Friday, 20 March 2026

The Chios Incident: Echoes of Pylos Humanitarian Disaster and Greece's Criminalization of Solidarity


 

Georgios Athanasiou, PhD Researcher, University of Antwerp

Photo credit: Julian Lupyan, via Wikimedia Commons

The Chios migrant boat shipwreck of 3 February 2026 exemplifies the acute tensions between Greece’s increasingly securitized border management and its obligations under EU law, the ECHR, and international maritime conventions such as the 1979 Search and Rescue (SAR) Convention. More specifically, off the coast of Chios island, a Hellenic Coast Guard patrol vessel collided with an inflatable boat carrying approximately 39 Afghan nationals, resulting in 15 deaths and 24 injuries, including 11 minors, and cases of miscarriage. All of the deaths were attributed to severe head trauma rather than drowning, per up-to-date autopsy reports, with survivors claim that the coast guard did not offer any prior warning or communication before ramming the migrant boat, contrary to official claims of the migrants’ speedboat initiating contact. Interestingly, the on boat cameras of the patrol vessel had been deactivated.   

This event parallels the 14 June 2023 Pylos shipwreck, Europe’s deadliest maritime migration tragedy, where over 500 lives (mainly Syrian, Pakistani, and Egyptian) were lost after the overcrowded trawler Adriana capsized, allegedly due to Coast Guard towing maneuvers following delayed rescue operation, despite prior distress alerts. The ongoing criminal proceedings in Greece have charged 17 Coast Guard personnel, including the rescue vessel captain, with felony offenses such as endangering lives and contributing to the shipwreck,. This development appears to be part of a systematic attempt to portray the eastern Mediterranean migration route as inherently life-threatening for asylum seekers, thus reflecting a pervasive securitization narrative guiding border policies of the Greek government that overshadows State accountability.

Legal Parallels and Accountability Gaps

From a legal aspect, both incidents implicate Greece’s positive obligations under Article 2 ECHR (right to life), requiring States to safeguard lives within their jurisdiction, including during maritime interceptions, and conduct effective, independent investigations into fatalities. The ECtHR has repeatedly held Greece accountable in analogous cases: in Safi and Others v. Greece, for inadequate protection and probing of a sunk migrant boat; Alkhatib and Others v. Greece, for excessive lethal force lacking “absolute necessity” and deficient regulatory frameworks for Coast Guard firearms use. Such repeated failures in border management operations seem to formulate a consistent pattern of action of the Greek authorities in handling migrant routes, in an attempt to not allow migrants to enter Greek territory/territorial waters. In this sense, although Article 3 ECHR (prohibition of inhuman/degrading treatment) further prohibits collective expulsions or pushbacks, this practice has been deemed systematic by Greek authorities in A.R.E. v. Greece (also see, here).

Under EU law, the Qualification and Asylum Procedures Directives respectively mandate the upholding of the principle of non-refoulement and individual assessments of asylum applications, while Article 4 of Protocol 4 of the ECHR and Article 19 CFR bar collective expulsions. Meanwhile, the SOLAS and SAR Conventions impose duties to render assistance “without delay” to persons in distress, disembarking them to a place of safety, irrespective of nationality or the existence of a right to enter the country. Hence, any form of interception framed as SAR mission cannot justify pushbacks or endangering the lives of migrants.

Greece’s Restrictive Policies and Criminalization of Solidarity

Domestically, this incident aligns with broader migration policy tendencies, as Greece has instrumentalized criminal law in an attempt to restrict migration, rendering irregular entry, stay and exit of the country a felony punishable with up to 5 years of imprisonment coupled with a minimum fine of €5,000 (Law 5226/2025 Government Gazette Α' 154/8.9.2025). Similarly, rejected asylum seekers face administrative fines up to €10,000, as well as up to five-year sentences or electronic ankle monitoring. Hence, the 2025 deportation law, hailed as Europe’s most stringent, essentially attempts to streamline expulsions of “economic migrants,” given that long-term regularization after 7 years of stay in the country is equally abolished.

This framework cannot be dissociated from Greece’s post-2019 migration hardening: escalated border fortifications (Evros 35 km wall), freezing of asylum applications, and systematic pushbacks exceeding 540 incidents between 2020-2022 (also see, here). Hence, high-seas shipwrecks, such as the Chios and Pylos lethal incidents, epitomize how this apparatus practically overrides positive obligations under the ECHR, as well as international humanitarian and maritime law, subordinating the protection of life at sea to national security imperatives.

The Greek Government defends its approach as prevention of illegal entry, invoking safe third country safeguards, especially for migrants arriving from Turkey, yet these yield no derogation from non-refoulement or collective expulsion bans. Hence, the Government’s approach in migration policies embodies a “fortress mentality,” which, coupled with its recent attempts to criminalize solidarity, further sets in danger the lives of migrants attempting to cross the Eastern Mediterranean route. A prominent example of this criminalization tendency include the recent Lesbos case against 24 rescuers, who were acquitted after years on charges, like espionage and smuggling, that carried up to 20-year sentences. Similarly, Norwegian activist Tommy Olsen faced an arrest warrant in February 2026 for documenting pushbacks via Aegean Boat Report, accused of criminal organization. Finally, a February 2026 migration law amendment (Law 5275/2026, Government Gazette Α’ 17/06-02-2026) makes NGO membership an aggravating factor, escalating misdemeanors (e.g., facilitation of stay) to felonies with fines exceeding €100,000, constituting the largest criminalization of solidarity in the EU.

Analysis

It is apparent that the Chios tragedy, when assessed alongside the Pylos shipwreck, does not constitute an isolated operational failure but rather indicative of a structural recalibration of border governance in the Eastern Mediterranean. The shift from the enforcement of search-and-rescue obligations to human rights violations at the EU’s external borders under SAR cover reveals a normative inversion: life-saving obligations are operationalized through a security prism that treats irregular entry primarily as a threat vector rather than a protection trigger.

At the doctrinal level, Article 2 ECHR imposes both substantive and procedural duties on States. More specifically, from a substantive aspect, States shall refrain from unlawful deprivation of life and adopt preventive operational measures where authorities knew or ought to have known of a real and immediate risk. Meanwhile, procedural obligations mandate to conduct prompt, effective, and independent investigations capable of leading to accountability. In both Chios and Pylos, the central legal question is whether Greek authorities fulfilled the due diligence threshold required during maritime interception. The reported deactivation of onboard cameras in Chios and the delayed rescue response in Pylos indicate the State’s unwillingness to comply with these operational standards.

In a similar vein, the 2026 legislative reform represents an internal consolidation of the securitization paradigm. By reclassifying irregular entry and facilitation-related conduct as felonies and elevating NGO affiliation to an aggravating factor, the Greek legal framework operationalizes criminal law as a migration-management instrument, fully adopting a “Crimmigration” approach, that is the convergence of criminal and immigration enforcement logics, as border management framework.

In this sense, the prosecution of humanitarian actors in the Lesbos case and proceedings against figures associated with monitoring networks reinforce a chilling effect on civil society oversight. When accountability mechanisms (NGO monitoring, documentation of pushbacks) are suppressed, the evidentiary architecture for fundamental rights protection is simultaneously weakened. In practical terms, criminalization of solidarity indirectly facilitates impunity.

