I. Introduction
The migration crisis and the European Union (EU) response from 2015 onward constitutes one of the most significant political and legal challenges in the history of European integration.Footnote 1 Triggered by several factors, including armed conflicts, climate change, and economic instability fueled by historical inequalities, over one million migrants and asylum seekers arrived in the EU in 2015 alone, largely through the Eastern Mediterranean and Western Balkan routes.Footnote 2 In response, the EU has progressively shifted toward a model centered on border management, externalization of migration control, and enhanced operational cooperation, culminating in the strengthening of the European Border and Coast Guard Agency (Frontex) mandate.Footnote 3 Ten years later, a glimmer of possibility for judicial review appears before EU courts, with the European Court of Justice decisions in WS and Ors. and Hamoudi.Footnote 4
Frontex’s current legal framework is primarily based on Regulation (EU) 2019/1896, which replaced earlier instruments, including Regulation (EU) 2016/1624,Footnote 5 the legal framework applicable in the WS and Ors. and Hamoudi cases. Acting on the basis of Articles 77 and 79 of the Treaty on the Functioning of the European Union (TFEU), which confer competence upon the EU in matters of border management and migration policy, Frontex coordinates joint operations, provides technical and operational assistance to member states, supports return procedures, and deploys standing corps personnel at the EU’s external borders.Footnote 6 Regulation 2016/1624 introduced the legal notion of “shared responsibility” between Frontex and member states in border management.Footnote 7 Member states retain “primary responsibility” for the management of their external borders, while Frontex supports the implementation of EU measures by reinforcing, assessing, and coordinating the actions of member states. Although member states formally retain responsibility for border control and migration decisions, successive reforms have considerably expanded Frontex’s operational role, transforming it from a coordination body into an agency exercising significant executive functions within systems of shared administration.
In the last ten years, shared responsibility has become a blame-shifting exercise, under which “there is always someone else to point the finger at.”Footnote 8 In a frequent pattern for the responsibility of international organizations, member states and Frontex defend themselves hiding behind a complex chain of command, between the EU institutional veil and attribution of competences.Footnote 9 Impunity is also fostered by the lack of judicial fora opened to victims. Currently, the EU legal system does not provide human rights accountability mechanisms before which victims of pushbacks and other violations related to the EU border management can seek justice. Similarly, the international legal system does not have judicial fora in which the activities of international organizations can be scrutinized. As a result, victims resort to a patchwork of remedies, suing states before regional human rights courts and EU agencies before EU courts. Fragmentation into different fora has favored the blame-shifting exercise that has resulted in impunity and prevented the realization of a comprehensive system under which forms of co-perpetration and dual responsibility could effectively emerge. This evolution has raised fundamental questions concerning accountability and the extent to which Frontex may incur legal responsibility for violations of EU fundamental rights during the operations it coordinates or supports.Footnote 10
In this context, on December 18, 2025, the Court of Justice of the European Union rendered its decisions in the cases WS and Ors. v. Frontex (WS and Ors.) and Hamoudi v. Frontex (Hamoudi),Footnote 11 overruling the orders of the General Court of the European Union of September 6 and December 13, 2023, respectively.Footnote 12 The two cases have been brought by victims of return (WS and Ors.) and pushback (Hamoudi) operations and have been debated together both by the Court and legal scholarship. In the context of the broader concern for the EU and member states de facto impunity for their migration policy, these decisions have been welcomed as a groundbreaking attempt to bring justice to victims.Footnote 13 In particular, they focus on two core legal issues that prevented reparations until today, the burden of proof and the causal link between damages and relevant conduct.
The General Court dismissed both cases in first instance in 2023. In WS and Ors., the Court rejected the claim of damages for the return operation on the basis that the applicants had not adduced evidence of a sufficiently direct causal link between the damage and the conduct. In Hamoudi, the General Court rejected the claim because the applicant had not demonstrated actual damage. The groundbreaking decisions of the Court of Justice on appeal open the way for a judicial remedy, interpreting EU liability law to enable individuals to claim compensation from Frontex. WS and Ors. focused on the blame shifting exercise between the EU and its member states, clarifying that responsibility of member states does not absolve Frontex. Hamoudi focused on the burden of proof, rejecting the “probatio diabolica” established by the General Court. Upholding the appeal, the Court of Justice sent back both cases to the General Court for reconsideration, but, independently from their outcomes in the specific circumstances, these decisions set a fundamental precedent for all future actions against Frontex and other EU agencies.
