Who Reviews the Conditions of Union Power? The Kövesi Litigation and a Blind Spot in the Judicial Review of Hybrid EU Governance

 

Joanna Demopoulou holds a PhD in International Affairs and is a former Executive in Residence at the Geneva Centre for Security Policy (GCSP). Her research examines the legal and institutional architecture of public authority and governance, focusing on how law structures, distributes, and constrains public authority across European and international institutions.

 

Photo: Laura Kövesi, by AGERPRES, cropped by Ionutzmovie, via Wikimedia Commons

 

EU law can examine almost any exercise of public power. Where it struggles is with the decision, taken a step earlier, about whether that power may exist at all. The Kövesi litigation is where the gap becomes visible.

 

On 24 June 2026 the Administrative Plenum of the Greek Court of Cassation dismissed, as inadmissible, an application brought by the European Chief Prosecutor, Laura Kövesi. She had challenged a Greek decision renewing three European Delegated Prosecutors (EDPs) for two years, where the College of the European Public Prosecutor's Office had wanted five. The vote was seventy-two to ten. Reported as news, the case looks like a routine standing problem: a supranational official has no locus, under domestic law, to contest a Member State's handling of its own judicial officers, and national procedure ran its course. That reading is not wrong. It is just incomplete.

 

What the dismissal actually protects is worth stating plainly. European prosecutorial independence was shielded here not according to what it does, but according to who signed the act that curtailed it. Nobody designed things to work this way, which is rather the trouble.

 

The case in brief

 

The College favoured a five-year renewal for three prosecutors in the Greek national section. The Supreme Judicial Council, the organ of national judicial self-government, granted two. Kövesi sought annulment. The Plenum found she lacked standing. Under Greek law on the service status of judicial officers, recourse lies only with the affected officers themselves, and only where the underlying Council decision carried a dissent of at least two members. This renewal had been unanimous.

 

The consequence is what matters. The problem was not that the wrong applicant had turned up. On these facts, no applicant could have brought a challenge at all. A minority of ten judges would have referred the underlying question, who fixes the duration of an EDP's mandate, to Luxembourg; the majority made no reference. And the forum matters: this was the Administrative Plenum in closed session, the Court acting in its self-governing rather than its adjudicative capacity. A national body, applying national rules, set a term that would bound the independence of a Union prosecutor.

 

EU law reviews the exercise of power well

 

The EPPO Regulation shows how far Union law has come in policing how European Union power is used. Operational acts, meaning the procedural steps of an investigation that can affect third parties, are reviewable under Article 42(1). Administrative acts of the Office fall to the Court of Justice under Article 42(8). And that Court has read these provisions generously. In EPPO v I.R.O. and F.J.L.R. (C-292/23), decided by the Grand Chamber on 8 April 2025 on a reference from a Spanish court, it insisted on effective review even of a witness summons issued by a European Delegated Prosecutor, looking past the formal label of the act to its effect on a legal position. The Regulation itself reveals an asymmetry: Article 42(8) expressly provides review for decisions dismissing EDPs under Article 17(3), but says nothing about national decisions shortening the office on which the Union mandate depends. Where European Union power is exercised, in short, a court can generally be found to look at it.

 

The Greek renewal decision is not that kind of act. It exercises no Union power and administers no Union office. What it does is change the duration of the national judicial status a person must hold in order to serve as an EDP at all (Recitals 32–33). Article 17, which entrusts appointment and dismissal to the College, treats that status as a standing precondition of the Union mandate, not as the thing the mandate acts upon. The act does not use the prosecutor's independence and it does not formally end it; it determines whether that independence can go on being held at all. What is being decided, in other words, is a precondition of the Union mandate rather than anything the mandate does, and preconditions of that sort are where EU judicial review is thinnest.

 

Why it falls between the regimes

 

The Regulation's review architecture sorts acts along two axes: institutional origin (an act of the EPPO, or of a Member State) and character (operational or administrative). A decision of the kind at issue here fits none of the resulting boxes. Because it is national in origin, the national track applies and Article 42 does not. Because it is neither operational nor administrative in the Regulation's sense, no Union regime claims it either. It is left to ordinary national service-status litigation.

And in Greece that litigation led nowhere. The unanimity bar, anchored in Article 90(3) of the Greek Constitution (whose compatibility with the right to effective judicial protection is itself already contested), meant that no review was available to anyone. The independence of a Union prosecutor was capped by a national hand, and no court, Greek or European, was in a position to examine the cap.

 

The obvious objection

 

The natural response is that this is simply a national act, and that dressing it up as anything else misreads the bargain on which the EPPO was built. The Office was designed to stay embedded in national structures. Article 42 was written for the Office's operational output, not for Member States' decisions about their own judges.

 

Take that objection at its strongest and it still runs into the facts. To succeed, it has to treat the decision's effect on a Union status as carrying no independent legal weight, the two-year cap as, in Union terms, simply invisible. But the College's five-year act and the Council's two-year act, pulling in opposite directions over the same office, put exactly that in doubt. If national tenure can quietly cap the Union mandate, a formally national act is shaping a condition of Union authority with no Union-level review; if it cannot, the act was never purely national in effect. Either way the binary the architecture relies on does not hold, and incorporation of the national status into the Union scheme, even if one grants it, says nothing about who may review a decision that shortens the incorporated condition.

 

The argument does not turn on how Kövesi ends

 

It is worth being precise about what this is not. There is a substantial literature on judicial review in composite or integrated procedures, where national and Union authorities each contribute to a single act and review fractures because each court sees only its own segment. The Greek renewal is a different animal: not a national input into a composite Union act, but a self-standing national decision that governs a precondition the Union scheme takes as given. The composite-procedures problem is who reviews which part of a joint act. The problem here is that the review architecture offers no settled framework for a decision about whether the Union mandate can subsist at all.

 

None of this denies that EU law has built review doctrines in adjacent terrain. Article 19(1) TEU, in the line running from the Portuguese Judges case, obliges Member States to uphold the independence of national courts that may rule on Union law. But that doctrine asks whether independence is protected; it does not supply a route for reviewing a national decision that fixes the tenure on which a Union mandate depends. The guarantee runs to independence as a value, not to the reviewability of the act that conditions the office.

 

Suppose the Court of Justice were eventually to hold, here or in a later case, that a renewal of this kind is purely national and beyond Union reach. That would not defeat the point; it would illustrate it. The claim is not that such decisions must become Union acts and be reviewed in Luxembourg. It is narrower and more awkward: EU law currently reads such decisions through the origin of the act, so that their bearing on Union power does no independent work. A ruling that an act of this kind is 'purely national' would be the clearest demonstration of that reflex. The thesis is about the criterion of review, not the disposition of this dispute, which is why it survives whichever way the dispute is resolved.

 

Beyond the EPPO

 

European Union constitutional law has grown confident at reviewing how public power is exercised. Whether it has built anything equally coherent for reviewing the legal conditions on which that power rests is far less clear, and those conditions multiply as Union governance turns hybrid, its authorities European but their underpinnings national.

 

So Kövesi is not really a story about three prosecutors in Athens. It is the point at which the standard question of judicial review, who exercised this power, starts to miss something. EU law has developed an answer to the first question. It has not yet developed one to the second, who set the conditions under which the power could be exercised at all.

 

*A note on sources. At the time of writing the full reasoning of the Administrative Plenum was not publicly available. This post relies on the EPPO Regulation (Council Regulation (EU) 2017/1939) and CJEU case law as primary materials, and on contemporaneous reporting for the Greek proceedings. The characterisation of the majority and minority positions should be checked against the official text once it is released.

