EU Immigration and Asylum Law and Policy Droit et Politique de l'Immigration et de l'Asile de l'UE

Galina Cornelisse, Professor of Courts and Transnational Justice, VU Amsterdam

European border agency Frontex’ involvement in human rights violations has received considerable attention over the last few years. In a political order based upon the rule of law, any arguable claim that individual rights have been violated should be subject to an effective judicial remedy. However, when it comes to Frontex’ operational activities, such remedies are not readily available. Case T-600/21, in which the General Court (GC) dismissed a claim for damages by a Syrian family against Frontex, did not only suffer from a faulty application of the legal criteria for establishing liability of the EU – which can and hopefully will be righted by the Court of Justice upon appeal. Much more fundamentally, it exemplifies the urgent need for the EU judicial order to provide remedies that acknowledge the fundamental changes that have occurred in the institutional set-up of the EU, exemplified as they are in the conferral of coercive powers to an EU agency in the area of border control and return.

Case T-600/21: Faulty Application of the Criteria for EU Liability

In October 2016, a Syrian family arrived on the Greek Island of Milos. They were transferred to a reception centre in Leros, where they indicated that they wanted to apply for international protection. Nonetheless, 6 days later they were deported to Turkey in a joint return operation carried out by Greece and Frontex. As the family was afraid that Turkey would send them to Syria, they travelled onwards to Iraq where they are currently still residing. The applicants bring an action for damages on the basis of Article 268 TFEU, claiming before the GC that Frontex is liable for the material and non-material damages incurred by them, because it failed to act in conformity with its human rights obligations before, during and after the joint return operation. For the legal basis of these obligations, they refer to various provisions in Regulation 2016/162  (this regulation has since been repealed by Regulation 2019/1896, but at the time of the events, it was still in force) and the EU Charter of Fundamental Rights. 

The GC reiterates the three conditions for liability of the EU for its conduct, or those of its institutions or agency’s: unlawful conduct, actual damage and a causal link between the two. As these conditions are cumulative, if one of them is lacking, it is not necessary to examine the other two: instead the action can be dismissed in its entirety. The GC therefore immediately narrows the legal question down to whether there is a causal link between the conduct that Frontex is accused of and the damages incurred. The applicants had claimed material damages consisting of the amount spent to travel to Greece; the cost of renting a house in Turkey and of buying furniture; the expenses incurred in fleeing to Iraq; the costs for rent electricity and subsistence in Iraq, as well as the children’s school fees there; and the cost of legal aid in the procedure. They had also claimed immaterial damages related to feelings of anguish, fear and suffering, such as felt by the children when they were separated and prohibited from talking during the flight and the family as a whole when they undertook the difficult and dangerous journey to Iraq because they feared that Turkey would return them to Syria. 

The GC finds no causal link between any of these damages and the conduct of Frontex. Its first finding – that the expenses linked to their initial travel to Greece cannot be seen as damages linked to the conduct of Frontex– is rather straightforward (para 61). However, in the subsequent paragraphs, the argumentation is curious and misguided: The GC confuses causality with questions of attribution and legal qualification of the conduct, and it fails to consider crucial questions of joint liability. Essentially, the GC argues that because Frontex has no competences as regards the taking of return decisions and the examining of asylum applications, there is no direct causal link between the damage suffered and the conduct of which Frontex is accused. According to the GC, as Member States are the ones taking these decisions, the conduct of which Frontex is accused of cannot be the “determining” cause for the damage that was suffered. 

The way in which the GC misapplies EU liability law here has been commented upon extensively here and here and here and here. Essentially, these posts together show clearly that the GC has confused causality, attribution and qualification of the conduct, and that it has completely ignored the possibility of joint liability of Frontex and of the Member States, as well as Frontex’ positive obligations relating to the protection of human rights. Let us not repeat their arguments here, except from adding that the causality which is at stake in liability proceedings concerns a causal relationship, not between damages and conduct per se, but between damages and unlawful conduct. 

Crucially, the fact that the causal relationship to be established concerns the relationship between unlawful conduct and damages, means that the GC should never have skipped the question which precise aspects of Frontex’ conduct are unlawful. This is so because only the “unlawful nature of the act alleged” can serve as the basis for an assessment of causality. Especially when assessing causality in a complex situation of joint conduct, it is therefore logically impossible to employ an “argumentative short-circuit”, or ignore questions of (un)lawfulness from the perspective of “judicial economy.” On the contrary, in order to assess causality between unlawful conduct and damages, it is necessary to establish, first which actor has done what (attribution) and secondly, how that particular conduct can be qualified under the law (i.e. as unlawful or lawful). Only after this has been clarified, it is possible to assess the causality between the unlawful nature of the act and damages.