This review essay aims at analyzing these groundbreaking decisions in the context of the international responsibility of the EU and of its member states for their migration policy, in the framework of theories on co-perpetration and the relationship between international law and EU law. I will focus on the relevance of the two decisions in the context of the debates on shared responsibility, distinguishing between causality and the burden of proof. I will conclude by framing the issue as one of so-called “co-perpetration,” defining its main characteristics.
II. WS and Ors. v. Frontex: The Causal Link Between the Conduct and the Damage
In WS and Ors. v. Frontex, the applicants are a Syrian family of Kurdish origin composed of six persons who arrived on the Greek island of Milos in October 2016 among a group of refugees from Türkiye, during the period following the EU-Türkiye Statement on migration.Footnote 14 The EU-Türkiye Statement was designed to stop irregular migration flows from Türkiye to Greek islands in exchange for financial support, refugee assistance, and other cooperation measures. It is widely criticized for being morally wrong and conducive to gross human rights violations.Footnote 15 Upon arrival, WS and his family were transferred to a reception and identification center in Greece, where they expressed the intention to seek international protection.Footnote 16 According to the applicants, contrary to being told that they would be transferred to Athens the next day, they were subjected to a joint return operation coordinated by Frontex and returned to Türkiye.Footnote 17 It was only when they disembarked from the plane and were handed to Turkish officials that they realized that they had been taken back to Türkiye.Footnote 18 They received temporary protection documents from Türkiye, but on fear of being returned to Syria, they moved to Iraq and settled in Erbil.Footnote 19
In January 2017, the applicants filed a complaint before Frontex concerning the illegality of the joint return operation. This was followed by a second complaint in June 2018 due to the inertia of the first proceeding. In November 2019, the applicants were informed that after internal disciplinary investigation, no liability was established and the case was closed. In the interim, the applicants sued Greece before the European Court of Human Rights, claiming violations of the prohibition of torture and the right to an effective remedy. The case was settled out of Court with Greece agreeing to compensate the applicants with 75,000 Euro for material and non-material damages.Footnote 20
In September 2021, the applicants brought an action for damages against Frontex before the General Court based on TFEU Article 268, the second paragraph of TFEU Article 340, and Article 60(3) of Regulation 2016/1624. An action for damage is based on civil law standards for non-contractual damages, and it is not a human rights complaint mechanism, nor does it establish international responsibility.Footnote 21 EU law requires three cumulative conditions to be fulfilled to grant compensation: a serious breach of an EU law that confers rights on individuals; the causation of damages; the causal link between the breach of the obligation and the damages.Footnote 22 The applicants claimed that non-contractual damage derived from Frontex’s failure to comply with its obligations under EU law, including the EU Charter of Fundamental Rights. In particular, the applicants made eight pleas, accusing Frontex of failing to prevent the breach of fundamental rights during the preparation, in the course of the joint return operation, and when handling the subsequent complaints.Footnote 23
In September 2023, the General Court dismissed the action in its entirety, claiming that the applicants had not proved a sufficiently direct causal link between the damage invoked and Frontex conduct.Footnote 24 In particular, the Court contended that Frontex’s fundamental rights obligations are applicable only in the performance of its tasks, which are limited to providing technical and operational support to member states and not to deciding on the merits of return decisions adopted by member states.Footnote 25 Since Frontex is not competent regarding the merits of the return decision, nor for international protection applications, a causal link between the damage and Frontex’s conduct could not be established.Footnote 26 The Court also stressed that the decision to move to Iraq after receiving temporal protection in Türkiye was made by the applicants and could not be considered a direct consequence of Frontex’s conduct.Footnote 27
The applicants appealed the judgment to the Court of Justice, who partially ruled in their favor in December 2025 with the decision under discussion. Reversing the first instance decision, the Court focused on the causal link between the damage and the conduct, holding that Frontex is under an obligation to verify that the competent member state issued a return decision and that an omission triggers its liability.Footnote 28 Frontex’s competence for coordinating joint return operations implies a requirement of supervision to ensure that return decisions exist, in order to ensure that they comply with EU law and the principle of non-refoulement.Footnote 29 The General Court erred in considering that Frontex’s role consists solely of providing technical and operational support to member states, without the power to verify the existence of return decisions. While member states have the competence of deciding on the merit of the return decisions, Frontex should have verified the existence of such decisions and omissions may trigger its liability. Consequently, the Court explained that the requirement of a causal link between the conduct and the damage has to be examined in the light of Frontex obligations.
