Wednesday, 29 April 2026

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.


No comments:

Post a Comment