Greece’s approach cannot be decoupled from the broader EU externalization strategy. Financial and operational support through Frontex, coupled with political endorsement of deterrence metrics (reduced arrivals as “success indicators”), generates structural incentives that privilege interdiction over protection. In this context, it appears that the Eastern Mediterranean has become a testing ground for this hybrid governance model of the EU’s external borders. This primarily includes operational opacity (restricted access, disabled recording systems), normative elasticity (expansive security justifications), and penal reinforcement (domestic felony frameworks). In other words, the legal tension at stake is not merely compliance with international human rights law but the hierarchy of values underpinning EU border management. If border integrity consistently supersedes the core values of life and human dignity, the doctrinal architecture of human rights law is functionally subordinated to security rationales.

Accordingly, the Chios incident should be analyzed not only as a maritime tragedy but as a constitutional stress test for the EU human rights regime. The decisive issue is whether accountability mechanisms, domestic courts, the ECtHR, EU oversight bodies, will be able to effectively recalibrate operational practice toward a life-preserving baseline or tacitly normalize deterrence-driven fundamental rights erosion.

Conclusion

Greece exemplifies a broader European paradigm: a so‑called “success story” for deterrence‑based migration control, yet in reality a humanitarian catastrophe for those seeking protection. The country’s migration policies mirror a wider EU strategy that prioritizes border fortification over human life. Hence, a rights‑first recalibration is urgently required. This entails independent and transparent investigations into all reported maritime incidents, such as the full public release of the Chios and Pylos footage, and unhindered support for NGOs engaged in SAR operations, paired with the domestic decriminalization of humanitarian assistance to migrants. In the absence of these measures, the prevailing doctrine of “prevention at all costs” will perpetuate watery graves, turning the Mediterranean into an open cemetery and rendering the protection of migrants’ fundamental rights mere eulogies in default.

 

Monday, 16 March 2026

The Aleb judgment on ‘safe third countries’ in asylum law: the CJEU’s answer to EU legislative amendments?

 


 

Tamta Gventsadze, PhD candidate in law, UNITUS

Photo credit: Mstyslav Chernov, via Wikimedia Commons

 

Introduction

On February 5, 2026, the Court of Justice delivered its judgment in Case C-718/24, Aleb concerning the interpretation of Articles 33(2)(c), 38 and 46 of Directive 2013/32 in the context of the ‘safe third country’ concept and the right to an effective remedy, Article 47 of the Charter of Fundamental Rights of the European Union.

This analysis examines the Aleb judgment in light of the ‘safe third country’ concept under Directive 2013/32 (Asylum Procedures Directive, APD) and its replacement by Regulation (EU) 2024/1348 (Asylum Procedures Regulation, APR) from 12 June this year. After mapping the factual background and the Court’s clarification of the cumulative safeguards governing the presumption of safety, it then considers the impact of the 2026 amendments and assesses their implications for judicial control and fundamental rights protection.

I. Facts of the Case

The applicant, NP, is a Syrian national and an unaccompanied minor who lodged an application for international protection in Bulgaria on 2 November 2023. During the interview conducted on 1 December 2023, he stated that he had lived in Aleppo (Syria) and had left two to three months earlier with his brothers because of the war. Before “illegally entering the Bulgarian territory” (para. 19), NP stayed in Türkiye for approximately one month, where his brothers remained with three of his sisters who already lived there.

By decision of 18 June 2024, the Chairperson of the Bulgarian National Refugee Agency rejected the application, refusing to grant him both refugee and humanitarian status. The authority accepted that Syria was affected by internal armed conflict and indiscriminate violence and acknowledged that the applicant was exposed to a real threat to his life or person. Nevertheless, it refused to grant protection on the ground that Türkiye constituted a ‘safe third country’ in which the applicant could seek protection. The decision relied, inter alia, on the fact that the applicant had already lived in Türkiye for about a month without suffering harm, had close family members there, and that Syrian nationals in Türkiye benefited from temporary protection and protection against forced return; finally, their basic needs [were] satisfied” (para. 20).

The referring court expressed doubts as to the compatibility of this approach with Directive 2013/32, especially considering the absence of both a defined methodology for applying the ‘safe third country’ concept and the lack of explicit procedural guarantees under Bulgarian law allowing the applicant to challenge the existence of a sufficient connection with Türkiye (paras 24-26).

II. The Presumption of Safety under Article 38 of the Directive

The judgment deals with the legal nature and limits of the presumption underlying the ‘safe third country’ concept. The Court begins by recalling that the application of Article 33(2)(c) of the APD (ie, providing that asylum applications are inadmissible where the ‘safe third country’ principle is applied) is conditional upon strict compliance with the requirements of Article 38 thereof (ie the definition of ‘safe third country’ and the conditions related to it). In para. 48 of the judgment, it is expressly stated that the conditions laid down in Article 38 are cumulative, with the result that the inadmissibility ground cannot be applied where any one of those conditions is not satisfied.

This formulation makes clear that the presumption of safety is neither automatic nor self-standing, it is legally constructed and constrained by a series of substantive and procedural safeguards (para. 46). More precisely, Article 38(2) requires Member States to regulate the safe third country concept through national law and that the national rules ensure: (i) there is a sufficient connection between the applicant and the third country so that return there is reasonable; (ii) must define a methodology for applying the concept, which includes either a case-by-case safety assessment or the designation of generally safe countries; (iii) must guarantee an individual examination and allow the applicant to challenge both the safety of the third country in their specific circumstances and the existence of the required connection.

Therefore, the key requirement in the safe third country designation is the existence of a “connection” between the applicant and the third country. The Court emphasizes that Article 38(2)(a) obliges Member States to define in national law criteria enabling authorities to determine whether such a connection exists and whether return to that country is reasonable (paras 51–52). Since the Directive does not define “connection,” Member States retain discretion to specify the criteria, but within EU limits.

Importantly, the Court further reiterates its prior case law that mere transit through a third country cannot, on its own, justify the conclusion that return there is reasonable (para. 54). This statement substantially narrows the presumption. It prevents Member States from relying on minimal factual links and requires a qualitative assessment of the relationship between the applicant and the third country, considering factors such as duration and circumstances of stay and family ties. It is evident that the presumption cannot be based merely on the “transit” criterion. Even where national law relies on the notion of “stay,” national courts must assess, in light of all circumstances, whether that stay genuinely establishes a sufficient connection (para. 55).

The Court also addresses national lists of safe third countries. Member States may, in principle, designate safe third countries by general act. However, such designation does not dispense with the obligation to conduct an individual assessment. National law must provide a methodology for a case-by-case evaluation of both the country’s safety for the applicant and the existence of a sufficient connection (para. 65). The presumption must remain rebuttable, and the applicant must be able to challenge the existence of that connection.

The judgment firmly situates the safe third country concept within the framework of effective judicial protection. Article 38(2)(c) must be read in conjunction with the ‘effective remedy’ rights in both Article 46 of the Directive and Article 47 of the Charter (paras 69-74). Even if national law does not explicitly confer such power, a court hearing an appeal must verify whether a sufficient connection exists.

Accordingly, the Court recalls that Article 46(1) of the Directive guarantees a right to an effective remedy in asylum cases and that Article 46(3) requires a full and ex nunc examination of both facts and law in asylum appeals. This standard applies even in inadmissibility cases and does not necessarily require a substantive assessment of protection needs, but it does require full judicial scrutiny of admissibility conditions. Furthermore, Article 47 of the Charter enshrines the principle of effective judicial protection and is directly applicable, thereby requiring national courts to conduct comprehensive review consistent with EU fundamental rights standards.