Moreover, the Court of Justice ruled in favor of the appellants’ challenge of the General Court ruling that liability for any violation would lie exclusively on member states.Footnote 30 The Court of Justice confirmed that under EU regulation a host state is liable for any damage caused by Frontex’s officials, which operate under the instructions and in the presence of state officials. However, this does not exclude any liability on the part of Frontex for any damage caused by its conduct in the performance of its duties. In the WS and Ors. case, the Court of Justice found that Frontex is under the duty to ensure compliance with fundamental rights in the context of the return operations it coordinates and co-perpetration cannot be excluded.Footnote 31
This revolution was initiated by Advocate General Tamara Ćapeta in her opinion in WS and Ors., clarifying that the General Court failed by excluding the relevance of Frontex’s obligations and its omissive conduct.Footnote 32 She clearly distinguished between the position of member states and the position of Frontex, opening the way to a form of co-perpetration, in which more than one subject is involved in parallel wrongful acts. The return order carried out by member states does not exclude the parallel omission of Frontex, which also contributed to the damage.Footnote 33 She explained that the General Court confused the active conduct of the member states with the omission of Frontex and erred in dismissing the claim for the lack of a causal link with the damage. She concluded by claiming that Frontex can be liable in parallel with a member states for the violation of shared obligations.Footnote 34
Advocate General Ćapeta further asked whether the shared responsibility of Frontex and member states may lead to joint and several liability. As the Court explains, joint and several liability means that each of the co-perpetrators is required to compensate the damage in its entirety and the injured party is entitled to claim from each of them full compensation until satisfaction. Furthermore, the person liable who has made full reparation may bring an action for contribution against any other co-perpetrator.Footnote 35 Recently, the Court of Justice has referred to joint and several liability even when it is not expressly provided for by EU legislation.Footnote 36 As such, joint and several liability may arise in the context of multi-actor activities between EU agencies and member states. However, in WS and Ors. it refrained from determining whether Frontex and Greece are joint and severally liable, because it considered it a new inadmissible question posed by the applicants only at the appeal stage.Footnote 37
III. Hamoudi v. Frontex: The Burden of Proof
In Hamoudi v. Frontex, the applicant is a Syrian national who arrived in the Greek island of Samos in April 2020 from Türkiye with a group of twenty-two refugees seeking asylum in Greece.Footnote 38 He claimed that during his time at sea, a surveillance airplane operated by Frontex flew over the scene twice. Once disembarked in Samos, the migrants asked the local population to find police officers.Footnote 39 When they arrived, the police confiscated their phones and drove them back to the beach and put them on board a life raft. Once at sea, the group was forced to change boats twice more before the Türkiye coast guard picked them up and brought them back to Türkiye. Following the push back operation, Hamoudi was transferred to a detention center in Türkiye where he was detained for ten days before receiving an expulsion order and the confiscation of his Syrian passport.Footnote 40 Under the threat of refoulement to Syria, he managed to arrive in Germany, where he lodged an application for international protection.
The applicant claimed before the General Court that Frontex failed to comply with its obligations under EU Regulation 2019/1896 and the EU Charter of Fundamental Rights.Footnote 41 In particular, he made three pleas, covering Frontex failures to exercise its supervisory role, including precautionary measures, and direct involvement in the pushback operation, which led to violations of human dignity and integrity, torture and inhuman and degrading treatment, and deprived him of the right to asylum and protection from collective expulsion. Alternatively, he claimed that Frontex should be considered responsible for aiding and abetting Greek authorities in the violation of fundamental rights.Footnote 42
The General Court dismissed the claim, arguing that the evidence produced by the applicant was manifestly insufficient to demonstrate that he was present and involved in the incident.Footnote 43 Indeed, in order to prove his claim, the applicant relied on his own witness statement, a Bellingcat media article published online, two YouTube videos included in that article, and four photographs that are color screenshots taken from the videos featured in that article. Based on the evidence produced, the judges contended that Hamoudi did not prove any damage, and, consequently, dismissed the case without addressing attribution of conduct and causality.