 

Advocate General Ćapeta’s Opinion in WS and Others v Frontex before the Grand Chamber: The End of Frontex’s Shielding? Joint Liability of Frontex and Member States in Return Operations

 

Antje Kunst*

* Antje Kunst is an international lawyer and barrister of Garden Court North Chambers, admitted to the Bar of England and Wales, and the Bar of Berlin, advising and representing individuals in a wide range of matters related to fundamental rights within the CFSP and other fields. She has appeared in numerous cases before both the Court of Justice and the General Court, within the Court of Justice of the European Union.  

Photo credit: Asurnipal, via Wikimedia Commons

 

Introduction

On 12 June 2025 Advocate General (AG) Tamara Ćapeta, following a hearing by the CJEU Grand Chamber in February 2025, rendered an important opinion (‘Opinion’) on the appeal of WS and Others v. Frontex,  in Case C-679/23 P against the General Court’s judgment of 6 September 2023.  The General Court’s judgment has been rightly heavily criticised (see for example here, and here). Two authors went as far as stating that the shielding Frontex from meaningful responsibility is not only harmful to victims of fundamental righta breaches committed by Frontex, but that it also undermines the institutional framework of EU agencies and affects the legal credibility of the Court (see here).

The clarity that was lacking in the General Court’s judgment is made up for by the AG, who addressed the issues directly and even offered two possible readings of what the General Court may have intended to find, including that it might have failed to attribute the unlawful conduct to Frontex. (see paras. 60-67 and 84 of the Opinion)

As outlined  here WS and Others v. Frontex involves a Syrian refugee family with four children, who are ethnic Kurds. They arrived on the Greek island of Milos in October 2016 by boat with the intention of claiming asylum. Even though they formally declared their interest in applying for international protection, just six days later, they were deported by flight to Turkey in a joint return operation conducted by Greece and Frontex. Because of the fear of being returned to Syria by the Turkish authorities (para. 59 of the GC’s judgment), the family decided to flee from Turkey to Iraq, where they reside to this day.

The applicants filed a claim for compensation, alleging that Frontex had violated its own obligations under the then-applicable Frontex Regulation (the 2016 Frontex Regulation) and several fundamental rights under the EU Charter during the return operation.

The significance of the case lies in the fact that, for the first time, the Court of Justice examines the joint liability of Frontex and Member States for breaches of fundamental rights arising from joint return operations. There is also an opportunity for the Court to develop its case-law on the causal link between the alleged unlawful conduct of an EU agency -specifically Frontex- and the damage suffered by individuals. This blog post focusses on these two issues.

 

Admissibility of the action

Any credible claim that fundamental rights have been violated deserves access to an effective judicial remedy. In cases like this, that remedy is a compensation claim under Article 340(2) TFEU to make good any damage caused by EU institutions, bodies and agencies.

Despite Frontex’s objections on four grounds, the AG rightly concluded that the claim is admissible (see paras. 42- 55 of the Opinion). This includes the General Court’s failure to assess relevant facts like the existence of a return decision which are arguments of points of law. If the Court of Justice agrees, which is expected, it would set an important precedent, opening the door for similar actions against Frontex to be heard in the future.

 

Frontex has its own obligation to verify under the 2016 Frontex Regulation

The AG makes it clear: Frontex has its own independent obligations under EU law when it takes part in joint return operations. That includes a duty to verify whether a valid return decision exists for each person put on the plane (paras. 68–83 of the Opinion).

Crucially, she stresses that Frontex -like all EU bodies- must uphold fundamental rights while carrying out its tasks. It is a legal obligation under Article 51(1) of the EU Charter of Fundamental Rights. So, when Article 34(1) of 2016 Frontex Regulation sets out Frontex’s duty, it is simply restating what EU law already requires (para. 72 of the Opinion).

The AG takes a close look at the 2016 Frontex Regulation and shows that Frontex’s responsibilities during return operations are clearly spelled out therein.

First, Article 28 of the Regulation ties Frontex’s operations directly to the 2008 EU Return Directive, which strictly prohibits returning individuals without an enforceable return decision. Second, Article 4(h) of the Regulation sets out the same point. This is supported by consistent case law from the Court of Justice. The idea is simple but crucial: the existence of return decision helps ensure that the return does not violate the principle of non-refoulement.

The AG makes an important -if obvious- point: the 2016 Frontex Regulation clearly intends that joint return operations coordinated or organised by Frontex must only involve individuals who are the subject of individual, enforceable return decisions.

She goes on to conclude that Frontex bears its own responsibility in this process- again an obvious point despite Frontex's consistent denials. By verifying that return decisions exist for everyone on board, Frontex is fulfilling its legal duty to ensure respect for fundamental rights, including the principle of non-refoulement, in the performance of its tasks, as required by Article 34(1) of the 2016 Frontex Regulation.

Consequently, the General Court erred in law that there could not be a causal link between the allegedly unlawful conduct of Frontex and the damage suffered ( see para. 83 of the Opinion)

 

Joint and several liability of Frontex and the host Member State?

In paragraphs 84–97 of the Opinion, the AG observes that the General Court may have erred in finding that Frontex and a Member State could not be held jointly and severally liable for the same damage. She clarifies the concept of joint and several liability: when more than one actor contributes to a single harm, each may be held fully liable, and the injured party has the right to claim compensation from any one of them for the entire damage (see para. 85 of the Opinion). The issue of shared liability is a much-debated topic in legal scholarship (e.g., here). Can both Frontex and a Member State be liable for separate breaches of EU law that give rise to a single harm? According to the AG, the answer is yes.

She goes on to clarify that under the EU’s integrated border management system, Member States can be held liable for any action or omission. Frontex, on the other hand, is accountable only for actions or failures that fall within the scope of its assigned tasks under EU law (para. 88 of the Opinion).

She offers an important clarification regarding the Court’s earlier ruling in Kočner v Europol (para. 90 of the Opinion), which concerns joint liability of an EU agency and Member States: that case did not establish a general principle that joint and several liability must be expressly provided for in EU legislation, as it was in Kočner.

In other words, joint and several liability can apply even when not explicitly laid down in legislation. At the same time, she distinguishes Kočner from the present case. In Kočner, it was not possible to attribute the unlawful conduct- namely, the data processing, to Europol or a Member State. By contrast, in the current case, she explains, it is possible to attribute the same omission - failing to ensure a valid return decision - to both Frontex and Greece (para. 91 of the Opinion).

She concludes correctly had either Frontex or the Member State conducted the necessary verification the Syrian family would have been excluded from the joint return operation, assuming there was no return decision. Consequently, responsibility for the damage in this case can be attributed to both, since either one had the capacity to prevent it.

 

Member State’s primary responsibility does not exclude Frontex’s liability

The AG rejected Frontex’s argument that the agency cannot be held liable simply because Member States bear primary responsibility for return procedures under the 2016 Regulation. She emphasizes that this cannot mean Frontex is absolved of liability for failures -here the omission to verify whether a valid return decision existed. She rightly pointed out that accepting such reasoning would effectively shield Frontex from accountability for unlawful actions or omissions during joint return operations. (para. 92 of the Opinion)

Ćapeta reaches this important conclusion: in situations where both Frontex and Member States share obligations in joint return operations, Frontex can be held liable for damage resulting from a breach of those obligations, even if a Member State may also be held liable for the same damage (para. 93 of the Opinion) This position should be adopted by the Court of Justice; otherwise, Frontex will continue to hide behind a Member State to evade liability.

 

Causal link - broken through own decisions?