Therefore, when reviewing a decision declaring an application inadmissible on ‘safe third country’ grounds, national courts must conduct a full and up-to-date examination of whether the third country is safe for the applicant and whether all cumulative conditions, including the connection requirement, are fulfilled (para. 75). The Court thus subjects the presumption of safety to meaningful judicial scrutiny grounded in Article 47 of the Charter.

Finally, the Court confirms that an application may be declared inadmissible on ‘safe third country’ grounds even where the applicant faces a real risk of serious harm in the country of origin (ie, as distinct from the ‘safe third country’). This confirms that the ‘safe third country’ concept functions as a jurisdiction-allocating mechanism rather than as a substantive denial of risk in the country of origin. Precisely because it allows displacement of protection despite such risk, strict compliance with Article 38 safeguards is imperative.

III. Relevant changes of legislative framework

It is further necessary to underline that while the Court has provided meaningful clarifications regarding safe third country concept and judicial protection in light of corresponding provisions under Directive 2013/32 in a few months’ time this instrument will be replaced by another secondary EU legislation in the form of Regulation (EU) 2024/1348 of the European Parliament and of the Council of 14 May 2024 establishing a common procedure for international protection in the Union and repealing Directive 2013/32/EU (already discussed in great detail here).

Upon a comparative assessment of these two contrasting instruments, several illustrative differences emerge between the provisions concerning safety presumptions, and those of appeals.

Article 33 of the APD addresses inadmissible applications, framing them as an exception to the obligation to examine the substance of an application, meaning that Member States are not required to assess the merits of qualification for international protection where one of the exhaustively listed grounds for inadmissibility applies. The list is limited: Art. 33(2) includes protection granted by another Member State, first country of asylum, ‘safe third country’, subsequent applications without any new elements, and applications lodged by dependents who had consented to be included in another application.

Article 38 of the APR restructures this framework, by separating the decision on admissibility from the decision on the merits. Article 39(3) explicitly governs rejection as unfounded following substantive examination pursuant to the ‘Qualifications’ Regulation (EU) 2024/1347. This structural clarification seems to strengthen the distinction between inadmissibility and unfoundedness, which had already been emphasized by the Court.

In addition, relevant to the current analysis, a notable development concerns the first country of asylum and ‘safe third country’ grounds of inadmissibility. In the APD, Article 33(2)(b) and (c) refer to Articles 35 and 38 without expressly conditioning inadmissibility on the likelihood of admission or readmission. The Regulation introduces an explicit safeguard, with Article 38(1)(a) and (b) essentially stating that inadmissibility may apply unless it is clear that the applicant will not be admitted or readmitted to the third country. Therefore, changes regarding admissibility grounds combine expansion with procedural tightening and partial codification of judicial safeguards.

To further detail the elements regarding the concept of ‘safe third country’, it seems clear that the evolution of the rules on this principle from Article 38 of the APD to Article 59 of the 2024 Regulation, especially as amended in 2026, reflects an apparent shift in structure and scope, even if several aspects remain unchanged.

First off, at the level of safety criteria, both instruments require absence of threats to life or liberty on Convention grounds, absence of serious harm, respect for non-refoulement, and protection against removal contrary to international law. The Regulation clarifies the content of “effective protection” through reference to Article 57 and to Regulation (EU) 2024/1347. As defined by Article 57, effective protection can be attained in a country that has ratified and respects the Geneva Convention on Refugee Status, within any permitted reservations or limitations. However, where a geographical limitation applies, or where the Convention has not been ratified, protection must be assessed against minimum criteria of permission to remain on the territory, access to sufficient means of subsistence to ensure an adequate standard of living, access to healthcare and essential treatment, access to education under general national conditions, and the availability of protection until a durable solution is found. Interestingly, these minimum criteria resemble those of subsidiary protection guarantees, but the context refers to international protection.

Secondly, additional significant changes concern Union and national designation mechanisms introduced by the Regulation. Article 59(2) and (3) of the Regulation, which have no previous equivalent in the Directive, allow partial designation for specific territorial parts or identifiable categories of persons and require reliance on a broad range of sources. Importantly, Art. 59(5)(b) explicitly maintained, that the ‘safe third country’ concept may only be applied where “there is a connection between the applicant and the third country in question on the basis of which it would be reasonable for him or her to go to that country” (the same standard of ‘reasonable’ connection as employed by the Court in Aleb).

However, the 2026 amendments substantially reshape the connection criterion, by removing it as a mandatory condition and introducing two additional bases. The concept may now apply where the applicant transited through the third country (explicitly opposing to what the Court prohibits in Aleb). It may also apply where an agreement or arrangement exists requiring the third country to examine protection claims. As already wittily named elsewhere, this ‘Rwanda clause’ permits transfer without prior connection or transit. Moreover, the removal of automatic suspensive effect in safe third country appeals in the 2026 amendments seem to further intensify a restrictive shift, although there is still a possibility for requesting a judicial suspension.

The most problematic aspect seems this possibility of transferring an individual to a state where they might have never even been to, but the secondary law would permit it solely due to the existence of an agreement (often in non-legally binding form of MoUs) between a EU Member State or the Union and possibly any third country. It is true that an applicant will maintain a possibility to appeal this decision, however, they might have to first ask for a suspensive effective of the appeal, which further complicates an already arduous procedure. In theory, a person might end up being transferred to such a supposedly ‘safe’ third country before a decision is made upon their request to remain, which could create a rather unclear legal consequence; would a person potentially have to be brought back, shall their appeal succeed? It is true that this particular scenario might be extremely rare in practice, but does this rarity allow for its legality?

Nevertheless, despite this expansion in scope, the APR maintains several core safeguards, explicitly stating that admission or readmission must be ensured and that individual assessment remains required (which could become more restricted in practice). Special guarantees also apply to unaccompanied minors, including best interests and prior assurances of protection; and the ‘Rwanda’ clause cannot apply to unaccompanied minors at all.

As confirmed by the Alace judgment, designations of ‘safe countries of origin’ must remain subject to judicial review under Article 47 of the Charter. The Court held that national courts must be able to examine compliance with material designation criteria and to rely on independent sources of information. This reasoning applies by analogy to ‘safe third country’ designations and considering that the Charter has the same legal value as primary EU law, legislative attempts to narrow judicial review cannot override it.

In this context, the Aleb judgment constitutes an additional message. It reaffirms that application of the ‘safe third country’ concept is subject to cumulative conditions and full judicial scrutiny. Even as the legislature broadens the concept and limits suspensive effect, the Court insists on effective judicial protection and strict assessment of safety criteria.

IV. Concluding Assessment

The Aleb judgment does not abolish the presumption of safety inherent in Article 38 of the Directive. However, it subjects the presumption to cumulative substantive conditions, mandatory individualized assessment, and full judicial review.

In more detail, the ECJ allows that the determining authority may apply the concept of ‘safe third country’ on the basis of information from publicly available sources and rely on a national list of safe third countries, if such exists, but this is only provided that national law also defines the methodology applicable for assessing, on a case-by-case basis, according to the particular circumstances of the applicant for international protection, whether the third country concerned satisfies the conditions for being regarded as safe for that applicant and the possibility for that applicant to challenge the existence of a connection, within the meaning of Article 38(2)(a).