The appeal brought against this decision insisted on the failure by the General Court to correctly apply the burden of proof, imposing a standard that is impossible to meet in the context of pushback operations.Footnote 44 This resulted in a probatio diabolica that infringed fundamental rights, including the right to an effective and complete system of remedies. The Court of Justice ruled in favor of the appellant, insisting that the guarantee of effective judicial protection of individuals includes rules on the burden of proof and the taking of evidence.Footnote 45 To this effect, courts may adopt measures requiring the parties to produce all documents and supply all information they consider necessary. The Court also contended that Frontex is likely to possess relevant information to prove the pushback, given its competence and presence in the relevant area.
Reversing the decision, the Court of Justice claimed that the General Court failed to adapt the burden of proof to the specific circumstances of the pushback, in which the person subject to it faced significant vulnerability and difficulty collecting evidence. These persons cannot be requested to adduce conclusive proof, and the failure to adapt the burden of proof might hinder all legal actions against Frontex. Consequently, the respect for the right to an effective remedy requires a lower standard of proof.Footnote 46 Once prima facie evidence is provided by means such as videos, testimony, and NGO reports, the Court should investigate further, adopting measures to obtain from Frontex all relevant information to clarify the facts of the case. The General Court should not have dismissed the case, as Hamoudi provided prima facie evidence of a pushback that should be further examined.
The Hamoudi decision complements WS and Ors. by determining that access to justice should not be impaired in situations in which it is extremely difficult to provide conclusive proof on the causation of the damage. Rejecting the probatio diabolica established by the General Court, it allows victims of multi-actor activities to demand reparations based on prima facie evidence that courts are required to corroborate through further investigation. A new case (FM v. Frontex) is already pending before the General Court, in which the principles established in Hamoudi may have a broader impact.Footnote 47
IV. Co-perpetration as the Fundamental Model for EU and Member States’ Responsibility
In this last part, I will reflect on the outstanding implications of these decisions for the EU international responsibility. I will first situate the discussion in the relationship between EU liability, international responsibility, and shared responsibility. Then, I will present how the two decisions open the way to frame the relationship between Frontex and EU member states in terms of co-perpetration. Co-perpetration remained at the margin of the discussions on shared responsibility, which focused more on complicity and other forms of responsibility in connection with the act of another person. Conversely, co-perpetration will be defined as a situation in which two or more international legal persons, states, or international organizations jointly commit parallel and identical wrongful acts, either through a single integrated course of conduct or through multiple independent actions that occur in a coordinated context.
A. The Limits of the Available System of Remedies and the Relationship Between EU Liability, International Responsibility, and Shared Responsibility
The available system of remedies against the EU migration policy has so far proven incapable of dealing with situations in which the EU and member states cooperate and may assume joint responsibility. An individual can either initiate proceedings before international human rights courts for the responsibility of member states or before EU courts for the responsibility of the EU, in the absence of joint proceedings. Within the EU, the lack of venues to challenge Frontex activities limits accountability systems to three options: an action for annulment under TFEU Article 263; holding Frontex responsible for failing to act under TFEU Article 265; or seeking non-contractual damages under TFEU Article 340 as in the WS and Ors. and Hamoudi decisions.Footnote 48
Concerning actions for annulment, litigation against Frontex has focused on access to information. This is not a proper forum to demand accountability for the EU migration policy, but, particularly in the context of complex chains of command, access to information to prove the facts of the case is pivotal to giving victims a chance of redress, as evidenced in the Hamoudi decision. In Naass and Sea-Watch v. Frontex, the applicants sought the annulment of a Frontex decision to reject a request for information in its possession on its alleged involvement with a Libyan operation within the Maltese search and rescue zone on July 30, 2021.Footnote 49 Frontex rejected the requests based on a public security exception. The applicants argued that Frontex is required to explain why disclosure of information “could specifically and actually” undermine public security and that the associated risk must be “reasonably foreseeable and not purely hypothetical.”Footnote 50 The EU General Court distinguished between the obligation to state reasons and the merit of the reasoning. The obligation to state a reason is a procedural requirement that must be appropriate to the measure in question and brevity may be justified by the need not to undermine the interests protected.Footnote 51 On the merits of the decision, the Court argued that the applicants have not challenged the reasoning of the decision,Footnote 52 and rejected the action except for what concerns two documents containing more than one hundred photos, and only because Frontex did not mention the existence of the photographs in question.Footnote 53 This action for annulment shows how it can be difficult for victims to prove the facts of the case when all information is in the hands of Frontex and Frontex does not release it. The requirement of a lower prima facie burden of proof discussed in the Hamoudi case should also be interpreted in this context and it may lead to compelling Frontex to share information on its operations.