The AG disagrees with the General Court’s conclusion that the harm suffered by the appellants -their forced relocation to Türkiye, flight to and eventual settlement in Iraq- was the result of their own choices breaking the causal chain, and not any action or omission by Frontex. She finds this correctly to be an error of law.

Regarding factual causality, the AG refers to the ‘but for’ test and to the General Court’s finding that the alleged unlawful conduct being a conditio sine qua non for the damage to arise is not sufficient to establish a causal link.  

The AG is of the view that the ‘but for test’ is fulfilled in the present case on the assumption that there was no return decision: had Frontex fulfilled its duty to verify whether a valid return decision existed, it might have discovered that none was in place. The appellants likely would not have been included in the joint return operation and would not have suffered the alleged damage. (see paras. 103 and 106 of the Opinion).  In this context the AG notes that, from an early stage, the Court’s case law has imposed limits on the ‘but for’ test when assessing liability, particularly by excluding the liability of EU institutions or bodies for damage deemed too remote (so-called legal causality) (para. 109 of the Opinion).

As set out here, in the present case, however, the damage is not too remote: Frontex’s operation constitutes the most proximate and direct cause of the harm suffered by the family.

As the AG explains, a finding of damage of being too remote often comes into play when others, sometimes even the injured party, have contributed to the harm (paras. 110-112 of the Opinion). In such cases, the causal chain may be considered broken, and liability may not attach. The AG refers to past case law where the injured party’s own actions or decisions were seen as breaking the chain of causation, predominantly in market cases where the injured parties were economic operators. But in this case this reasoning must not apply (paras. 114 to 116 of the Opinion).

A Syrian Kurdish family - two parents and their four children - fled Syria at the height of the war. What followed was a serious breach of their fundamental rights, most notably the principle of non-refoulement. As the AG points out, the family was undeniably in a position of vulnerability.

After being returned to Turkey, they eventually made their way to Iraq - a place they considered safer. While that move may have technically been their own decision, it cannot be called a ‘free choice’. As AG Ćapeta puts it, it was a ‘difficult life-choice decision’.

On that basis, the Court of Justice should follow Ćapeta’s reasoning and find that the factual chain of causation was not broken by the appellants’ actions. Without the unlawful omission by Frontex, they would never have been forced to make such a difficult decision in the first place.

 

Conclusion: The Shielding of Frontex Should End

The AG’s Opinion should mark the beginning of the end of Frontex’s shielding from accountability in EU joint return operations. It is now up to the Court of Justice to carry this principled stance forward. Based on the questions raised by the judges during the oral hearing, there is a strong indication that the Court will overturn the General Court’s judgment - though perhaps not entirely.

The Court may well agree with the AG that a final ruling is not yet possible and that the case should be referred back to the General Court for a fuller factual assessment. Nonetheless, a clear and thorough judgment will be essential - not only to guide the outcome of this case, but also to provide legal clarity for similar future cases. What emerges already from the facts presented and the Court’s existing case law is this: the EU judiciary must eventually take the necessary step of establishing Frontex’s non-contractual liability for the unlawful deportation of the family and award appropriate damages.

Pyrrhic victory for the Greek government: the CJEU rules on Turkey as a “safe third country”

Professor Steve Peers, Royal Holloway University of London

Photo credit: Mstyslav Chernov, via Wikimedia Commons

Hamster idiom credit: Adam Sharp

(Thanks to Zoe Gardner for discussion of the case; the following analysis is my own)

 

Introduction

Remember the ‘refugee crisis’ of 2015-16? One part of the EU’s response to it was to negotiate a ‘joint statement’ with Turkey (see comments and criticism here, here and here) that, among other things, led to Turkey agreeing to take back asylum-seekers from the Greek islands. Greece facilitated the application of this process by determining that Turkey was a ‘safe third country’ (ie that asylum-seekers could and should have applied there instead, given that this other country is ‘safe’) under EU asylum law, initially for Syrian asylum-seekers in the eastern Greek islands. Faced with a legal challenge to this determination, the Greek Council of State ruled in 2017 that the determination was legal, deciding by a 13-12 vote not to ask the CJEU to interpret the issue of EU law at stake. For its part the EU General Court (upheld on appeal by the CJEU) rejected legal challenges to the joint statement, on the frankly bizarre ground that the European Council had nothing to do with the EU/Turkey joint statement (which referred to EU law issues such as visa waivers).

Subsequently in 2020, Turkey decided to refuse all readmission of asylum-seekers from Greece, stating that (as quoted in note 10 of the Advocate-General’s opinion):

 Responding to repeated requests from the Greek authorities and the European Commission regarding the resumption of return operations, [the Republic of Türkiye] has stated that no return operation would take place unless the alleged pushbacks along the Turkish-Greek border stop and [the Hellenic Republic] revokes its decision to consider [the Republic of Türkiye] a Safe Third Country.

Greece nevertheless decided to list Turkey as a ‘safe third country’ more broadly in 2021, for asylum-seekers from Syria, Afghanistan, Pakistan, Bangladesh or Somalia; it added Albania and North Macedonia to its ‘safe third country’ list soon afterwards. In light of the EU law rules on ‘safe third country’ (discussed further below), two NGOs (the Greek Council for Refugees and Refugee Support in the Aegean) decided to challenge the Greek government decision to list Turkey, resulting this time in the view of the Council of State in February 2023 (by an 18-4 vote) that the listing of Turkey was invalid, on the grounds that EU law precluded designating a country which refused to readmit asylum-seekers as a ‘safe third country’. (Other EU law arguments against the designation of Turkey were rejected by the national court and not referred to the CJEU: see the analysis by Minos Mouzourakis).  But unlike in 2017, there was great deference to the view of the (much smaller) minority (two judges apiece argued that the issue of readmission was relevant instead when ruling applications inadmissible, or when enforcing that decision). So the national court decided to ask the CJEU questions about the interpretation of EU law on the readmission point before giving its final ruling. The CJEU delivered its judgment last Friday.

This is not only the first time that the CJEU has been asked about an aspect of the EU/Turkey asylum relationship. Remarkably, despite the huge number of asylum seekers that have come to Greece over the last decade and more, this is also the first time that the CJEU has been asked any questions about EU asylum law by any Greek court.

 

EU legal framework

Legislation

The definition of ‘safe third country’ for the EU is currently set out in Article 38 of the EU’s asylum procedures Directive. (This is about the ‘safety’ of non-EU countries, ie it is distinct from the EU Member States regarding each other as ‘safe’ countries, under the Dublin rules) First, Article 38(1) provides that the (optional) ‘safe third country’ concept can ‘only’ be applied if Member States are satisfied that certain ‘principles’ are ensured for asylum-seekers in that country: (a) ‘life and liberty are not threatened on account of race, religion, nationality, membership of a particular social group or political opinion’; (b) ‘there is no risk of serious harm as defined in’ the EU qualification Directive; (c) ‘the principle of non-refoulement in accordance with the Geneva [Refugee] Convention is respected’; (d)  ‘prohibition of removal, in violation of the right to freedom from torture and cruel, inhuman or degrading treatment as laid down in international law, is respected’; and (e) ‘the possibility exists to request refugee status and, if found to be a refugee, to receive protection in accordance with the Geneva Convention’.