Can the Court of Justice “Sort out this Mess”?

As Gareth Davies writes, it is now for the Court of Justice on appeal to sort out this mess. Under Article 256 TFEU and Article 58 of the Statute of the Court of Justice, the Court of Justice may only be asked to rule on the existence of an error of law committed by the GC. It is therefore pivotal to separate the findings of fact by the GC (which cannot be appealed) from its findings of law. In this light, it is perhaps a blessing in disguise that the GC has based its curious decision on very few findings of fact, therewith opening up the case for a full re-examination. Without a doubt, the errors which it made in applying the criteria for liability of the EU are errors of law, which can therefore be appealed and righted. However, to what extent may the Court of Justice answer the legal questions at hand with regard to the attribution and the unlawfulness of Frontex’ conduct itself? And, can it actually give a final ruling in this case? 

The answer to this depends on whether the Court, in order do so, would have to carry out a complex factual assessment or not. At least as regards establishing the liability of the EU, the factual assessment to be made in this case is quite simple. To see this, we need to zoom in on one of the claims brought forward in the proceedings by the applicants, that no return decisions had been taken in their cases. The absence of return decisions in this case is not disputed by Frontex. On the contrary, Frontex’ negligence in checking “the paperwork” has been acknowledged openly by Frontex’ current Executive Director on Dutch television. The Executive Director explained that the Syrian family was brought onboard last minute, and that Frontex acted on the statements of a Greek official instead of checking for itself that the required procedures had been followed. As such, in the appeal procedure, the Court of Justice can certainly establish the simple fact that Frontex neglected to check the existence of return decisions before carrying out the removal of the family. What are the legal implications of this simple fact? 

If Frontex has participated in the forced return of persons who were not subject to return decisions (or with regard to whom it has not itself ascertained that return decisions had been taken), the GC’s argument that Frontex cannot go into the merits of a return decision, or decide on asylum applications is an error of law: the question regarding (the attribution of) unlawful conduct in this case has nothing to do with the division of responsibilities between Frontex and Greece with regard to the taking of return decisions or the assessing of asylum claims. Instead, it revolves around the question whether Frontex may or may not participate in the forced return of persons, if it is not clearly established that return decisions have indeed been taken. Put differently, the question to be answered for establishing liability for the EU in this case is whether it is unlawful for Frontex’ to exercise its operational powers in return operations with regard to persons who are not verifiably subject to a return decision. If the answer to that question is yes, the subsequent question is whether this is a sufficiently serious breach of EU law to engage the liability of the EU. 

Certainly, to answer these two questions, Frontex’ obligations as regards the protection of human rights when it is exercising its operational powers in forcibly returning people need to be assessed. These are put down in Regulation 2016/1224, which should be read in conjunction with the relevant provisions of the EU Charter of Fundamental Rights, most notably those protecting people from collective expulsion, non-refoulement and inhuman and degrading treatment. It has been pointed out that the extent to which Frontex is under positive obligations, which in turn relies on its awareness of the risk that human rights violations could occur, are important considerations in establishing liability in this case. That is certainly correct, but a crucial argument has been overlooked in previous posts. Frontex’ operational powers in the field of return and the accompanying legal obligations that it should adhere to, should be interpreted not only in light of Regulation 2016/1624 and the Charter, but also with regard to other provisions of EU secondary law, most notably the Return Directive. Interestingly, it is precisely where the coherence of EU law is at stake, that the appeal procedure before the Court of Justice finds its rationale

The Return Directive is relevant because Frontex is the “operational arm of the common EU return system”. That common system is standardized by the Return Directive, which “sets out common standards and procedures to be applied in Member States for returning illegally staying third- country nationals, in accordance with fundamental rights as general principles of Community law as well as international law, including refugee protection and human rights obligations” (Article 1 of the Return Directive). Against this background, it makes perfectly sense that Regulation 2016/1624 mentions the Return Directive in a number of its provisions. 