An action under TFEU Article 265 for failing to act is another improper way to seek Frontex accountability in the absence of an effective judicial venue. In the case SS and ST v. Frontex, the applicants claimed that human rights violations occurring during Frontex’s activities in the Aegean Sea compel the application of Article 46(4) of the 2019 European Border and Cost Guard Regulation, requiring the withdrawal of financing or suspension or termination of any activity if the Frontex executive director considers that serious violations of fundamental rights are happening.Footnote 54 The Frontex executive director rejected the request, contending that the agency’s activities in the Aegean Sea region had been carried out in strict compliance with the applicable legal framework. Consequently, the applicants brought a case before EU Courts for failure to act. The EU General Court declared the action inadmissible considering that Frontex had defined its position and, therefore, was not failing to act.Footnote 55 This second attempt to make Frontex accountable encountered severe limitations due to the application of administrative law procedures such as the failure to act, that cannot be considered as human rights accountability mechanisms.
Finally, the last way that EU law permits making Frontex accountable is through an action for damages under TFEU Article 340, as presented in the WS and Ors. and Hamoudi cases. Again, this is not a human rights complaint mechanism, and it is based on civil law standards for non-contractual liability, including a serious breach of EU law, the production of damages, and a direct causal link between the breach and the damages caused. However, the importance of the appeal decisions under discussion goes beyond the EU non-contractual liability regime and involves broader issues of international responsibility, particularly in terms of attribution of conduct and causality, on which key problems of multi-actor operations lie.
The EU liability regime and international responsibility are not mutually exclusive. The former does not supersede the latter. They are not in a relationship regulated by the lex specialis derogat generali principle.Footnote 56 The EU legal system is not lex specialis in relation to international law and EU law does not supersede international law. They constitute two different legal systems connected by the factual element that EU fundamental treaties are instruments of international law, like all constitutive instruments of international organizations.Footnote 57 The dual nature of constitutive instruments, acting as both a treaty and a constitution, affects the legal systems they produce and establishes a relationship between institutional law and international law that is not reducible to a lex specialis/lex generali dynamic. The EU liability regime constitutes a different (institutional) accountability mechanism parallel to international responsibility. Moreover, EU non-contractual liability is structurally different from international responsibility. As mentioned, non-contractual liability requires the causation of damages. Conversely, international responsibility only consists of attribution of conduct and violation of an obligation. Under international law, attribution of conduct is a technical operation to link conduct to its author(s), through criteria that can be based on an institutional link (such as a state organ) or a factual link (such as effective control exercised by a state over a non-state actor).Footnote 58 Conversely, causality links the wrongful conduct to the damage, and, under international responsibility, it is only necessary to determine the obligation to make a full reparation.Footnote 59 The attribution of the conduct that violates an international obligation is enough to establish international responsibility.
Despite the difference between attribution and causality, recent debates on so-called shared responsibility have relied on causality and the production of an indivisible injury as an essential element to identify situations in which more than one legal person is involved. Indeed, “shared responsibility” has been defined as “the commission by multiple international persons of one or more internationally wrongful acts that contribute to an indivisible injury.”Footnote 60 As such, shared responsibility becomes an alternative concept to international responsibility. The causation of an indivisible injury as an essential characteristic of shared responsibility makes it different from international responsibility and more similar to the EU non-contractual liability regime.