Secondly, Article 38(2) requires Member States to subject the application of the ‘safe third country’ rule to national law rules, including: (a) ‘requiring a connection between the applicant and the third country concerned on the basis of which it would be reasonable for that person to go to that country’; (b) ‘rules on the methodology by which the competent authorities satisfy themselves that the safe third country concept may be applied to a particular country or to a particular applicant. Such methodology shall include case-by-case consideration of the safety of the country for a particular applicant and/or national designation of countries considered to be generally safe’; and (c) ‘rules in accordance with international law, allowing an individual examination of whether the third country concerned is safe for a particular applicant which, as a minimum, shall permit the applicant to challenge the application of the safe third country concept on the grounds that the third country is not safe in his or her particular circumstances’. Also, an asylum-seeker must ‘be allowed to challenge the existence of a connection between him or her and the third country’, as referred to in point (a).

Next, Article 38(3) requires Member States to inform the applicant that the principle is being applied to them, and also to give the asylum-seeker a document informing the supposedly ‘safe’ third country that the merits of their asylum claim have not been examined.

Crucially for this case, Article 38(4) then provides that:

Where the third country does not permit the applicant to enter its territory, Member States shall ensure that access to a procedure is given in accordance with the basic principles and guarantees described in Chapter II. [ie an ‘ordinary’ examination of the merits of an asylum claim]

Finally, Article 38(5) obliges Member States to inform the Commission of their use of the concept.

The practical relevance of applying the ‘safe third country’ notion is (among other things) that it is a ground of (optional) inadmissibility under Article 33 of the Directive. And if the case is inadmissible, the asylum application is not further considered on the merits and has failed in that Member State (subject to an appeal of the inadmissibility decision, or the limited prospect of a ‘repeat application’, discussed further below). The failed asylum-seeker can then deprived of support as an asylum seeker under the reception conditions Directive and detained and removed under the Returns Directive. Of course, the impact of this is in theory ameliorated in ‘safe third country’ cases because the supposedly ‘safe’ country will readmit the person concerned and consider their asylum application properly. In practice, as we have seen, Turkey was explicitly and unambiguously refusing to readmit asylum-seekers from Greece in the context of the ‘safe third country’ concept.

Case law

The CJEU had interpreted the ‘safe third country’ rule in some prior case law, all concerning Hungary. First of all, in Case C‑564/18 (paras 29-51 of the judgment), the CJEU confirmed that the grounds of inadmissibility listed in the Directive were exhaustive (‘only’), and ruled that a national law applying a ‘safe third country’ principle to non-EU States where the asylum-seeker ‘as not exposed to persecution or a risk of serious harm, or in which a sufficient degree of protection is guaranteed’ was not consistent with the requirements of Article 38(1) of the Directive, in particular because it omitted the ‘non-refoulement’ requirement. Moreover, mere transit through the non-EU country was not enough for there to be a ‘connection’ with that country as required by Article 38(2).

Next, the Court reiterated these points in its judgment in Joined Cases C-924/19 and C-925/19 PPU (paras 148-165); but questions about what happens when the non-EU State refuses to readmit the asylum-seekers – the issue since addressed in last week’s judgment – were ruled inadmissible (paras 166-174). However, the Advocate-General’s opinion in that case discusses the issue (paras 107-127), taking the view that the refusal of readmission is only relevant at the point of enforcement of a decision based on the principle.  

Finally, in Commission v Hungary (Case C‑821/19), along with ruling on the criminalisation of assistance to asylum seekers, the Court again reiterated that the Hungarian law breached Article 38(1) and (2) of the Directive. The existence of a (then) pending proposal to amend the Directive on this point (since adopted as part of the 2024 asylum pact – but without some of the changes originally proposed) was irrelevant in determining whether a Member State was in breach of EU law at the time of the alleged breach.

 

Greek practice

As noted above, Greece had listed Turkey as a ‘safe third country’ despite Turkey’s refusal to readmit anyone in that context. While the CJEU judgment says nothing about the practice of the authorities, and the Advocate-General’s opinion says little (paras 65-68 of the opinion), one can find rather more information at the bottom of the ivory tower, in the recent Pro-Asyl report on the ‘safe third country’ principle in Greece, and in the annual AIDA report on Greece (‘safe third country’ section).

The reports state that over 10,000 asylum claims have been found inadmissible on the basis of the Greek listing of Turkey as a ‘safe third country’, although around 12,000 asylum claims have been found admissible (in some cases because the asylum-seekers argued that Turkey had not allowed them to make an asylum claim, or due to past exposure to refoulement, ie return to an unsafe country, or because they lacked a sufficient ‘connection’ with Turkey, on the basis of the national law implementing Article 38(2) of the Directive). Also, the reports are highly critical of the conclusion that Turkey is ‘safe’ – although the recent CJEU judgment does not address this issue, because the national court did not ask the CJEU about it.

Like the Advocate-General’s opinion, the Pro-Asyl study reports that Greek authorities apply the readmission test at the point of enforcement of a ‘safe third country’ decision. When Turkey then refuses readmission, the Greek bodies refuse to withdraw their prior decisions, leaving asylum-seekers to make a ‘repeat application’. But there are delays in registering these applications, and the authorities often reject them because they do not believe that the refusal of readmission is a ‘new element or finding’ (as required for a repeat application to be admissible under the EU Directive). But at least some courts rule against detention in such cases, given that there is no reasonable prospect of return (as required by the Returns Directive).

 

The judgment

The CJEU’s judgment began by answering the first question referred by the national court, rephrasing it as asking ‘in essence, whether’ the ‘safe third country’ rule in the Directive, ‘read in the light of Article 18 of the Charter’ on the right to asylum, ‘must be interpreted as precluding legislation of a Member State classifying a third country as generally safe for certain categories of applicants for international protection where, despite the legal obligation to which it is subject, that third country has generally suspended the admission or readmission of those applicants to its territory and there is no foreseeable prospect of a change in that position’.

First, the Court reiterated its case law that the grounds for inadmissibility of asylum applications in the procedures Directive (including ‘safe third country’ as defined in the Directive) are ‘exhaustive’, ie the Member States cannot add more grounds of their own (such as a divergent national definition of ‘safe third country’). It pointed out that the EU law definition ‘is subject to compliance with the cumulative conditions laid down’ in Article 38 of the Directive, as summarised above. But as the national court had already ruled that Greek law complied with the basic principles in Article 38(1), the CJEU did not examine that point. But it summarised the requirements for national law set out in Article 38(2) and the information obligations in Article 38(3), then recalled the Article 38(4) rule that ‘where the third country concerned does not permit the applicant for international protection to enter its territory, the Member States are to ensure that access to a procedure is given in accordance with the basic principles and guarantees described in Chapter II of that directive’.

In the CJEU’s view, it followed from this first from the wording of the Directive that a Member State can designate a country as ‘generally safe’ by ‘an act of general application’. The references to national law in Article 38(2) referred to such a designation, as did the preamble to the Directive. Also, the Directive did not make the validity of such a general act ‘subject to the condition that it be proved that the applicants for international protection concerned will actually be admitted or readmitted to the territory of that third country’. This was not among the national law rules in Article 38(2); and the wording of Article 38(4) implied that the designation of the third country as ‘safe’ preceded the consideration of the issue of whether that country actually readmitted asylum seekers.

The CJEU also believed that this followed from the context of the Directive, in particular by comparison with the wording of the rules on ‘first country of asylum’ (ie where the asylum seeker already had protection, as distinct from where they could have sought it) on the readmission issue.