• In the first place, Article 2 of the Regulation refers to Article 3 of the Return Directive for the definition of ‘return’: “the process of a third-country national going back — whether in voluntary compliance with an obligation to return, or enforced — to his or her country of origin” (or, in some situations, another country). 
• Article 2 of Regulation 2016/1624 defines a ‘return decision’ also in conformity with Article 3 of the Return Directive as “an administrative or judicial decision or act, stating or declaring the stay of a third-country national to be illegal and imposing or stating an obligation to return”.
• Article 2 of Regulation 2016/1624 continues with the definition of a ‘returnee’, as an illegally staying third-country national who is the subject of a return decision issued by a Member State.
• Lastly, Article 2 of Regulation 2016/1624 defines a return operation as an operation which is coordinated by Frontex and involves technical and operational reinforcement being provided by one or more Member States under which returnees from one or more Member States are returned either on a forced or voluntary basis.

As regards the sharing of responsibility between the Member States and Frontex, Article 5 of Regulation 2016/1624 stipulates that European integrated border management is to be implemented as a shared responsibility of the Agency and of the national authorities. One component of European integrated border management is the “return of third-country nationals who are the subject of return decisions issued by a Member State” (Article 4 Regulation 2016/1624).

Reading all these provisions together, it seems unequivocal that legally, return operations may only concern returnees, i.e. individuals who are subject to return decisions. Article 28 of Regulation 2016/1624 concerns return operations in particular, stipulating that the Agency, albeit “without going into the merits of return decisions” shall provide the necessary assistance and, at the request or with the approval of one or several participating Member States, ensure the coordination or the organisation of return operations. While the GC was keen in underlining that the Regulation thus specifies that Frontex should not go into the merits of return decisions, it forgot to pay heed to the other significant linkages between Regulation 2016/1624 and the Return Directive, not in the least Article 28, which requires clearly that the Agency’s involvement in return operations should be “in accordance” with the Return Directive. 

As a consequence, in carrying out its operational powers (i.e. assisting, organising or coordinating return operations), the Agency at the very least needs to make sure that joint return operations only concern persons who are clearly subject to a return decision. In other words, Frontex is under a clear obligation to check if it is assisting, organising or coordinating the return of actual returnees, i.e. persons who are subject to return decisions. Moreover, Articles 8 and 9 of the Return Directive provide detailed rules for the enforcement of return decisions, including its postponement if there is a risk of refoulement or on account of the physical condition or mental state of the returnee (see also Article 5). It is only logical that Frontex, when it is lending its operational support to the enforcement of return decisions by Member States, is equally bound by these rules, which is, as we saw above, duly confirmed by Article 28 of Regulation 2016/1624. In conclusion: by not checking properly if it was exercising its operational powers vis-à-vis persons with regard to whom return decisions had been taken, Frontex has committed a clear breach of EU law. 

Now, as regards the question whether this is a sufficiently serious breach of EU law to establish the liability of the EU, the reasoning can be brief. If the operational arm of the common EU return system removes persons without making sure they have received return decisions, it can literally return anyone. The requirement that persons should receive a return decision before they can be removed from the territory of the Member States, protects them against refoulement and it serves crucial due process guarantees. It leaves no room for discretion whatsoever and puts clear and unequivocal obligations on Frontex, all the more so when read against its positive obligations with regard to protecting human rights, as also provided for in Regulation 2016/1624.

As regards causality: if Frontex had checked the existence of return decisions in this case, the Syrian family should not have been returned to Turkey in this particular return operation. As such, Frontex’ unlawful conduct is the decisive cause for damages that are suffered as a consequence of this return operation. It might be true that if Frontex had acted lawfully, i.e. refused to lend its operational support to the enforcement of the return of the Syrian family, they might have been returned on another day, on a different flight, this time not in a joint operation, i.e. without Frontex’ involvement. But that is not relevant for establishing the EU’s liability in this particular case, seeing that the damage suffered by the family as a result of this particular removal is a direct consequence of the misconduct of Frontex. While it is true that Frontex’ unlawful conduct exists alongside unlawful conduct of Greece (which seems to have acted in violation of almost every provision of the Common European Asylum System and the Return Directive), that does not exclude the liability of the EU. It merely means that in an appeal, when zooming in on the causal link with particular damages claimed (which are quite a few, as we saw above), the Court of Justice is likely to engage with one of two theories of causation which can be applied where the same damage has more than one cause.