As the WS and Ors. and Hamoudi cases show, causation is extremely difficult to prove and may lead to impunity if the chain of command is difficult to establish and the burden of proof is not shared with the entity in possession of relevant information. Indeed, individuals harmed by the activities of international organizations are often in the almost impossible situation of proving the relationship between member states and the organization.Footnote 61 For instance, unfortunate decisions before Dutch courts contended that the family of the victims of the Srebrenica genocide have the burden of proving the causal link to establish attribution of conduct to the Dutch military. Ignoring organic control between a state and its military, Dutch courts included causality in the interpretation of “effective control” as the factual criterion to attribute the conduct of the Dutch military in Srebrenica to the United Nations and/or to the member state. Dutch courts claimed that effective control means establishing a link between the damage and its causally proximate form of control from which the wrongful conduct arose.Footnote 62 Under this argument, victims—for instance, the families of the estimated 10,000 victims of the cholera outbreak in Haiti—have the burden of proving the link between the damage and the causally proximate form of control from which the wrongful conduct arose.Footnote 63 This is again a probatio diabolica that makes attempts of accountability almost impossible. To rely on causality to attribute the conduct to an international organization or to its member states is contrary to the customary rules of international responsibility, do not define effective control over the conduct as requiring an assessment of the eventual damage caused.Footnote 64 It is a conflation of attribution of conduct and causality that derives from the notion of shared responsibility, requiring the causation of an indivisible injury.
B. EU International Responsibility and the Emergence of Co-perpetration
The same conflation of attribution of conduct and causality affects the debates on EU international responsibility, stemming from attempts to “externalize” the EU regime into international law, positions advanced by the EU Commission before the International Law Commission and international tribunals.Footnote 65 By externalization, I mean the attempt to rely on the characteristics of the EU as a self-proclaimed special international organization to adapt the rules of international responsibility to its benefit. Meaning, it intends to make international responsibility more similar to the regime applicable to EU non-contractual damage. In particular, the EU Commission has favored criteria of attribution of conduct that set aside customary international law. It has argued that the internal division of competences between the EU and its member states should determine who between the state or the EU is internationally responsible and not the customary attribution of conduct based on institutional or factual criteria.Footnote 66 For instance, as fishing is an exclusive competence of the EU, the EU should be responsible for all conduct of member states in this domain.Footnote 67 Contrastingly, as border management is a shared competence, member states are primarily responsible. In a similar formulation, the normative control exercised by the EU on its member states should determine attribution of conduct in favor of the former.Footnote 68 Normative control implies the possession of exclusive competences, and both are based on the same conceptualization of the EU as an exceptional organization that requires special rules of responsibility.
This position of the EU Commission reproduces at the international level the argument applied by the General Court in WS and Ors. in the context of EU non-contractual liability.Footnote 69 Indeed, the General Court claimed that the applicants did not prove a causal link between the conduct and the damage because under the EU division of competences, Frontex has a support and coordination role only, and the member states bear all responsibility. Shared competences lead to a form of shared responsibility that, as envisaged by Regulation 2016/1624, becomes a shield from accountability because it impairs dual attribution. Indeed, none of the criteria of attribution of conduct based on causality advanced by the EU Commission envisage the possibility of multiple attribution to both the EU and its member states. By setting aside this argument, the Court of Justice rejected the relevance of causality in the test for the attribution of conduct, rightly distinguishing it from the causation of damages. This argument also opens a new scenario for international responsibility, clearly distinguishing between attribution of conduct and causation, and explaining that dual attribution is a viable option and not an exception, especially in multi-actor operations in which more than one legal person is involved.
The inclusion of causation as an essential element makes the cases of shared responsibility particularly complex, without adding benefit.Footnote 70 In line with the WS and Ors. appellate decision, Frontex’s joint operations can be considered cases of co-perpetration that can be defined as a situation in which two or more international legal persons, states, or international organizations jointly commit parallel and identical wrongful acts, either through a single integrated course of conduct or through multiple independent actions that occur in a coordinated or concurrent context. The active conduct of member states is parallel to the omissive conduct of Frontex, and they both concur to the violation of identical obligations.
Under co-perpetration, each actor is individually responsible for committing a wrongful act that is substantively identical to the acts of the other(s), even if those acts are not functionally interdependent. This stands in contrast to complicity or other forms of “responsibility in connection with the act of another,” which rely on establishing a causal link and/or intent to support the primary actor’s conduct.Footnote 71 Co-perpetration, by contrast, does not require that one act be dependent on the other. The focus is on the independent attribution of wrongful conduct to each actor and the violation of an international obligation that is either formally or substantively identical across the actors. As such, co-perpetration focuses on the application of Chapter II of the Articles on Responsibility of States for Internationally Wrongful Acts and the Articles on the Responsibility of International Organizations, and not Chapter IV.Footnote 72 It mainly focuses on the attribution of conduct, rather than on the responsibility of one subject in connection with the act of another subject.