Finally, in the CJEU’s view, this interpretation also did not conflict with the objective of the Directive, as set out in the preamble, that in the interests of asylum-seekers in particular, asylum applications ‘should be made as soon as possible, without prejudice to an adequate and complete examination being carried out’. This objective was addressed by Article 38(4), requiring that a Member State ‘ensure’ access to an asylum decision on the merits, in accordance with the usual rules, if the ‘safe third country’ refused readmission. And then a key point:

54      It follows that, where it is established that the third country designated as generally safe by a Member State does not in fact admit or readmit the applicants for international protection concerned, that Member State cannot reject their applications for international protection as inadmissible on the basis of Article 33(2)(c) of Directive 2013/32. Furthermore, that Member State may not unjustifiably postpone the examination of those applications and must, inter alia, ensure that that examination is conducted on an individual basis, in accordance with Article 10(3)(a) of that directive and in compliance with the time limits set out in Article 31 thereof.

This interpretation, in the Court’s view, ‘is likewise not such as to deprive of any practical effect the right of an applicant for international protection, as enshrined in Article 18 of the Charter and given specific expression by that directive, to obtain the status of beneficiary of international protection, provided that the conditions required by EU law are met.

Having ruled, in effect, that the designation of Turkey was valid, the CJEU felt it was unnecessary to answer the second and third questions referred by the national court, on whether the refusal of readmission was a factor to be considered at the admissibility or the enforcement stage. But as we have seen, the Court had already de facto answered the second question: if it is ‘established’ that the ‘safe’ third country ‘does not in fact’ readmit the asylum seekers concerned, a Member State ‘cannot’ find their applications admissible, and must consider those applications on the merits (moreover, without ‘unjustifiable’ postponement, and subject to the deadlines in the Directive). And that renders the third question moot.

 

Comments

Greece can have a dog; but it will be wagged by its Turkish tail. That is the effect of the Court’s judgment: in more technical legal language, Turkey can remain on Greece’s ‘safe third country’ list, but the practical impact of this is zero as long as Turkey refuses to readmit asylum-seekers from Greece, because asylum applications cannot be found inadmissible on this ground. The wheel of designation of Turkey as a ‘safe third country’ keeps spinning, but the hamster of inadmissibility of asylum applications is dead. Let’s unpack this further, examining first the reasoning of the judgment and then its potential consequences.  

Judicial reasoning

The CJEU begins by, with respect, answering a ‘straw man’ question that it had not been asked.  The national court did not ask the CJEU if it was possible to designate ‘safe third countries’ at all; rather it asked if those countries could be designated if they failed to readmit asylum-seekers. The references to common principles on ‘safe third countries’ in the preamble to the Directive do not answer the question as to what those common principles are; a rule that failure to readmit precludes listing a country would be a ‘common principle’ too. However, the Court’s points that Article 38(2) on national law and (especially) Article 38(4) on readmission presume that the country concerned is already on a list are more convincing.

Most importantly, the Court’s finding that applications cannot be inadmissible on ‘safe third country’ grounds where the country concerned refuses to readmit, requiring a Member State to consider the merits of such applications, is particularly convincing.  Given that between Greece and Turkey the readmission will not take place, and in light of the evidence that repeat applications are difficult to make and are often rejected, this is the best way to ensure that the obligations set out in Article 38(4), coupled with the principle of considering applications as soon as possible, are met.

Unlike some judgments, this is not so much a ‘liberal’ or ‘activist’ interpretation of EU asylum law by the CJEU, but rather the overdue termination of conservative activism by the Greek authorities and lower courts. In light of the clear obligation in Article 38(4) to consider the merits of an asylum application if the asylum seeker is not readmitted to a ‘safe third country’, in the context of the absolute refusal of Turkey to readmit, the failure of the authorities to consider the merits of many of those applications, and of the courts to require them to do so, is unjustifiable and frankly shocking.

Consequences in practice

First and foremost, what happens to the 10,000 plus asylum seekers whose applications were already rejected as inadmissible? As noted above, some had previously tried to make repeat applications, but faced the difficulty that they had to invoke ‘new elements or findings’ in order to do so, which was often difficult to do. But their ability to invoke the ‘repeat applications’ clause in the Directive has now, as a consequence purely of the Court of Justice’s recent judgment, undergone a Copernican revolution. This is because the CJEU has previously ruled (see Joined Cases C-924/19 and C-925/19 PPU, paras 175-203, confirmed in Case C-216/22) that its own judgments on EU asylum law are a new element or finding in filing a repeat application, where those judgments are relevant and add significantly to the likelihood of an asylum application being successful. For those asylum-seekers whose application was found inadmissible on ‘safe third country’ grounds despite Turkey’s refusal to readmit them, last week’s judgment is obviously highly relevant; and overturning the finding that their applications were inadmissible equally obviously significantly increases the odds that their application will be successful. (Moreover, the judgment in Joined Cases C-924/19 and C-925/19 PPU is especially relevant by analogy here, since the prior national judgments which conflicted with a later CJEU judgment in that case also concerned an unlawful application of the ‘safe third country’ rules).

Therefore, if the rejected asylum-seekers apply again, the Greek asylum system will now have to decide on the merits of thousands of asylum applications that it wrongly ruled were inadmissible in the first place. A pending CJEU case referred from Ireland may be relevant by analogy in establishing whether Greece has to pay damages too. Finally, if any Afghan women asylum-seekers were among those whose applications were considered inadmissible, they can also rely on another CJEU judgment from the same day (discussed here) as a ‘new element or finding’, supporting the merits and fast-tracking of their gender-based persecution claims.

As for current or future applicants, their applications cannot be rejected as inadmissible on the grounds that Turkey is a 'safe third country' at all. Nor can they be rejected on the grounds that Turkey is a ‘first country of asylum’ for them: without even looking at the substance of that concept, the CJEU judgment explicitly stated that this principle cannot even get off the ground if the country concerned will not readmit the asylum seeker.

What about applicants after the asylum pact takes effect? The Court of Justice judgment from last week only mentions the new asylum procedures Regulation (discussed here) in passing, but as the Advocate-General’s opinion pointed out (in para 64), the interpretation of the Directive ultimately followed by the Court is explicitly set out in the Regulation, now being built into the definition of inadmissibility – also as regards the ‘first country of asylum’ principle (these grounds of inadmissibility can be invoked for either principle ‘unless it is clear that the applicant will not be admitted or readmitted to that country’). (See also recital 53 in the preamble to the Regulation). Whether the planned review of the ‘safe third country’ rules in the Regulation in 2025 affects these issues remains to be seen.

Finally, the Court’s judgment understandably only dealt with situations where the refusal of readmission is certain at the time of considering the possible application of the ‘safe third country’ rule. It did not deal with the scenario where refusal was merely a possibility – as that was hypothetical on the facts. How should such cases be dealt with?

The starting point is that the current Directive (recital 44 in the preamble) implies the possibility of an inadmissibility decision in that scenario: Member States do not have to examine the merits of an application where ‘there are grounds for considering that the applicant will be admitted or readmitted to that [‘safe’ third] country’. The inadmissibility rule in the new Regulation points in the same direction: it can’t be applied if ‘it is clear that’ readmission is not going to happen (see also recital 43 in the preamble to the Regulation). But this implies that the inadmissibility rule can be applied if it is not clear that readmission is ruled out. On the other hand, the recent judgment points to the importance of deciding on applications as soon as possible, without unjustifiable postponements, while ensuring the application of the Charter right to asylum. How best to reconcile the possibility of the adoption of inadmissibility decisions with the need to take decisions speedily and afford an effective prospect of having an asylum application decided on the merits, if there is uncertainty about whether readmission will actually take place?

