On 18.12.2025, the Court of Justice of the European Union (CJEU) issued its judgment on the appeal against the General Court’s Order in W.S. and Others v Frontex. The judgment is uniquely important as the CJEU deviated from the academics’ view that the causal inquiry shall be conducted in two steps and used instead the causal test of directness. Given that adjudicating fora are increasingly relying on less demanding causal tests, in particular the test of foreseeability in cases of multi-actor involvement in a refoulement, the CJEU should implement this method where Frontex and the EU Member States are involved in refoulement-related damage.

Background of W.S. and Others v Frontex

W.S. and Others v Frontex concerned several Syrian nationals, who, a few days after their arrival in Greece and despite having expressed their intention to apply for international protection, were deported to Türkiye in a joint return operation undertaken by Greece and Frontex (paras. 2–16). The applicants sought compensation for the damage flowing from Frontex’s failure to confirm the existence of a return decision concerning them. However, their action for damages was dismissed, after the General Court found that Frontex was not competent to assess return decisions or international protection applications, and, thus, that there was no causal link between Frontex’s conduct and the damage suffered by the protection-seekers (para. 66). Thereafter, the applicants appealed to the CJEU.

In her Opinion, the Advocate General proposed the following. When the CJEU assesses the fulfilment of the conditions for the determination of the EU’s non-contractual liability (i.e., the existence of an unlawful act committed by Frontex, the existence of actual damage suffered by the appellants, and the existence of a causal link between the two), it ought to first explore whether wrongful conduct had been committed by Frontex. Only thereafter it should proceed with exploring the factual and legal causation (i.e. whether Frontex’s wrongful conduct was a cause in fact and in law of the damage suffered). However, the CJEU proceeded otherwise. It held that since the conditions for the determination of the EU’s non-contractual liability are cumulative, the EU judicature does not necessary need to consider all of them, if one of them is not met (para. 61). Henceforth, it can directly assess the existence of a causal link between the alleged wrongful conduct and the damage even without explicitly identifying the wrongful act and the damage.

The CJEU acknowledged that under the 2016 Frontex Regulation, Frontex bears the obligation to confirm the existence of return decisions for all persons included in return operations (paras. 101–102). However, in the CJEU’s view, Frontex’s violation of this obligation did not automatically give rise to Frontex’s liability. To this end, the existence of a direct causal link between Frontex’s omission and the damage alleged would need to be further examined (paras. 111–112, 148).

The CJEU reiterated that a direct causal link exists when the damage flows sufficiently directly from the unlawful conduct, i.e. when the latter is the determining cause of the damage (para. 150). Nonetheless, the Court further recalled that even when this causal link is established, it may be broken, inter alia, by a contributing act of the affected person, which occurs between the conduct complained of and the damage suffered, if that act constitutes the determinant cause of the damage (para. 151). This could be the case also in the instance of the appellants, who upon their return to Türkiye decided to travel to Erbil.

However, the Court further underlined that the assessment of whether such a contributing factor is capable of breaking the causal nexus requires a consideration of the circumstances, under which the decision was made, and of the particular vulnerability of asylum-seekers, which is likely to affect their judgment, due to their migration and the traumatic experiences (paras. 152–156). In such exceptional situations, the decision of the asylum-seekers can be regarded as a reasonable response to the risk of being subjected to refoulement and the causal link between the unlawful act of Frontex and the occurred damages may remain unbroken (paras. 157–158). In this light, the CJEU held that the General Court erred in concluding that the appellants’ decision broke the causal link, because it was their own choice, without examining the context of that choice (para. 161).