Moreover, the Court’s insistence, particularly in paragraphs 48, 54 and 65 of the judgment, on the cumulative nature of the safeguards, the insufficiency of mere transit, and the necessity of a defined methodology, significantly limits the discretionary space of national authorities. The ECJ reinforces the procedural containment of the ‘safe third country’ mechanism, by linking these cumulative requirements to Article 46 of the Directive and Article 47 of the Charter, and by requiring national courts to verify the existence of a connection even where national law is silent in this regard.

Therefore, the Court ties the existence of the ‘safe third country’ presumption to two co-existing elements: clearly defined methodology underlining individual assessment (which includes the existence of “reasonable” connection between the applicant and the safe third country) and a possibility for judicial review of the connection requirement. It seems evident from this judgment, that the safety presumption, be it national or supranational level, would otherwise be invalid. In a way, Aleb strengthens the doctrinal link between inadmissibility decisions and effective judicial protection. The presumption of safety is permitted, per se, but only as a structured, reviewable and rebuttable legal construction embedded within the broader guarantees of EU fundamental rights law.

The combined effect of Alace and Aleb indicates that the Court of Justice does not seem to be prepared for relaxing the standards governing the designation and review of ‘safe third countries’. Nevertheless, it remains to be seen whether and how the Court will respond to the legislative changes, considering that they aim to abolish the mandatory connection element and restrict safeguards when challenging it.

Tuesday, 10 March 2026

Race, Housing, and the Limits of EU Anti-Discrimination Law: A Commentary on the CJEU’s Judgment in the Danish ‘ghetto law’ case

 



 

Karin de Vries (professor of fundamental rights law, Utrecht University) and Sarah Ganty (JSD candidate, Yale Law School; FNRS Post doc fellow, UCLouvain)

 

Photo credit: Kristoffer Trolle, via Wikimedia Commons

 

The so-called Danish ‘ghetto law’ case has been one of the most closely watched judgments of 2025 before the Court of Justice of the European Union (CJEU), long awaited by lawyers working in EU anti-discrimination law and beyond. Its resonance has extended well beyond legal circles, attracting sustained attention from the media and NGOs.

Much has already been written on the judgment, from different perspectives (for example, here, here, here and here). The focus of this contribution is on how the judgment shapes the definition of racial discrimination in EU law, including the emerging tension between objective and social understandings of ethnic origin, the Court’s application of the concepts of direct and indirect discrimination and its reticence to address the structural racism at the roots of the Danish housing policy. Before proceeding to our analysis we first offer a brief outline of the case and the Court’s reasoning.

Background: the Danish ‘ghetto law’

At the heart of the case lies a practice of state-driven gentrification as part of the Danish housing policy. The Danish Law on public housing seeks to reduce the proportion of public housing in certain designated neighbourhoods, officially labelled “transformation areas”. Under the scheme, authorities may sell buildings to private developers, demolish existing housing, or convert family dwellings into accommodation for young people. These measures may entail the unilateral termination of tenants’ leases, resulting in their forced displacement.

To determine which neighbourhoods are subject to restructuring, an earlier version of the Law on public housing introduced the labels of “ghettos” and “hard ghettos”. These labels have since been replaced by the less openly yet still stigmatising terms “parallel societies and “transformation areas”, without however incurring any substantive changes to the scheme. A neighbourhood classifies as a parallel society when it satisfies at least two out of four socio-economic criteria—relating to unemployment, education levels, criminal convictions, and average gross income—and at least 50 per cent of its residents are “immigrants and their descendants from non-Western countries”. Where such a classification persists for five consecutive years, the area is classified as a “transformation area” and becomes subject to the abovementioned far-reaching restructuring measures. The Law on Public Housing also identifies “vulnerable areas”, neighbourhoods that meet the same socio-economic criteria as parallel societies but do not have a majority of “non-Western” residents. These areas are not eligible for restructuring. Hence, the possibility of restructuring turns explicitly on the “Western”/“non-Western” distinction, making the risk of forced relocation contingent upon the demographic composition of a neighbourhood.

The applicants are residents of designated “transformation areas” in the municipalities of Slagelse and Copenhagen. The applicants from Slagelse challenged the termination of their leases, whereas the applicants from Copenhagen sought invalidation of the ministerial decision approving the development plan for their neighbourhood. Both cases reached the High Court of Eastern Denmark, which referred two preliminary questions to the Court of Justice First, whether the criterion “non-Western immigrants and their descendants” constitutes a distinction based on ethnic origin within the meaning of the Race Equality Directive (RED); and second, whether the Danish legislation gives rise to direct or indirect discrimination.

 

The Judgment

 

The Opinion of the Advocate General (Ćapeta) answered both questions in the affirmative, concluding that the distinction is based on ethnic origin and that the difference in treatment amounts to direct discrimination, although it could also be qualified as indirect discrimination. While the Grand Chamber does not fully adopt the AG’s Opinion, much of its reasoning is reflected in the judgment. The CJEU, first of all, concurs with the AG that a public housing scheme such as that existing in Denmark is capable of falling within the material scope of the RED (§ 66). It then proceeds to examine whether the Danish legislation constitutes direct ethnic discrimination. While leaving it to the referring court to ultimately decide this issue, the Court offers detailed guidelines which allow very little room for a negative answer. Still, in case the referring court should conclude that there has been no direct discrimination, the CJEU also provides guidelines to determine whether there has been indirect discrimination. Here the Court interprets the requirement of Article 2(2)(b) RED that persons of a racial or ethnic origin must be put at a ‘particular disadvantage’, as well as the criteria for objective justification. The Court leaves to the referring court to decide, notably, if the Danish legislation puts persons belonging to certain ethnic groups at a particular disadvantage (§ 143) and, if that is the case, if the restructuring measures are appropriate, necessary and proportionate in relation to the public interest of promoting social cohesion and integration (§§ 164-165).

 

On ethnic origin: a step forward but not quite there yet

A central issue in the judgment is whether the category of “non-Western immigrants and their descendants” falls within the prohibited ground of racial or ethnic origin under the RED. The Court of Justice has long displayed an uneasy relationship with the concepts of “race” and ethnic origin — most notably with the former, which it has traditionally been reluctant to invoke expressly. In CHEZ, the Grand Chamber held that ethnic origin ‘has its origin in the idea of societal groups marked in particular by common nationality, religious faith, language, cultural and traditional origins and backgrounds’ (§ 46). While this definition in itself is not unnecessarily restrictive, its application in subsequent cases demonstrated a very restrained understanding of the concept. In Jyske Finans as well as Land Oberösterreich v KV the Court’s focus was on establishing the presence of objective characteristics rather than identifying social processes of racialisation (here) and othering through which such characteristics gain social meaning and become markers of racialized social hierarchy. Moreover, it was suggested that the requirement of a ‘particular’ disadvantage in Article 2(2)(b) RED meant that only ethnically homogeneous groups could seek protection under the Directive, thus excluding differential treatment of heterogeneously composed groups such as ‘third-country nationals’. This approach has been widely criticised in the literature (for example here, here,and here). Commentators have argued that the Court’s understanding of ethnic origin risked hollowing out the protective scope of EU anti-discrimination law and failed to account for the lived realities of racialised exclusion suffered by migrants in particular.