The 2024 procedures Regulation already expressly provides (in recital 53 in the preamble) that access to the Member State’s asylum system must be offered again if readmission does not take place:

…if the applicant is eventually not admitted or readmitted to the third country after the application has been rejected as inadmissible, the applicant should again have access to the procedure for international protection in accordance with this Regulation

How should that principle apply in concrete terms? The best approach is to infer an obligation upon the Member State to act with due diligence to attempt to secure the readmission, starting as soon as it takes the inadmissibility decision (after any appeals, or the prospect of appeals, against that decision are spent). From that point, the Member State should have a reasonable time frame to request readmission from the non-EU country concerned. By analogy with the procedures Regulation time limit for deciding on inadmissibility, two months should be sufficient. If the non-EU country refuses readmission, or does not reply within the time limit set out in the readmission arrangements in force, or accepts readmission in principle but does not carry it out in practice within the relevant time limit, then the Member State’s obligation to consider the application on the merits where readmission does not take place must be triggered. Procedurally either the inadmissibility decision must be considered as withdrawn from that point (it makes sense to consider it as only provisional until readmission is actually carried out), or a repeat application must be considered – the requirement of new developments being satisfied by the inability, subsequent to the inadmissibility decision, to ensure readmission in practice within a reasonable time frame. The inadmissibility decision was based on the presumption that readmission would be secured within a reasonable period; that presumption has since been rebutted.

The final word on the impact of last week’s judgment should go to the asylum NGOs who brought the case:

…the CJEU’s judgment is of major importance because it overturns an arbitrary and abusive practice of the Greek authorities which has been in force for many years, of wholesale rejecting asylum applications as inadmissible under the safe third country principle, and puts an end to the denial of the rights of thousands of applicants, by prohibiting their rejection when there is no possibility of their readmission to Türkiye and dictating their individual examination in accordance with the Directive.

‘Can a man serve two masters’? The Court of Justice decides whether monks can be banned from being lawyers

Rebecca Zahn, Senior Lecturer in Law, University of Strathclyde

Lawyers occupy a unique position in the European Union. Within their Member States, they play a vital role in providing access to and administering justice, and upholding the rule of law. As with other professionals such as doctors or architects, access to the profession is tightly regulated by national bodies. Yet unlike other professionals, lawyers are inherently immobile. The knowledge required to be a lawyer is closely linked to the jurisdiction within which an individual trains and qualifies. An understanding of the national language as well as of a particular legal system’s customs and practices is vital for the exercise of the profession. This makes the facilitation of free movement of lawyers difficult. Lawyers can make use of the general system of Directive 2005/36/EC, which leads to full integration in the profession of the receiving Member State following the successful completion of an aptitude test. The European Commission’s Regulated Professions Database suggests that 7,506 lawyers have availed themselves of this route to gain admission to the profession in a host State.

Yet in recognition of their special status, lawyers are the only professionals who benefit from an additional regime designed to facilitate temporary and permanent free movement.  Directive 77/249/EEC permits lawyers registered in one Member State to provide temporary cross-border services in another Member State without any prior notification. There are no statistics available on the use of this right but a 2012 study for the European Commission evaluating the Legal Framework for the Free Movement of Lawyers suggested that there was a large market for the temporary cross-border provision of services albeit with limited physical movement; often such services are provided at a distance, for example by e-mail or telephone.

Finally, Directive 98/5/EC allows lawyers to register to practice law on a permanent basis in a host Member State. Article 3 of that Directive mandates that lawyers wishing to practice in a host Member State register with the competent authority by presenting their registration certificate. The competent authority shall register the lawyer initially under their home-country professional title which allows the lawyer to give advice on his home law, EU law, international law or the law of the host Member State (Article 5(1)). Under Article 6, the host Member State’s professional rules of conduct will apply to the lawyer’s exercise of activity in that State. Following three years of practice in the law of the host Member State, the lawyer can apply to be admitted to the profession of lawyer in the host Member State (Article 10).

The Directive thus harmonises access to the profession under Article 3 while granting Member States discretion when it comes to integrating individuals through its own rules of professional conduct. However, this hybrid nature of the Directive can lead to difficulties as was made clear in a judgment handed down by the Grand Chamber of the Court of Justice of the European Union on 7 May 2019 in Case C-431/17 Monachos Eirinaios v Dikigorikos Syllogos Athinon.

The case concerned a Greek monk, Monachos Eirinaios, who lives in a monastery in Greece. He is also a qualified lawyer and a member of the Cyprus Bar Association. In June 2015, he relied on Article 3 of Directive 98/5 to register with the Athens Bar Association as a lawyer who has acquired his professional title in another Member State. The application was refused on the basis of the national rules on incompatibilities contained in Article 8(1) of the Presidential Decree 152/2000 and Article 6 of the Lawyers’ Code which prohibit a clergyman or monk from being a lawyer in Greece. The Athens Bar Association argued that monks were barred from becoming lawyers due to the absence of guarantees regarding their independence, doubts as to their ability to occupy themselves fully with their functions and whether they can handle contentious cases, the requirement for actual establishment in the geographical area of practice and the obligation not to provide services without remuneration. Monachos Eirinaios appealed the decision and the Council of State referred a question to the Court of Justice asking whether Article 3 of Directive 98/5 required the Athens Bar Association to register Monachos Eirinaios even though he would not be allowed to practice once registered.

In a short judgment, the Court reiterated that Article 3 of Directive 98/5 harmonises the rules under which a lawyer can establish himself in a host Member State. Provided that the lawyer is fully qualified and has the requisite certificate from the home Member State, the host Member State authorities must register him regardless of whether he is subsequently able to practice. National authorities are not permitted to impose additional conditions for registration other than those contained in article 3. This follows from the decision in Case C-58/13 Torresi (discussed here) where the Court established the mutual recognition of the professional titles of lawyers. The Court recognised the absurdity of the situation in Monachos Eirinaios where it was requiring the competent authorities to issue a registration certificate with the knowledge that the individual would not be able to practice.

However, the Court drew a clear distinction between access to the profession on the one hand and its practice on the other hand. The Court accepted that in relation to the latter national authorities may wish to impose certain requirements in order to maintain professional standards. The Court did not comment on the nature of the complete ban on monks becoming lawyers (saying only that it could not be a prerequisite for registration) but instead required the national court to determine whether the ban complied with the principle of proportionality (without giving specific guidance on the criteria to be applied).

The decision in Monachos Eirinaios follows on from the Court’s previous jurisprudence on Directive 98/5 which has focussed on facilitating the free movement of lawyers. In doing so, it is to be welcomed. Given the sensitive nature of the topic both in terms of access to a highly regulated profession and, in the specific Greek context, of the evolving relationship between church and state, the decision also strikes a subtle balance between facilitating access while maintaining Member State discretion. It is therefore perhaps not surprising that the case was heard by the Grand Chamber. However, by circumventing any discussion of the complete ban on monks and clergymen becoming lawyers, the Court has indeed created an absurd situation; namely, that the Athens Bar Association is required to register Monachos Eirinaios and can then immediately strike him off. More detailed guidance on the proportionality test to be applied in these circumstances could have been beneficial especially as the Advocate General engaged in a more nuanced discussion of the topic.

The Advocate General suggested that the rule bans all individuals with particular characteristics from practicing law and thus should not fall under professional conduct. This is a well-spotted flaw in the rules. The Advocate General suggests instead that the Bar Association should be required to see how a lawyer conducts himself in practice before removing registration (following the reasoning in Case C-225/09 Jakubowska where it was recognised that failure to comply with host State rules may lead to a foreign lawyer being removed from the register).