Conflating factual and legal causation

The difficulty for adjudicating fora in addressing overdetermination – namely, situations in which multiple causes (such as the acts of multiple wrongdoers) contribute to a single damage (such as damage arising from refoulement) – has been repeatedly highlighted. Part of this difficulty lies in the conflation between the two steps of the causal inquiry, i.e. of factual and legal causation. In the specific case of W.S. and Others v Frontex, the distinction between the step of the factual and the legal causation inquiry would require the CJEU (and earlier the General Court) to assess the following. At the first step, whether Greece’s wrongful failure to provide Frontex with return decisions concerning the appellants (para. 109), and Frontex’s failure to comply with its obligation to verify the existence of return decisions (paras. 101–102) were factual causes of the damage suffered by the appellants. In other words, whether they were necessary elements of a set of conditions that was sufficient to give rise to the appellants’ damage (see: Katsoni, pp. 74-77, 350-353). Then, at the second step, the CJEU would need to perform its legal causation inquiry, i.e. to assess whether Frontex should be also held legally accountable for its contribution to the damage suffered by the refouled appellants. Or whether – despite its factual contribution to the damage through its failure to verify the existence of return decisions – it should bear no liability for some other legal reason (such as due to the appellants’ contributory fault: see paras. 150-151 of the judgment).

The CJEU’s obiter on how it thinks that a causal inquiry should be conducted reveals that the Court does not embrace this two-step approach. The Court would have complied with this two-step inquiry, if it had acknowledged that the General Court erred in law by failing to assess first which wrongful acts had been committed by Frontex and Greece, as well as whether these wrongful acts were necessary for the rise of the appellants’ damage. Only then should it have assessed whether Frontex should be held accountable for (part of) the appellants’ damage. Yet the CJEU decision to uphold the merge of the factual and the causal inquiry, and that the EU judicature is not required to examine the conditions for the rise of EU’s liability in any particular order (para. 61), is not a novelty of this particular judgment.

The practice of not distinguishing explicitly between these two steps has been noted as a characteristic of fora that wish to conceal the precise rationale behind the conclusions of their causal inquiry. In such cases, the conflation of factual and legal causation may not even affect the outcome of the inquiry. Indeed, it would be hard to believe – and certainly harder to argue – that the well-read CJEU judges could not comprehend that the causal inquiry is in fact a two-step process. By preserving the mysticism surrounding the causal inquiry, the CJEU granted the EU judicature with flexibility as to the parameters that it can take into consideration throughout this inquiry, to the detriment of transparency and foreseeability. Be that as it may, the CJEU’s judgment in W.S. allows us to deduce further hints as to the Court’s view of a correct causal inquiry in instances of overdetermined refoulement.

Selective application of a “one size fits all” approach

The CJEU underlined that for the EU to bear liability there shall exist a direct causal link between the EU’s wrongful conduct and the damage alleged (paras. 112, 148). Doing so, the Court pointed to “directness” as the appropriate test of causation in this context. Yet this test has been characterised as more difficult to be met than the rest of the causal tests developed in jurisprudence (see Lanovoy, pp. 47–54), as well as rather simplistic, given its limited capacity to capture within its scope multiple causes that contribute to a damage (ibid, pp. 53–54). The Court indicated its understanding of the meaning of a “direct causal link” descriptively, holding that such a link does not exist when the damage is a remote consequence (i.e. not a sufficiently direct one) of the wrongful conduct (paras. 149–150). At the same time, it also underlined that the direct nature of the causal link must not be understood restrictively (para. 150).

In this way, the CJEU seems to be supporting a flexible understanding of the causal test of directness, which has been also upheld by the International Court of Justice (ICJ) in its recent Advisory Opinion on climate change. There “the standard of ‘a sufficiently direct and certain causal nexus’ between an alleged wrongful action or omission and the alleged damage” was found to be “flexible enough to address the challenges arising in respect of the phenomenon of climate change” even in instances of overdetermination (para. 436). However, the CJEU also noted that for the damage to be captured within this test, the EU’s wrongful act shall be the determining cause of the damage in the sense that the damage would not have arisen in the absence of that conduct, which again points to a narrow application of the test of directness (para. 153). This remark points to a restrictive understanding of directness.

These vague remarks, which partly support a narrow perception of directness and partly a broad understanding thereof, preserve some latitude of discretion for the General Court in its assessment of whether there is a causal link between Frontex’s omission to verify the existence of return decisions and the applicant’s damage in W.S. and Others v Frontex. Although the CJEU’s decision to allow the EU judicature to remain flexible while applying the causal test of directness is not itself problematic, the Court’s selective consideration of normative factors in some – thus not all – parts of its proposal of how the causal inquiry should be done, is. These normative factors were essentially policy-based and value-laden factors that should – in the Court’s view – play a role in the determination of whether responsibility and liability should be expanded or delimited (Stoyanova, pp. 16-17).