Despite AG Ćapeta’s invitation (Opinion, § 69) the Court does not expressly recognize ‘ethnic origin’ as a socially constructed category. It confirms, however, that the term ‘racial or ethnic origin’ in the RED must be understood in light of the International Convention on the Elimination of All Forms of Racial Discrimination (ICERD) and of Article 14 European Convention of Human Rights (ECHR) and the case law of the European Court of Human Rights (ECtHR). This confirms the Court’s growing tendency to engage with external sources, especially its Strasbourg counterpart, even as EU accession to the European Convention on Human Rights remains pending. The reference to ECtHR case law is moreover significant as that court has previously espoused a less rigid understanding of ethnicity, notably in the Grand Chamber judgment in Biao v. Denmark where it held that ‘Danes of foreign origin’ were subject to ethnic discrimination. The CJEU also recalls the ECtHR’s consistent qualification of racial discrimination as a ‘particularly invidious form of discrimination which, in view of its perilous consequences, requires […] special vigilance and a vigorous reaction’ (§ 79). It thus makes clear that the issue at stake is one of racial discrimination and that no relevant distinction exists in this regard between the concepts of “race” and ethnic origin. Still, it fails to draw any strong doctrinal consequences from this acknowledgement in the form of a more explicitly social conception of racial discrimination.

A very welcome aspect of the case is the Court’s clarification that ethnic origin—as previously defined in CHEZ—need not be determined by reference to a single or homogeneous ethnic group (§§101-104; §§134-140). Returning to its first judgment on the RED, Feryn (§103), the Court clarifies that the notion of ethnic origin may apply to broadly formulated criteria such as ‘allochtones’, ‘foreigners’ or, indeed, ‘non-Western immigrants’. Still, it continues to insist, as it did in Jyske Finans, that a single characteristic such as nationality or country of birth, can never be indicative of ethnic origin (§ 86). Instead, ethnic origin must always be determined based on a combination of factors. As we argued earlier, this insistence on multiple characteristics denies the reality of racial discrimination in which a single feature (such as someone’s nationality, surname or skin colour, for that matter) may be the reason for differential treatment if that feature functions, in the given context, as a marker of “race” or ethnicity. The Court thus keeps open a significant loophole in EU anti-discrimination law—one that Member States may exploit by designing exclusionary measures around formally non-protected criteria, while shielding them from scrutiny under the prohibition of racial and ethnic discrimination.

In the present case, however, the single characteristic requirement does not stand in the way of a finding of ethnic discrimination. The Court is satisfied that the criterion of ‘immigrants from non-Western countries and their descendants’ is based on a ‘complex combination of criteria’, including country of birth, the nationality and country of birth of the parents and whether those countries of birth and/or nationality are considered, under Danish law, as ‘non-Western’ countries (§ 100).


On direct and indirect discrimination

A distinction based on ethnic origin constitutes direct discrimination if it involves a person being treated less favourable than another in a comparable situation (Art. 2(2)(a) RED). In line with the AG’s Opinion, the Court identifies two types of less favourable treatment that could result from the Danish legislation. The first is that tenants of public family housing in “transformation areas” face a higher chance of having their lease terminated compared to tenants with a similar lease in “vulnerable residential areas”, which are comparable to “transformation areas” in socioeconomic terms but where “non-Western immigrants and their descendants” do not form more than half of the population. It follows from the judgment that the risk of a lease being terminated early is, in itself, sufficient to constitute less favourable treatment, there is no requirement that this risk must have materialised. The Court moreover confirms what it had already decided in CHEZ, namely that there is direct discrimination when less favourable treatment results from a distinction based on ethnic origin, even if among those affected there are people who do not belong to the targeted ethnic group (§ 107). After all, tenants who are not themselves “non-Western” immigrants face the same risk of losing their homes as their “non-Western” neighbours. What is at stake here is a form of discrimination by association, although this concept is not mentioned explicitly by the Court.

Besides the risk of having their leases terminated, the Court notes a second form of harm that the applicants may have suffered, which is stigmatisation. Here the Court shows itself much more deferential to the Danish courts: it is up to the referring court to investigate if ‘the very name “transformation area”, which for the residents of areas classified as such gives rise to an increased risk of early termination of their lease, and which replaced the name ‘hard ghetto area’ is, at national level, offensive and stigmatising’ (§ 126). It is interesting that, at this point, the Court does show consciousness of the fact that the meaning of certain terms is socially constructed and can therefore differ over time and place, in this case leaving the Danish courts in a better position to determine the stigmatising connotations of the notion of “transformation areas”.

After having thus applied a broad definition of direct discrimination, which acknowledges both redistributive and recognitional harm, the Court moves on to examine the possibility of indirect discrimination. According to the Court itself, this analysis is necessary in case ‘the referring court concludes that the national legislation at issue […] does not constitute direct discrimination’ (§ 130). While this may at first sight seem a logical step to take, the Court’s consideration of indirect discrimination becomes less obvious if it is taken into account that the previous part of the judgment leaves very little to no room for the referring court to find that there has not been any direct discrimination. Indeed, once direct discrimination is established, the analysis can end there: discrimination is either direct or indirect, but not both, at least not on the same ground.

This move is troubling. The scheme at issue is plainly not one of indirect discrimination, and introducing this layer of analysis risks generating confusion rather than clarity for national courts. For one, if the referring court concludes that there is no direct discrimination because, after all, the criterion concerning “non-Western immigrants and their descendants” does not result in differential treatment based on ethnic origin, this would raise the question of what would then be the ethnic group (or groups) that could be put at a particular disadvantage, as required by Article 2(2)(b) RED. This is a question on which the judgment remains silent. The risk of confusion is further exacerbated by the fact that the Court’s own conceptualization of indirect discrimination is fraught with difficulties, as illustrated most clearly by its case law on religious symbols and headscarves (see here and here).

On the upside, the judgment does provide valuable clarification as to how the case should be assessed if viewed through the lens of indirect discrimination. This may prove practically significant, as experience shows that even where direct discrimination is formally removed, discriminatory effects often re-emerge—consciously or unconsciously—through ostensibly neutral criteria. One can easily imagine, for instance, that the Danish legislator might in the future abandon the explicit reference to “non-Western immigrants and their descendants”, while continuing to target specific neighbourhoods through socio-economic indicators that would, in practice, place that very group at a particular disadvantage. Some key takeaways from The Court’s analysis are: 1) that the aim of “ensuring successful integration of third-country nationals (TCNs)” is legitimate in principle; 2) the potential of the Danish scheme to actually promote social cohesion and integration is viewed as doubtful, not least because it does not apply to “vulnerable residential areas” that are socioeconomically comparable to “transformation areas” but without a majority “non-Western” population (§§ 162-163); 3) in addition to the requirements of appropriateness and necessity, which are expressly mentioned in Article 2(2)(b) RED, the Court examines the measures’ proportionality stricto sensu, something that has not commonly done before (see here and here); 4) regarding this proportionality stricto sensu, it must be taken into account that the right to respect for the home is a fundamental right protected by the EU Fundamental Rights Charter (Article 7) and that, according to ECtHR case law, the loss of one’s home constitutes “a most extreme form of interference” (§ 170).