In this, she is correct. In effect, the Bar Association is making an assumption that monks are incapable of fulfilling the rules of professional conduct required of lawyers because they are monks. There is no need for any assessment as to whether the monk (or clergyman) is practicing his religion. Indeed, there is also no clear definition of the scope of the definition of monk or clergyman for the purposes of the Directive (what if, for example, someone is a lay preacher in another Member State – does this classify as a clergyman in this instance?). If Directive 98/5 is to facilitate free movement of lawyers then there must be a minimum harmonisation of what we understand ‘rules of professional conduct’ to be under Article 6. Blanket bans which relate to characteristics (and do not allow conduct to occur) do not facilitate free movement and render the right to registration in effect nugatory.

Moreover, the procedural guarantees in the case of disciplinary procedures in Article 7 or the right of recourse to a court in Article 9 of the Directive only apply once an individual has practiced as a lawyer which is rendered impossible in this case. If we abstract this case from its specific facts then accepting such a blanket ban in relation to a characteristic is a potentially problematic assumption which could undermine the rule of law. What if a Member State introduces rules which require political party affiliation in order to practice as a lawyer under the guise of professional rules of conduct? In a different factual context, the lack of minimum harmonisation of what we understand rules of conduct to encompass could undermine lawyers’ ability to provide access to and administer justice. In skirting around this topic, the Court of Justice has missed an opportunity to further clarify the scope of Directive 98/5/EC.

Barnard & Peers: chapter 14

Photo credit: InHouse Legal

The Case of the Administrative Arrangement on Asylum-Seekers between Greece and Germany: A tale of “paraDublin activity”?

Stathis Poularakis, Legal advisor - Advocacy Officer Médecins du Monde – Greece*

* Reblogged from the EDAL blog. An earlier version of this article was published in Greek on immigration.gr blog. The views and opinions expressed in this article are those of the author and do not necessarily reflect the official policy or position of Médecins du Monde – Greece. Special thanks go to Evangelia Tzironi, PhD Candidate at the Law School of the National and Kapodistrian University of Athens for proofreading the article.

In mid-August 2018, Germany, Greece and Spain agreed on the sketchy details of the initial migration compromise deal that was reached on the sidelines of the EU Summit in Brussels in late June 2018. In this context, the Ministers on Migration of Germany and Greece reaffirmed their commitment by exchange of letters, to work towards common European solutions and to avoid any unilateral measure with respect to migration and asylum.  

In particular, both countries underlined their support for concluding the revision of the Common European Asylum System by end of 2018 “with the aim of achieving the goal of fair sharing of responsibility and solidarity”. In the event of a crisis – defined as a situation where asylum applications have surpassed a certain percentage e.g. 140% of a State’s fair share of asylum applications based on objective criteria (total population and GDP) – Germany committed to support Greece in the Council (of the European Union), especially on the adoption of additional support measures at European Union level. Germany agrees also that further supportive and development measures need to be adopted on the five Greek islands of Eastern Aegean, where the hotspot approach is implemented, in order to assist local communities. Finally, Germany affirms its commitment to increase the human resource capacity in Greece, through the EU Asylum Support Office (EASO), with the aim to strengthen the asylum system.

The final operational details of the aforementioned political agreement were annexed to the letters, under the Title “Administrative Arrangement”. This blog post aims to outline the key points of this “Arrangement”, to examine its legal nature arguing that argue that this document is a bilateral treaty whose scope extends beyond the Dublin Regulation, and to critically assess its impact on the EU asylum policy.

The content of the Administrative Arrangement

The Administrative Agreement is comprised of 15 articles and divided into three sections. The first part includes arrangements for the readmission to Greece of persons identified in the context of temporary checks at the German-Austrian border, having previously applied for asylum in Greece. The second part includes provisions for concluding pending Dublin cases of family reunification from Greece to Germany, and finally, the third section includes provisions on the review of the implementation of the "agreement", the mutual dispute settlement and beginning and ending of the aforementioned cooperation between the two countries.

In particular, under the "Administrative Arrangement", the following are agreed:

Germany will return to Greece, any adult third-country national who has been identified during a check at the German-Austrian border and wishes to apply for international protection if he/she has already applied for asylum in Greece i.e. when an entry in Eurodac (the EU system of exchanging fingerprints of asylum-seekers) indicates that the person has already requested protection in Greece, and such entry is dated from July 1st, 2017 onwards. Unaccompanied children (under the age of 18) at the time of the identification are excluded from readmission to Greece.

The German Authorities will provide notification of the refusal of entry to the Greek responsible authority using the form annexed to the Agreement via fax or email. The return should be initiated no more than 48 hours after the person has been apprehended, unless the Greek side objects to the return within six hours from the automatic confirmation of the receipt of the notification of the refusal of entry, demonstrating why the conditions of the administrative agreement have not been met. Within 7 days of the readmission, if the Greek authorities demonstrate that the above conditions had not been met (refusal of entry in error), Germany will readmit the person concerned without delay.

Germany commits to swiftly concluding family reunifications from Greece in the framework of the Dublin III Regulation by the end of 2018, with respect to “take charge” requests already accepted by the German Dublin Unit before 1 August 2018. It should be noted that due to the high number of Dublin transfers from Greece to Germany over the last year, asylum seekers entitled to be transferred to Germany under the relevant provisions of the Dublin III Regulation, were “blocked” in Greece for periods exceeding the six-month deadline provided by art. 29 of the Regulation. Germany also undertakes to examine all pending “take charge” requests that have been submitted before 1 August 2018 that have not been examined yet within two months of the beginning of the cooperation between the two parties based on this Administrative Agreement. The number of people to be transferred from Greece to Germany is capped at 600 people per month and family reunifications should be completed by December 2018.

At the same time, the German authorities will examine and reply "without undue delay" to all requests for re-examination submitted before 1 August 2018. To this end, the Greek authorities will provide a relevant sortable list with case numbers and submission dates. Passports and ID cards, marriage and birth certificates, family booklets, as well as the transcript of the interview with the person concerned, will be considered as evidence, when submitted in relation to a currently pending request for re-examination. These documents will be submitted in their original language. A translation in English will be also submitted, if available. It is explicitly envisaged that the submission of documentation in its original language (without translation) may not be used as a justification for rejecting the request for re-examination.

The cooperation between the two countries based on the Administrative Arrangement starts from the day of its acceptance by the Greek Minister on Migration Policy (dated 18.8.2018). The two parties will review the implementation of the agreement on a three-month basis. Each Party may withdraw from the Agreement by notifying the other Party in writing at least three weeks in advance. Lastly, it is explicitly envisaged that the agreement will discontinue upon entry into force of the revised Common European Asylum System.

"Administrative Arrangement" or international treaty?

The first issue raised by this so-called "Administrative Arrangement" concerns its actual legal nature.

Based on the public statements made so far and the title given to the document, it could be argued that the latter could be understood to be an informal non-treaty instrument - an arrangement on operational issues and actions in line with EU law. According to this argument this text seems to be in conformity with Article 36 of the Dublin III Regulation, which stipulates that Member States may, on a bilateral basis, establish administrative arrangements between themselves concerning the practical details of the implementation of this Regulation, in order to facilitate its application and increase its effectiveness. Such arrangements may relate to exchanges of liaison officers and the simplification of the procedures and shortening of the time limits relating to transmission and the examination of requests to take charge or take back applicants.

However, taking as a point of departure that negotiating States have the autonomy to choose between complex or simplified forms of express consent to be bound by a treaty (in the present case by exchange of instruments under Article 13 of the Vienna Convention on the Law of Treaties (VCLT), and given its content, it could also be argued that the “Agreement” could be perceived as an international treaty concluded in simplified form. This agreement in fact sets forth the “technical” modalities of the earlier political agreement between Chancellor Merkel and Prime Minister Tsipras in Brussels last June, in the sidelines of the EU Summit. International treaties in simplified form are concluded without the need for ratification by the State, nor the consent of the parliament. They are usually signed by Ministers or any other body authorized under national law and are related to technical or administrative issues of cooperation, military pacts concluded on a battlefield or arrangements for the application of a binding (“typical”) international treaty.