Although the CJEU took such normative considerations into account in one aspect of its causal inquiry, i.e. during its reflection on why the appellant’s decision to travel to Erbil should not be perceived as breaking the causal chain between Frontex’s wrongful omission and the damage caused, and should, thus, not exclude the Agency’s liability, it did not consider any such parameters before opting for the causal test of directness. Instead, this choice was only made due to the Court’s reliance on the test of directness in its established jurisprudence (para. 60). As the following section will highlight, other fora have in similar cases taken into account normative considerations, while choosing a test of causation for their inquiries.

Normative considerations in other fora’s relevant causal inquiries

In its infamous judgment in Hirsi Jamaa and Others v. Italy, the Grand Chamber of the European Court of Human Rights (ECtHR) was confronted with an application concerning Italy’s involvement in the exposure of protection-seekers to the risk of arbitrary repatriation in Libya. The ECtHR took into consideration the scope and nature of the primary rule at issue (i.e. non-refoulement) and examined the foreseeable consequences of Italy’s removal of the applicants to Libya (para. 117). It held that Italy knew or should have known about the absence of asylum procedures in Libya and that, by returning protection-seekers thereto, Italy breached Article 3 of the ECHR because it exposed them to the risk of arbitrary repatriation (paras. 154–158). Similarly, in ZT v Australia, the Committee Against Torture highlighted that in instances of chain refoulement, the initially deporting State will bear responsibility if it transfers protection-seekers to an intermediate state, from where it is foreseeable that they may be deported to a third state that will subject them to the risk of torture (para. 6.4).

By relying on the causal test of foreseeability, the above fora have not only used the most appropriate causation test for cases involving overdetermination, but have also shown consideration of the importance of non-refoulement, as the primary rule that prohibits the harm inflicted upon the applicants. Adopting the causal test of directness for the assessment of whether the EU’s wrongful act contributed (alongside a Member-State’s conduct) to (the damage stemming from) a refoulement and to (the damage stemming from) a loss in the value of bonds, disregards that refoulement is a wrongful outcome that countless conventional, customary and even peremptory norms seek to prevent. Different wrongful outcomes that have different legal value require the Court’s engagement with a different test of causation. In other words, they require the Court to deviate from its standard practice of using the test of directness in all of its judgments, regardless of whether these judgments concern the EU’s involvement in a refoulement or its involvement in a loss in the value of bonds.

The need to choose different causation tests when normative considerations so require finds support in jurisprudence and in relevant scholarship. The ICJ has hinted to the need to choose the causation test that will be used depending on inter alia the harm suffered, holding that “the ‘causal nexus’ between the wrongful act and the damage in question is not static in nature, and may vary depending on the primary rule violated and the nature and extent of the injury” (Advisory Opinion on Climate Change, para. 436). Although this remark was made in the ICJ’s climate change Advisory Opinion, it reflects the ICJ’s solid view on how the causal inquiry shall be done in instances of overdetermination (see also: Armed Activities on the Territory of the Congo, para. 93). These points have also been raised in academics’ writings on the same issue (see: here, p. 59; here, pp. 60-62; here, p. 478, and here, pp. 80-81).

Concluding remarks

The CJEU’s judgment in W.S. and Others v Frontex brought to the forefront the Court’s problematic persistence to the use of the causal test of directness even in cases concerning overdetermined refoulement. Considering the CJEU’s important role as a forum where pushback survivors can seek the acknowledgment of Frontex’s wrongful involvement in their refoulement, and considering the relevant jurisprudence of other fora, which reveals a coherent consideration of normative factors during the selection of the test of causation that these fora employ in their causal inquiries, it is essential that the CJEU (and by extension also the General Court) follows a similar approach. As the blogpost highlighted, other fora that have taken into account such normative considerations in all parts of concerning overdetermined refoulement have been led to the test of foreseeability in their causal inquiries. By following this approach, the CJEU would not only align its jurisprudence with the relevant jurisprudence of other fora, but it would also contribute to a coherent jurisprudential oversight over the observance of non-refoulement in practice.