Structural discrimination

In sum, there is a clear suggestion that the Danish scheme would also have to be considered as indirectly discriminatory – at least if a particular disadvantage to one or more ethnic groups can be established. The judgment can thus be readily qualified as an important blow to Denmark’s “ghetto” policy. Still, from a perspective of structural equality several problematic aspects remain – we round off this blog by mentioning two of them. The first, already raised by one of us elsewhere, is that the judgment entrenches the existing distinction between EU citizens and third-country nationals whereby only the latter are deemed in need of integration. This becomes evident in the indirect discrimination analysis, where it is pointed out that “integration” is a legitimate public interest only where third-country nationals are concerned (§ 151) – the subtext being that EU citizens are by definition already integrated.

Second, the Court seems to leave open the possibility that, under certain circumstances, “gentrification” measures such as those imposed by Denmark might be proportionate. This suggestion is troubling, given that those measures reflect a discriminatory logic that runs even deeper than the use of stereotypes in legislative documents or the harm suffered by individual tenants who are forced to leave their homes. At the basis of Denmark’s (or perhaps any) integration policy lies the rationality of Goldberg’s ‘racial state’ – the notion that at its core the modern state serves to protect public order by excluding the racialised other. It follows that any proportionality analysis cannot meaningfully abstract from the fact that the policy is premised on the stigmatization of racialised and impoverished communities. Yet is seems that the Court failed to recognise this, unlike in CHEZ where it expressly acknowledged the offensive and stigmatising nature of the practice at issue (§§84 and 108).

In short, there are several positive points to be noted about the judgment. These are, in particular, the Court’s acknowledgement that race discrimination does not turn on the presence of a homogeneous ethnic group, its recognition of the stigmatisation inherent in the so-called ‘ghetto law’ and its engagement with international instruments and proportionality in the indirect discrimination analysis. Still, the Court stops short of addressing the central issue at stake: the systemic and structural character of racialised exclusion. By avoiding a direct confrontation with structural racism, the Court ultimately does not get to the bottom of what makes the Danish ‘ghetto law’ so deeply problematic.

 

Monday, 23 February 2026

Asylum Pact 2.0: the EU amends the rules on ‘safe third countries’ and ‘safe countries of origin’

 


Steve Peers, Professor of Law, Royal Holloway University of London

Photo credit: Paula Reister, via Wikimedia Commons


Introduction

The EU has finally adopted two amendments to the EU’s asylum pact, entailing significantly more stringent rules as regards ‘safe countries’ of origin’ (adopted text here) and particularly ‘safe third countries’ (adopted text here).

This blog post summarises the new laws in their context, and then examines how much they impact upon the ability to claim asylum in the EU in practice – in particular the introduction of a ‘Rwanda clause’ in EU law, mirroring the last UK government’s attempt to designate that country as ‘safe’ even for asylum-seekers who had not travelled through it.


Background

The history of the two ‘safe countries’ concepts in EU law goes back initially to 1992, in the form of the London Resolutions of Member States’ ministers (see here and here). The concepts then appeared in the initial 2005 asylum procedures Directive, since replaced by the current 2013 asylum procedures Directive. The 2013 law will be replaced, as from 12 June 2026, by a new Regulation on asylum procedures, adopted in 2024 (the ‘2024 Regulation’, which I previously discussed here), part of the EU’s asylum pact. While the 2024 Regulation already makes the rules more stringent than in the 2013 Directive, the 2026 amendments now make those rules more restrictive again.

(For earlier blog posts on the 2026 amendments, see my analysis of the Commission proposal here, and my analysis of the Council and Parliament positions).


‘Safe countries of origin’

2024 Regulation

The 2024 asylum pact Regulation retains the definition of ‘safe country of origin’ from the 2013 Directive, referring to a list of human rights standards. As in the 2013 Directive, the rule can only apply if an asylum-seeker is a national of the country concerned, or is a stateless person formerly habitually resident there; and the principle remains a presumption, which the applicant can rebut in individual cases.

However, there are several changes. First of all, the rule will become mandatory for Member States, instead of an option. Secondly, it will be possible to designate a non-EU country of origin as ‘safe’ in part, either geographically (exceptions for ‘specific parts of its territory’) or for some groups of people (‘clearly identifiable categories’). Recent Court of Justice judgments have confirmed that such designations are not possible under the 2013 Directive (see here and here).

Thirdly, the 2024 Regulation makes it possible for the EU to adopt a common list of ‘safe countries of origin’, alongside Member States’ designation of such countries. Countries can be taken off the common EU list in the event of ‘significant changes’ there, initially by means of a delegated act adopted by the Commission, then via legislation. Member States cannot put the country back on their national list of ‘safe countries of origin’ while a delegated act suspending it from the list applies; and for two years after the country is dropped from the common EU list via legislation, Member States need the Commission’s approval to reinstate it on their national lists.

Fourth, in place of a general reference to speeding up consideration of ‘accelerated’ cases (including ‘safe country of origin’ cases) in the 2013 Directive, there will be a three-month deadline to decide on such cases, although if an authority considers the asylum application to be too complex, it can default to the usual six-month deadline to decide on the merits. (This general deadline has not been amended, although the possibilities of extending it have been partly curtailed).

Fifth, a new ground for accelerated procedures, overlapping with the ‘safe third country’ ground, has been added: Member States must also fast-track cases where the most recent annual recognition rate for asylum claims for the asylum-seekers’ nationality is below 20% EU-wide at first instance (ie before appeals), ‘unless the determining authority assesses that a significant change has occurred in the third country concerned since the publication of the relevant Eurostat data or that the applicant belongs to a category of persons for whom the proportion of 20 % or lower cannot be considered to be representative for their protection needs, taking into account, inter alia, the significant differences between first instance and final decisions.’ It is also mandatory, and subject to the same deadlines.

Sixth, while the ‘safe country of origin’ concept remains a ground for applying special border procedures, as under the 2013 Directive, those border procedures have been changed. They are now mandatory in principle (for a specific number of asylum-seekers per Member State) and can now last for up to twelve weeks (including both the administrative and judicial phase of the case), instead of four weeks. Presumably the Court of Justice case law, requiring that asylum-seekers be released from detention (if that was the only ground for it) and admitted on to the territory at the end of the border procedure deadline, still applies. Unaccompanied minors will be exempt from the revised border procedure, unless they are ‘national security’ or ‘public order’ risks.

Finally, it will be mandatory, instead of optional, for Member States to deny automatic suspensive effect to appeals in ‘safe country of origin’ cases, and several safeguards related to the lack of automatic suspensive effect are dropped. However, it will still be possible for asylum-seekers to ask a court for permission to stay during the appeal. On appeals, there is nothing to suggest that the case law of the Court of Justice on the 2013 Directive (Alace) has been overturned. In that case, the Court said that: a Member State’s designation of a ‘safe country of origin’ has to be subject to judicial review; the sources of information used for the designation must be provided to the applicant and to courts; and courts must be able to examine other sources of information.

2026 amendments

First of all, the new Regulation allows (as an option) the early application (ie before June 2026, from the day after publication of the new law in the EU Official Journal) of the possibility of partial designation of a country as a ‘safe country of origin’ (either geographically or for certain groups of asylum-seekers), as well as the new ‘20% recognition rate’ rule. In particular, this is intended to allow for the earlier application of the Italy/Albania arrangements on processing asylum-seekers on Albanian territory, which were in effect frustrated by the Court of Justice’s interpretation of the current Directive. Of course, other Member States might have a use for earlier application of the rules too.

Furthermore, the 2026 amendments similarly allow for early application of the partial designation of countries as ‘safe third countries’.