 

In order to draw a safe conclusion on the nature of the “administrative arrangement”, all the circumstances around the exchange of letters should be meticulously examined. “Treaty” means an international agreement concluded between States in written form and governed by international law, whether embodied in a single instrument or in two or more related instruments and whatever its particular designation (Article 2(1)(a) of the VCLT). Whatever its designation (convention, accord, protocol, exchange of letters, agreed memorandum, memorandum of understanding), the intention of negotiating parties to create binding legal obligations marks the difference between treaties and informal instruments. To wit, even an exchange of letters could be considered as an international treaty.

A closer look on the actual content of the "Arrangement" suggests that this is not just an informal international non-binding agreement or arrangement on technical details of cooperation between the two Parties for the implementation of the Dublin Regulation, but rather a legal instrument setting forth “new” binding rules whose scope moves beyond the obligations established under the Dublin III Regulation.

In fact, the first part of the “Arrangement” contains clauses that are usually included in readmission agreements. No such an agreement has been ever signed between Germany and Greece. In particular, the clauses of the Arrangement lay down, among others, the specific conditions for readmission, the competent authorities of each State, the procedure to be followed, the responsible State in terms of costs etc. A standard template form is also included so as to facilitate the return. In reality, the “Arrangement” does not establish simplified and streamlined procedure for “Dublin transfers” (the wording “transfer” in not used in the text) but rather a fast-track readmission procedure for persons refused entry at the German-Austrian borders, likewise in any other case of an irregular migrant refused an entry in Germany. 

Another crucial element regarding the legal nature of this “Arrangement” lies with the specific clauses on family reunification that undoubtedly enumerate commitments to which the Parties have consented and their intention to create legal rights and obligations that go beyond those already provided for in the "Dublin III” Regulation. Namely, assuming the obligation to accept a transfer of an asylum seeker from Greece to Germany upon expiry of the 6 months’ time-limit under Article 29 of the Regulation – where the respondent MS is relieved of its obligations to take charge or to take back the person concerned – is a “new” obligation for Germany enshrined in the "Administrative Arrangement".

The same also applies for the obligation to re-examine all rejected requests for “take charge”. To avoid any confusion, one should make a clear distinction between the potential responsibility of Germany for the infringement of the time-limits and conditions laid down in the Dublin III Regulation, e.g. for breaching EU law, and the legal obligation of Germany to actually accept transfers for which its responsibility has already ceased under “Dublin”. These obligations are now explicitly assumed by the aforementioned "Administrative Arrangement".

It is worth mentioning that last year the competent Ministers of Greece and Germany had reached to a similar informal arrangement by which Dublin transfers to Germany were capped at 70 persons per month for a certain period of time. This informal agreement, however, was rather a "gentleman's agreement" between the two states without creating clear binding obligations for both parties. In any case, the conformity of this agreement to the Dublin Regulation is open to discussion, since the Dublin Regulation does not foresee such caps on relevant transfers for administrative convenience reasons.

Finally, one could hardly argue that during Merkel and Tsipras΄ meeting in Brussels, where a common political agreement was reached on migration, the two leaders concluded a formal treaty under international law with the subsequent "administrative arrangement” serving as a treaty in a simplified form, laying down modalities for the application of the treaty concluded by the two heads of state.

For all these reasons, the "administrative arrangement" should not be considered as a "gentlemen's agreement", nor as an administrative arrangement under Article 36 Dublin III Regulation but rather as a binding bilateral treaty whose provisions establish obligations that go beyond the scope of obligations established under the Dublin III Regulation. The agreement thus is governed by international law; an international treaty between Greece and Germany. This contravenes EU law which does not allow legislation at national level or bi/multilateral inter-se agreements in policy areas of shared competence, to the extent that the EU has exercised its competence, as it did through the adoption of the Dublin III Regulation (see TFEU Art 2(2); TEU Art. 4(3), third indent, and Protocol 25 to the TFEU). 

Dublin revisited or further violation of EU law?  A glimpse into the future of EU asylum policy

Another key point of concern is the impact of this bilateral agreement on EU’s asylum policy.

Could this “Administrative Arrangement” be seen as a new, enhanced and adapted to current developments, version of the Dublin rules, that is to say a bilateral agreement between two Member States that aims at strengthening the Dublin rules and principles or just another agreement beyond the scope of Dublin Regulation governed by international law?

One could argue that such bilateral agreements are rather symbolic in nature. They indicate that compromises at European Union level are feasible. As talks on the revision of the Dublin III Regulation stall, the establishment of interim measures among those Member States willing to move ahead collectively at European Union level (“enhanced cooperation”) seems indeed to be the only possible way to actually make progress in this area. However, that enhanced cooperation should to be done under the framework of EU law, i.e. following the procedures of EU law and not bilaterally as an inter-se agreement.

But is this "administrative arrangement" actually such a measure of European Union cooperation? Although the responsible Ministers expressly commit to continue working towards common European Union solutions and avoiding unilateral measures in relation to asylum and migration, this agreement is in fact just another – deeper – derogation from/violation of the European Union acquis. In fact, through such agreements, Germany cooperates with Member States serving as a key point of entry in EU by creating a “Quasi-Dublin” system / a Dublin-like system creating obligations that go beyond the scope of the Dublin III Regulation (i.e. extension beyond the 12 month period for the responsibility for irregular entry) and limitations that are not foreseen in the Regulation (such as caps on the numbers of transfers under the family reunion clauses). A system which, on the one hand, follows the general philosophy of the Dublin Regulation (application of the first country of entry criterion), but, on the other hand, is "free" from all “obstacles” that could jeopardize swift returns – namely clauses and safeguards provided for asylum seekers by the Dublin Regulation, such as family unity criteria, right to appeal against the application of Dublin criteria and deadline to appeal against the transfer decision etc.

Undoubtedly, Southern Member States serving as the main gateways to Europe for third country nationals are in absolute need of European Union solutions on the basis of solidarity and fair responsibility sharing. In this context, the revision of the Dublin III Regulation and the introduction of a permanent allocation mechanism among European Union states is a key issue for Greece. One can hardly wonder why Germany would then be interested in reforming the Dublin Regulation if a more flexible "Quasi-Dublin" system- tailored to its needs- has already being established. Bilateral agreements, such as this, with one of the most prestigious EU countries, could draw away any possibility of a positive revision of the Dublin Regulation for Southern Member States in the near future.

In any case, this "administrative arrangement" marks an alarming development in the context of Greece’s migration policy. To date, Greece has signed and ratified a number of agreements on the readmission of persons irregularly entering or staying in its territory. For all these agreements, the usual process of concluding international treaties was followed: Signing of an agreement, protocol, or treaty and its ratification by the Parliament, according to Article 36 of the Greek Constitution. On the contrary, in the present case, it is the first time that a readmission agreement is concluded by Greece through an exchange of letters between Ministers. Though such an agreement is totally valid and binding under international law, the fact that it not only deals with international relations and migration policy but ultimately with human rights, is concluded away from parliamentary scrutiny and procedures – without even being published in the Government Gazette – raises important concerns on transparency and the rule of law.

Barnard & Peers: chapter 26

JHA4: chapter I:5

Photo credit: Fortune