Co-perpetration is conceptually simpler and normatively appropriate when actors engage in wrongful conduct independently but in parallel, next to each other. It does not require legal or factual subordination between actors, nor does it rely on proving that one actor’s behavior was contingent upon another’s. Each actor is directly responsible for their own conduct, and yet the collective effect of their actions constitutes a broader violation. There is no need to discuss “aid and assisting,” “direction and control,” “coercion,” or “circumvention of international obligations through decisions and authorizations addressed to members,” when responsibility can be effectively attributed as a form of co-perpetration applying the customary criteria on the attribution of conduct.
V. Conclusion
The judgments of the Court of Justice in WS and Ors. v. Frontex and Hamoudi v. Frontex mark a decisive turning point in the evolving architecture of accountability for EU border governance. While the cases formally concern the interpretation of TFEU Article 340(2) and the procedural conditions of EU non-contractual liability, their implications extend far beyond the confines of EU law. They represent the first serious judicial attempt to dismantle the structural impunity that has characterized the system of shared migration management between the EU and its member states.
At the core of both appeals lies a methodological correction. In WS and Ors., the Court of Justice rejected the General Court’s restrictive reading of causality, clarifying that the division of competences between the EU and member states cannot serve as a shield against liability. Frontex’s role as a coordinator does not reduce its legal obligations to a merely technical function. Where the agency is entrusted with supervising, organizing, and implementing return operations, it must ensure compliance with fundamental rights. The fact that a member state adopts the formal return decision does not preclude the agency’s concurrent responsibility for omissions within its sphere of competence. In this respect, the Court dismantled the blame-shifting logic that had previously insulated Frontex from judicial scrutiny.
In Hamoudi, the Court complemented this development by recalibrating the burden of proof. By rejecting the probatio diabolica imposed by the General Court, it reaffirmed that effective judicial protection under Article 47 of the Charter requires procedural standards adapted to the realities of pushback operations. Individuals subject to opaque, multi-actor border practices cannot be expected to produce conclusive proof of facts that lie largely within the institutional control of the Frontex. Prima facie evidence must suffice to trigger judicial investigation, particularly where fundamental rights are at stake. This adjustment is not merely procedural, but structural. Without such flexibility, the EU liability regime would remain practically inaccessible in precisely those contexts where accountability is most needed.
Taken together, the two decisions signal a conceptual shift from a paradigm of exclusive attribution to one of concurrent responsibility. The Court of Justice implicitly endorses the possibility that Frontex and member states may incur parallel liability for substantively identical breaches arising from coordinated operations. Although it stopped short of expressly recognizing joint and several liability, it made clear that dual responsibility cannot be rejected a priori. In doing so, the Court aligned EU liability law more closely with the logic of international responsibility while preserving the structural autonomy of the EU legal system.
The broader theoretical contribution of these rulings lies in clarifying the distinction between attribution of conduct and causation. By disentangling the two concepts, the Court avoided the conflation that has long complicated debates on shared responsibility in both EU and international law. Attribution concerns the linkage of conduct to an actor, while causation concerns the relationship between wrongful conduct and damage. The General Court’s error consisted precisely in collapsing these two inquiries into a single analysis grounded in the attribution of competences. The Court of Justice restored analytical clarity by reaffirming that multiple actors may independently breach overlapping obligations within the same operational framework.
This clarification opens the door to conceptualizing Frontex operations as instances of co-perpetration rather than complicity or other forms of responsibility in connection. In multi-actor border management, the agency and member states may each violate identical obligations without subordinating one to the other. Co-perpetration provides a more accurate and normatively coherent framework than models based on aid, assistance, or direction and control. It reflects the reality of integrated governance structures in which authority is distributed, and responsibility remains linked to the conduct of each participant.
Whether these doctrinal advances will translate into effective remedies in practice remains to be seen. The cases have been remitted to the General Court, and the ultimate outcome may depend on the specific factual findings. Yet independently of their final resolution, the appeal judgments establish an enduring precedent. By recalibrating causality and the burden of proof, the Court of Justice has weakened the structural foundations of EU border impunity and laid the groundwork for a more coherent system of responsibility in shared migration governance. The end of Frontex impunity may not yet be complete. But for the first time, it is no longer doctrinally implausible.