The early application of the 2024 version of these rules can also be applied at the borders, in the current version of the border procedure. This means that, among other things, the 2024 exclusion of most unaccompanied minors from the border procedure will not apply until the 2024 Regulation fully enters into force in June.

Secondly, the 2026 Regulation sets out a common EU list of ‘safe countries of origin’, as from the entry into force of the Pact: Bangladesh, Colombia, Egypt, India, Kosovo, Morocco and Tunisia, plus candidates for EU accession (Serbia, Montenegro, Albania, North Macedonia, Bosnia-Herzegovina, Turkey, Ukraine, Moldova and Georgia).

The accession candidates are, however, subject to special rules: they cannot be regarded as ‘safe’ if they have been subjected to EU foreign policy sanctions due to their actions ‘affecting fundamental rights and freedoms that are relevant to’ the definition of ‘safe country of origin’; or if their recognition rate is above 20% at first instance; or if there is a ‘serious threat to a civilian’s life or person by reason of indiscriminate violence’ in that country (which is an amended version of one of the grounds for ‘subsidiary protection’ in EU law; the requirement that the threat also be ‘individual’ does not apply in this context). The preamble draws the obvious conclusion that ‘Member States should not apply the concept of safe country of origin to applicants from a candidate country during the period in which the circumstances provided for by this Regulation persist.’ 

It appears that in place of the usual rules on suspending a country from the list, the accession candidates will be removed following a notification from the Commission to the Council, European Parliament and Member States. However, to trigger the ‘serious threat’ ground for removal from the list, the Commission will need approval from the Council first. Presumably the default voting rule of qualified majority in the Council will apply. According to the preamble, the Council’s role is necessary ‘in view of the potential implications for the external relations of the Union and the Member States’ in this scenario.

Several of these points (dropping the ‘individual’ criterion, qualifying the sanctions point, and creating a role for the Council) were changes from the Commission’s original proposal, made at the Council’s behest. The European Parliament’s proposed amendment (removing a country from the list if the temporary protection Directive applied – obviously referring to Ukraine) was not agreed, although the final version of the preamble says that the application (or termination) of temporary protection under that Directive should be taken into account when assessing whether a ‘serious threat’ exists. Moreover, the Parliament’s proposal to use delegated acts (ie the usual process) to remove candidate countries from the common list was not agreed either.

The Regulation also now provides for the possibility (again, at the Council’s behest) that a country on the common EU list can be de-listed in part, ie using the exceptions for parts of a country or groups of people which Member States can use.

Finally, at the behest of the Parliament, a clause in the preamble to the 2026 amendment appears to hint at limiting judicial review in individual cases:

[a ‘safe country of origin’] designation reflects the general situation in that country and is not affected by the individual circumstances, which may only be assessed for the purposes of determining whether the concept of safe country of origin should exceptionally not be applied in a specific case. Therefore, in the context of national judicial review, the detailed evidence regarding an applicant’s individual situation justifying the applicability of the concept of safe country of origin should be the main purpose of that assessment. In accordance with the Treaties, the Court of Justice of the European Union is competent to rule on any doubts on the validity of a designation of a third country as a safe country of origin at Union level.

This is slightly more equivocal than the EP’s proposed amendment on this point, which had read:

…national judicial review should examine the detailed evidence regarding an applicant’s individual situation justifying, in his or her case, the inapplicability of the concept of safe country of origin and not the designation as such.


‘Safe third countries’

2024 Regulation

The 2024 Regulation (before the 2026 amendments) makes a number of changes to the rules on this issue as compared to the 2013 Directive. First of all, it will be simpler to designate a country as a ‘safe third country’ in one respect: there need not be an opportunity to request and receive Refugee Convention refugee status there. Instead, if a non-EU country ‘has ratified and respects’ the Convention, within the limits of any reservations and limitations, that country ‘shall be considered to ensure effective protection’; but if has not ratified the Convention, or applies a geographical limit to it (ie Turkey), that country ‘only’ offers effective protection for people where, ‘as a minimum’, that country allows people to remain, offers subsistence, health care and education, and ‘effective protection remains available until a durable solution can be found’.  

Secondly, as with ‘safe countries of origin’, it will now be possible to regard a country as a ‘safe third country’ only partly, ie ‘with exceptions for specific parts of its territory or clearly identifiable categories of persons’. A country can even be a ‘safe third country’ for an individual applicant.

Thirdly, there will be new safeguards for unaccompanied minors: a ‘best interests’ test and a prior assurance that the non-EU country will take them in charge and give them immediate access to effective protection.

Fourth, there will now be a two-month deadline for deciding ‘safe third country’ applications (and most other inadmissible applications), with a possible extension of up to two months in certain circumstances.

Fifth, there will also now be a new prospect of a common EU list, which works the same way as the possible common list of ‘safe countries of origin’ (and removals from it) described above.

Sixth, the revised border procedure in the Regulation (described above), will also apply to ‘safe third country’ cases.

Several elements have not changed, though. It is still necessary that, in the country concerned, there is: no threat to [life or liberty’ on any of the grounds set out in the Refugee Convention; ‘no risk of serious harm’ as defined in EU law as regards subsidiary protection; respect of the non-refoulement principle in the Refugee Convention (ie not sending the asylum seeker to an unsafe country); and respect for ‘the prohibition of removal, in violation of the right to freedom from torture and cruel, inhuman or degrading treatment as laid down in international law’.

Also, the principle will still only be optional for Member States; and the rule still cannot be applied unless the asylum-seeker will be admitted to the country concerned. Also, the rule can only apply if the asylum-seeker has a ‘connection’ with the other country, ‘on the basis of which it would be reasonable’ for the asylum-seeker to go there; it must still be possible for the asylum-seeker to challenge the alleged safety of the country concerned for them, as well as their supposed connection with it. Finally, as before, appeals in ‘safe third country’ cases will still have automatic suspensive effect.

2026 amendments

First and foremost, the new Regulation (which will apply from June, along with the rest of the asylum pact) has dropped the requirement that the ‘safe third country’ principle can only be applied in the event of a ‘connection’ to the supposedly ‘safe’ third country. There are further explanations of that principle in the preamble:  

Member States should, while fully respecting the parameters set out in the case law of the Court of Justice of the European Union, be able to apply the concept of safe third country on the basis of a connection as defined in conformity with national law or practice, in so far as specifically defined therein. The connection between the applicant and the third country could be considered established in particular where members of the applicant’s family are present in that third country, where the applicant has settled or stayed in that third country, or where the applicant has linguistic, cultural or other similar ties with that third country.

It will now also be possible to apply the ‘safe third country’ principle in two other cases. First, it can be applied where the asylum seeker had transited through that country on the way to the EU, as further explained in the preamble:

…transit through a third country could include a situation where an applicant has passed through or stayed on the territory of a third country on the way to the Union, or where the applicant has been at the border or in a transit zone of a third country, where that applicant has had the possibility to request effective protection from the authorities of the third country concerned.

Secondly, the ‘safe third country’ principle can now apply to a Rwanda-type deal with a country that the asylum-seeker had neither a connection with nor transited through: ‘there is an agreement or an arrangement [with the non-EU country concerned]…requiring the examination of the merits of any requests for effective protection made in the third country concerned by applicants subject to that agreement or arrangement’.

Such an agreement or arrangement can either be negotiated by the EU or one or more Member States; the Member States might even negotiate it alongside non-EU countries (say the UK, or the US?). There are detailed rules on how any EU negotiations relate to Member State treaties on the issue.