Indeed, assessing the real and actual damage as a result of the unlawful nature of the act by Frontex may not only raise intricate additional legal questions, but it will most likely also require a complex factual assessment. The latter circumstance may warrant the Court of Justice to refer the case back to the GC. It will hopefully only do so after having established the liability of the EU in its ruling, in line with case law in which the Court has given “a decision at an early stage of the proceedings on the question whether the conduct of the institutions has been such as to entail the liability of the Community, reserving consideration of questions relating to causality, as well as those concerning the nature and extent of the damage, for a later stage”. But there is more: Case T‑600/21 shows that the current action for damages in Article 268 TFEU is not an effective remedy for human rights violations committed by Frontex in the exercise of its operational powers.

Frontex’ operational powers and EU judicial remedies as “misfits of EU law”

The difficulty in litigating questions around the lawfulness of the way in which Frontex has exercised its operational powers, provides a prime example of what Renaud Dehousse has called misfits of EU law: ‘the mismatch between law and reality’ caused by the fact that ‘traditional legal categories have ceased to adequately reflect the actual operation of the EU’s machinery.’ Indeed, the EU judicial system was designed from the “traditional point of view of executive federalism.” As a consequence, its legal remedies reflect the assumption that much of EU law is implemented or enforced by Member States, and “only where the latter encounter constitutional limits that affect the uniform application of European law will it centralize the enforcement of European law”. In this constellation, national courts work as decentralised EU law courts: if a person thinks that a Member State (or a private party) violates their EU law rights, they can appeal to these rights before a domestic court. The combination of the doctrines of direct effect and supremacy together with the genius of the preliminary ruling procedure ensures European integration and decentralised enforcement in one-go. For integration, it relies on individuals who have an interest in claiming their rights; for enforcement in conformity with higher EU rules it relies on domestic courts. 

Now, this decentralised system of protection of rights that relies on national courts to protect the EU rights of the individual against faulty or incomplete implementation of EU law by the state, doesn’t work when it is not the Member States (or private parties) that act contrary to EU law, but the EU itself. In such a case, EU law offers two principal direct remedies before the EU court system: annulment and compensation. However, the physiognomy of these two remedies is shaped by the assumption that powers of enforcement through coercion still lie at the level of the Member States. That makes holding Frontex accountable for human rights violations before the EU courts difficult.

In an action for annulment on the basis of 263 TFEU, the Court of Justice may review the legality of acts of bodies, offices, or agencies of the Union intended to produce legal effects vis-à-vis third parties. It has commonly been assumed that much of Frontex’ activities, such as organising or coordinating a return organisation, consist of purely factual conduct, and can therefore not be qualified as acts amenable to review by the Court of Justice. Similar considerations apply to the corollary remedy: the action for failure to act in Article 265 TFEU. The formal approach taken by the GC to the qualification of the act to be reviewed is apparent from Case T-282/21, also against Frontex, decided last year.

A number of NGO’s, acting on behalf of two asylum seekers, had asked the GC to determine that Frontex had acted in violation of EU law by failing to terminate its operations in the Aegean Sea despite serious, systematic, and widespread violations of fundamental rights related to the Agency’s activities. Before bringing the action, they had first asked Frontex to act (i.e. terminate the operation), in conformity with the second paragraph of Article 265 TFEU. Frontex had refused to do so, sending the applicants a letter in which it explained that ‘isolated incidents’ could not be a reason for a decision to terminate an operation. The GC dismissed the action, ruling that the applicants should have brought an action for annulment (of the position taken by Frontex in its letter to the applicants) under Article 263 TFEU, not an action for failure to act.

The ruling by the GC is not illogical: according to the consistent case-law of the Court, “any measure the legal effects of which are binding on, and capable of affecting the interests of, the applicant by bringing about a distinct change in his legal position is an act or decision which may be the subject [of an action for annulment]”. Especially as Frontex had clarified its position in response to the invitation to act by the applicants, an action of annulment could have been brought against this position. However, this formal approach disregards the fact that the alleged violations of EU law were not committed by the letter setting out the position of Frontex, but consisted in Frontex not terminating the operation despite alleged human rights violations. And indeed, many human rights violations, and certainly those committed in the exercise of operational powers, “do not occur in the form of a neatly stamped piece of paper.” As such, the action of annulment, as interpreted by the Court of Justice so far, is not capable of providing an effective remedy for these kinds of human rights violations. 

