in parallel for the same damage” (para 93). Her reasoning drew on a possible analogy with the joint liability of another EU agency, Europol (Kočner v. Europol). In the case of Europol, it might be difficult a posteriori for a person harmed by an abusive alert to determine whether the source was the EU agency or a Member State. In contrast, with Frontex there would be, in a sense, a priori responsibility on the part of each actor – both the State and the Agency – each being fully required to prevent any serious violation of fundamental rights. Both scenarios could lead, in similar fashion, to joint and several liability.

However, the CJEU held that the plea alleging possible joint and several liability was inadmissible on the ground that it had not been raised before the General Court (WS and Others, paras 80–89). Consequently, at least in the cases at issue, it is highly likely that – not least because of the division of competences – separate findings of liability will be made corresponding to each actor’s share of responsibility. This latter scenario could generate complex litigation, leading to parallel proceedings before Luxembourg and Strasbourg. In his commentary on the two judgments, Johan Callewaert highlights their systemic significance for protection under the European Convention on Human Rights (ECHR). The complex cooperative frameworks of a hybrid administration (such as joint operations) create new situations from the ECHR perspective. Certain actions carried out on the territory and under the jurisdiction of Member States, but by EU agencies such as Frontex, result in fragmented Convention protection: “patchwork coverage”. Some actions would remain covered by the Convention when attributable to national authorities, whereas those attributable to EU bodies would escape it and fall instead within the jurisdiction of the EU courts, with the result that the ECtHR would no longer be able to hold a State liable for the entirety of the damage arising from a joint operation. In other words, while the 2025 Frontex case law shed some light – by reducing the grey areas regarding EU liability – it also cast a shadow, as any light does, in this case concerning the extent of damage that the ECtHR may attribute to the responsibility of States. The way out of this paradox would, of course, be the completion of the EU’s accession to the ECHR. Yet that prospect still appears remote.

5.     Non-Justiciability of Human Rights Violations Arising from the Coordination of Search-and-Rescue (SAR) Operations at Sea

While the evolving case law concerning Frontex has led the two European courts – Luxembourg and Strasbourg – to review, concurrently if not jointly, compliance with fundamental rights in pushback operations at the borders, for the time being, so-called pullback operations still escape review by the Strasbourg Court. Even if the ECtHR emphasised that “entering into bilateral agreements on migration with third States has the effect of placing extremely vulnerable individuals at serious risk of infringements of their fundamental rights” (S.S. and Others v. Italy, para 110), such arrangements do not automatically bring those individuals within the jurisdiction of a Contracting State for the purposes of the Convention, thereby limiting access to Strasbourg protection.

The inadmissibility decision delivered in S.S. and Others v. Italy reveals the limits of Convention responsibility in the face of the EU policy of delegating migration control, including maritime control, to third countries. The involvement of EU Member States in SAR operations in the Mediterranean cannot be understood without reference to the central role entrusted, for more than a decade, to the border coast guards of third countries such as Libya or Tunisia. In the absence of an integrated EU SAR programme, the Union and its Member States have progressively externalised the management of SAR obligations, while adopting an increasingly restrictive approach toward humanitarian operations conducted by non-governmental organisations (NGOs). This development has been accompanied by financial, material and operational support to third countries: provision of vessels, training, technical assistance and the establishment of a coordination centre (on this topic see E. Frasca). Although the strengthening of SAR capacities in these countries is officially carried out in the name of saving human lives, it is structurally linked to the objective of preventing irregular Mediterranean crossings into the EU territory.

By relocating such actions, externalisation also makes it possible to shield these practices from scrutiny under the obligations arising from the ECHR. Yet this strategy unfolds in a context marked by interceptions at sea that systematically expose migrants to serious violations of their fundamental rights, including treatment contrary to Articles 2 and 3 of the Convention. Like the cases concerning Frontex before the CJEU, S.S. and Others v. Italy raises the issue of the dilution and segmentation of the exercise of powers and, consequently, of responsibilities in the conduct of SAR operations at sea. The Strasbourg Court confirms that its jurisdiction over extraterritorial actions of Contracting States remains exceptional. To fall within its jurisdiction, such extraterritorial actions require the establishment, in the control exercised over individuals, of a direct link with the respondent State. In this way, a dissociation occurs between factual causation and legal attribution, creating a procedural obstacle that is practically insurmountable (on this topic, see C. Ryngaert). In the present case, the Court held that the financial, logistical and operational support provided by Italy to strengthen the capacities of the Libyan coast guard does not, absent effective control or direct operational direction, amount to sufficiently decisive influence to trigger jurisdiction within the meaning of Article 1 of the Convention. In doing so, it effectively casts a veil of ignorance over any responsibility of a Contracting State under the ECHR. Any notion of indirect, functional, or remote control is rejected, even where the risks faced by the persons “rescued” at sea – only to be returned to Libya and subjected to torture and inhuman treatment that has been widely documented – are fully known.

Conclusion: A Rebuttable Presumption Grounded in Prima Facie Evidence

Rulings on matters of access to the EU territory are likely to become more and more significant, signalling a new focus not only by Member States but also by national courts and, consequently, by the CJEU and the ECtHR, on the control of the Union’s external borders and their heir harmful consequences on asylum seekers. The ECtHR’s inadmissibility decision may be read as consistent, confirming the Court’s settled case law on extraterritorial situations (Banković, Al-Skeini, M.N. v. Belgium). It may also be read as a retreat from an evolutive interpretation of fundamental rights in light of today’s conditions. Indeed, a well-established case law recognises indirect responsibility or secondary accountability (par ricochet) in cases of extradition or refoulement, and this restrictive interpretation of jurisdiction refuses to adapt this case law to the reality and contemporary transformations of migration control. There is a middle ground between the frequently criticised slippery slope of judicial overreach whereby Courts are deemed overly protective and the equally problematic territorial lock that shields fundamental rights violations of a serious nature from scrutiny. The middle ground advocated here may appear modest: a rebuttable presumption grounded in prima facie evidence. However, it is precisely in such narrow interpretative openings that the protection of fundamental rights may evolve, by adapting interpretation to the context – a context marked, on the one hand, by situations of extreme vulnerability and, on the other hand, by elements such as bilateral agreements and conspicuous funding which make it possible to presume a genuine link between the contested acts and a Contracting State. Just as the Luxembourg Court has done with regard to Frontex, the Strasbourg Court would be well advised to accept that prima facie evidence gives rise to a rebuttable presumption (juris tantum) requiring the Contracting State to produce evidence capable of rebutting responsibility through facts and documents that would exonerate it. “There’s a crack in everything. That’s how the light gets in”, sang the poet Leonard Cohen. In the present context, we believe that European courts and judges must now more than ever try to maintain a balance. That crack is the space within existing doctrine, mindful of the rights of the individuals concerned, that allows Courts to remain faithful to their jurisdictional frameworks while adapting to new realities of migration control.

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.