However, the new ‘Rwanda’ criterion will not apply to unaccompanied minors, although the Parliament had argued that it should apply to children who were security threats. The new transit criterion will apply to unaccompanied minors though, although the current safeguards for them in ‘connection’ cases (best interests of the child, certain assurances from the country concerned) in the 2024 Regulation will apply to ‘transit’ cases too. The preamble also refers to considering the family unity principle.

Finally, the new Regulation has altered the rules on appeals, so that there will no longer be automatic suspensive effect in ‘safe third country’ appeals, except for unaccompanied minors subject to the border procedure. The amendments also (at the Council’s behest) drop automatic suspensive effect in cases where the asylum seeker has international protection from another Member State, even though Court of Justice case law provides for the possibility of applying for international protection in another Member State to avoid harsh conditions in the Member State which granted such protection, where the latter Member State treats the beneficiaries of international protection so badly that it amounts to a breach of the EU Charter of Fundamental Rights. Nevertheless, it will still be possible to request a court to grant suspensive effect.

Other features of the ‘safe third country’ principle have not changed: it will remain optional for Member States, and the rules will not change as regards criteria defining the ‘safety’ of a country as such, the applicable deadlines, the related rules on the border procedure, or the requirement that Member States must consider the merits if the third country concerned refuses to admit the asylum seeker. Nor does it use the power to adopt a common EU list of ‘safe third countries’.

 

Assessment

The ‘safe country of origin’ amendments in particular are partly aimed at overturning case law of the CJEU, not only as regards allowing geographical and group exceptions from the principle, but also (on a more lasting basis, given that those exceptions would be allowed from June anyway) as regards limiting national judicial review of the national lists of such countries.

However, this apparent attempt to limit judicial review would, if the provision is interpreted as such, breach the EU Charter of Fundamental Rights. The Court of Justice’s Alace judgment refers several times to Article 47 of the Charter (the right to an effective remedy and a fair trial), when ruling that it must be possible to challenge the designation of countries as ‘safe countries of origin’ (underlining and bold text added):

…it should be noted that the obligation imposed on Member States by Article 46(1) of Directive 2013/32 to provide for a right to an effective judicial remedy for applicants for international protection, the scope of which is defined in Article 46(3) of that directive, corresponds to the right to an effective remedy guaranteed by Article 47 of the Charter… [para 53]

…although, in the absence of EU rules on the matter, it is for the national legal order, in accordance with the principle of procedural autonomy of Member States and subject to the observance of the principles of equivalence and effectiveness, to lay down the detailed procedural rules governing remedies for ensuring that individual rights derived from the EU legal order are safeguarded, Member States nevertheless have the responsibility to ensure observance in every case of the right to effective judicial protection of those rights as guaranteed by Article 47 of the Charter, the scope of that right being clarified, in the present case, by Article 46 of Directive 2013/32… [para 64]

…the choice, by a Member State, of the competent authority and the legal instrument effecting the designation, at national level, of safe countries of origin, in accordance with Articles 36 and 37 of Directive 2013/32, cannot affect its obligations under that directive. It is thus for each Member State, inter alia, to ensure respect for the right to an effective judicial remedy which Article 46(1) of that directive confers on applicants for international protection against decisions taken on their applications, the scope of which is defined by Article 46(3) of that directive. [para 65]

In that regard, the Court has held that, in accordance with Article 46(3) of Directive 2013/32, read in the light of Article 47 of the Charter, where an action is brought before a national court or tribunal against a decision taken on an application for international protection – examined in the context of the special scheme applicable to applications lodged by applicants from third countries designated, in accordance with Article 37 of that directive, as safe countries of origin – that court or tribunal must, as part of the full and ex nunc examination required by Article 46(3) of that directive, raise, on the basis of the information in the file and the information brought to its attention during the proceedings before it, a failure to have regard to the material conditions for such designation, set out in Annex I to that directive… [para 66]

Consequently, and having regard to the case-law referred to in paragraphs 62 and 63 above, the fact that a Member State has chosen to designate safe countries of origin by means of a legislative act cannot be such as to preclude the national court or tribunal seised in the circumstances set out in the preceding paragraph of the present judgment from reviewing, even if only indirectly, whether the designation of the third country in question as a safe country of origin complies with the material conditions for such a designation, set out in Annex I to Directive 2013/32. [para 67]

In the light of the foregoing, the answer to the first questions is that Articles 36 and 37 and Article 46(3) of Directive 2013/32, read in the light of Article 47 of the Charter, must be interpreted as not precluding a Member State from designating third countries as safe countries of origin by means of a legislative act, provided that that designation can be subject to judicial review as regards compliance with the material conditions for such a designation, set out in Annex I to that directive, by any national court or tribunal hearing an action brought against a decision taken on an application for international protection, which had been examined under the special scheme applicable to applications lodged by applicants who are from third countries designated as safe countries of origin. [para 68]

As the Charter has the same legal value as the Treaties (Article 6 TEU), any EU law adopted in breach of it would be invalid.

In any event, the litigation has not stood still. There are further cases referred from the Italian courts as regards aspects of the Italy/Albania arrangements, including Sedrata on detention issues, and Comeri (fast-tracked by the Court) on the external relations aspects of such arrangements. These pending cases may limit the possible application of the Italy/Albania agreement in the meantime, although it remains to be seen whether the changes to the law (including the 2024 version of the reception directive), might be relevant to the issues in these disputes post June. Of course, further legal challenges can be expected as the 2024 Regulation is partly rolled out early on the basis of the 2026 amendments, and applied fully from June.

The removal of automatic suspensive effect from ‘safe third country’ cases, while widening the scope of such cases in principle, may mean that it is harder to obtain an effective remedy in such cases – although the possibility of requesting suspensive effect from the national court will now become crucial. The same will be true of cases where a Member State is allegedly breaching the Charter rights of recognised refugees – an issue separate from ‘safe third (non-EU) countries’ where the legislators decided to carry out a ‘drive-by hit’ on appeal rights nonetheless.

As for the ‘Rwanda clause’ itself, at first sight it leaves open the possibility that (apart from unaccompanied minors) literally everyone who applies for asylum in the EU could be removed to such a country. In practice, though, this depends upon there being enough countries willing to admit the asylum-seekers. The EU’s current and planned leverage (via the visa code, recent visa waiver suspension rules, agreed amendments to trade policy laws, proposed changes to development aid and planned further changes to the visa code) relates to countries taking back their own citizens and those who transited through them. While the transit point is relevant to the widening of the ‘safe third country’ concept to include transit states, it does not cover the introduction of the Rwanda clause.

Even for those who are subject to the wider application of the ‘safe third country’ principle after the 2026 amendments, it will still be necessary to show that they will be admitted to the country concerned, and still possible to argue that the country is ‘unsafe’ in their particular situation. Moreover, it could be argued that – in parallel with the CJEU case law on challenging the listing of ‘safe countries of origin’, quoted from above – the EU Charter of Rights necessarily confers the possibility of challenging the listing of the country in question itself, comparable to the UK Supreme Court ruling that Rwanda was ‘unsafe’ as such. And since the EU Charter has a higher legal status than EU legislation, it would not be possible for the EU or its Member States to overturn such a ruling merely by passing legislation – unlike the ‘Safety’ of Rwanda Act passed under the previous UK government.