The action for damages, although it is capable of addressing factual conduct, is not an adequate remedy either. The object and purpose of the action for damages lies primarily in the award of compensation for unlawful conduct by the EU. There are two reasons why this is problematic from the perspective of effective remedies for violations of human rights. In the first place, a human rights claim does not equal a claim for financial compensation. On the contrary, in the ECHR, Article 41 on just satisfaction, complements the ECtHR’s jurisdiction to rule on the existence of a violation of the ECHR. When it comes to violations of human rights, especially those that concern the coercive power of the state and thus affect the life and limb of persons, financial compensation only constitutes a small part of the effectiveness of a remedy. Interestingly, the inadequacy of financial compensation on its own is recognised in Articles 34-37 of the ILC Articles on State Responsibility, which prioritise restitution over compensation (in the case under discussion in this post, restitution would entail bringing the applicants back to Greece!) and stipulate that if restitution or compensation is not possible, just satisfaction can be provided in the form of an apology or expression of regret.

From a more pragmatic view, the action for damages is not effective because it is characterised by high standards of proofs for litigants to determine the liability of the EU. Indeed, a special human rights court such as the ECtHR has under certain circumstances placed the burden of proof on Respondent States, once an applicant makes an arguable claim of ill-treatment by the authorities, or when the state is in possession of all the evidence. Against this background it makes no sense to deny the admissibility of evidence consisting of documents that were issued by Frontex itself (and other EU agencies),  as the GC did in its ruling. Moreover, placing the burden of proof for a causal relationship between damages and unlawful conduct in complex cases of joint liability on a migrant who has been forcefully expelled from the territory of the EU is tantamount to denying this person access to justice. In this light it is worth underlining that the ECtHR, once it has found a violation of the ECHR, does not require any proof of the non-pecuniary damage sustained.

Rethinking EU Remedies: Article 47 Charter and Holding the Court to its Own Standards

As current EU remedies in case of alleged violations of human rights by Frontex are dependent on litigants having to prove damages, or on the question whether an act has legal effects or not, these can hardly be called effective. In the absence of legislative action acknowledging the fundamental shift that has occurred in the EU system with the partial transfer of coercive powers from the Member States to an EU agency, it is possible, even necessary, for the EU courts themselves to update the existing remedies in Articles 263 and 268 TFEU when adjudicating Frontex’ operational powers. They could do so by applying mutatis mutandis to these procedures the Court’s own case law on domestic judicial procedures and the requirements of Article 47 of the Charter.

In the action for damages, the Court of Justice should (just as the ECtHR) apply a presumption of non-material damage when it has established that the EU has committed a violation of human rights. It should also lower the burden of proof for litigants, especially as regards causality in cases of joint conduct. Moreover, one can draw upon its own case law to argue that it should endow itself with an ex officio power to take into account all the necessary evidence when an arguable claim is put forward that core rights such as liberty, non-refoulment or the prohibition on torture are violated. 

As regards the action for annulment, it must be updated so as to be able to address purely factual conduct as well. The Court of Justice can do so by qualifying some of the human rights violations that Frontex has been involved in as bringing about a “distinct change in a person’s legal position”. When it concerns unlawful removals from the EU’s territory, this is not stretching the definition of a legal act at all: territorial exclusion is the prime manifestation of legal exclusion – the latter providing the rationale for the former. With regard to other human rights violations, such as push-backs on the high seas and other instances of the use of violence, it may be trickier to argue that factual conduct always brings about a change in the legal position of individuals. However, here again one can draw upon the Courts own case law to make a convincing argument that EU legislation “which does not guarantee any judicial review of the lawfulness of [such factual conduct] undermines the essential content of the right to effective judicial protection, guaranteed in Article 47 of the Charter, in that it absolutely prevents the Court from ruling on respect for the rights and freedoms guaranteed by EU law.” Accordingly, the Court, if it considers that applicants in an action for annulment have put forward an arguable claim of human rights violations committed in the exercise of operational powers by Frontex, should declare that it has “jurisdiction to examine the lawfulness of such [conduct], disapplying, where necessary, any [EU law] provision which prohibits it from proceeding in that way.”

Only by taking its own case law on Article 47 Charter seriously, can the Court of Justice thus align its remedies with a reality in which some elements of the monopoly over the legitimate means of violence have been transferred from the Member States to an EU agency. If not, the specific characteristics and autonomy of EU law cannot serve as a justification for delaying accession to the ECHR any longer, seeing that current judicial remedies do not do justice with the aspired constitutional structure of the EU which should guarantee that “the interpretation of […] fundamental rights be ensured within the framework of the structure and objectives of the EU.”