Safe for whom?

In his Opinion of 10 April 2025 in the joined cases C-758/24 and C-759/24, Advocate General (AG) Jean Richard de la Tour argues that a third country may qualify as a safe country of origin (SCO) even if it is unsafe for certain groups. In doing so, he revives a reasoning the ECJ had rejected in its ruling of 4 October 2024 (CV, C-406/22). This interpretative shift, embedded in the framework of the New Pact on Migration and Asylum, risks undermining essential safeguards for asylum seekers.

The AG’s interpretation of the revised Asylum Procedures Directive 2013/32 (APD) allows for partial SCO designations. This contradicts both the Directive’s protective purpose and the ECJ’s categorical rejection of such exceptions. This way, it opens the door to arbitrary exclusions based on personal characteristics. By the same logic, such personal exceptions should be considered impermissible under EU law. While Rocca has addressed the Opinion’s human rights implications, this piece focuses on partial SCO designations and the Directive’s intended safeguards, structure, and language.

Safe countries of origin and asylum seekers: procedural shortcuts, fundamental risks

The idea that some countries are “safe” is central to the EU asylum procedures. But who decides what qualifies as safe? This question has long preoccupied both international and EU refugee law. Guy Serle Goodwin-Gill once gave it sharp form: “Who then is to say that countries are safe? And by whose standards? Secret men in secret rooms reading secret memos? No.” (para. 2, Opinion).

In EU asylum law, the concept of SCO traces back to the London Resolutions 1992. It was later formalised in the APD 2005/85 and continues to shape asylum procedures under the revised APD 2013/32. The Directive sets out common rules for granting and withdrawing international protection (refugee status or subsidiary protection status) across EU Member States.

Articles 36, 37 and Annex I of the APD regulate the SCO mechanism. They allow Member States to subject applicants who are nationals of – or, in the case of stateless persons, were formerly habitually resident in – designated safe countries of origin to an exceptional examination regime. These procedures are frequently accelerated and may be carried out at borders or in transit zones.

This regime is built on a rebuttable presumption: applicants from SCOs are considered to be sufficiently protected against persecution and serious harm. On this basis, Member States may fast-track their applications.

However, these accelerated procedures come at a high cost. Compared to the regular asylum procedure, they drastically shorten appeal deadlines, permit detention at borders or in transit zones, and often remove the automatic suspensive effect of appeals, meaning a negative decision may be enforced even while under review. The procedural shortcuts pose serious risks to applicants’ fundamental rights: tight deadlines restrict access to legal advice, and the lack of suspensive effect increases the risk of refoulement, where someone is returned to a country where they may face persecution. The widespread use of detention also raises concerns under the right to an effective remedy under Article 47 of the Charter of Fundamental Rights of the European Union (CFR) and Article 6 of the European Convention on Human Rights (ECHR).

Can Italy’s designation of Bangladesh as a safe country of origin stand?

The recent Italy-Albania protocol, though not initiated by the EU, introduces a new element to European asylum practice. It allows for the transfer of asylum seekers – processed under the accelerated border procedure – to detention facilities in Albania, which are treated as EU border or transit zones. Although located outside Italian territory, these procedures fall under Italian jurisdiction. Accordingly, the EU asylum law, including the APD, remains fully applicable.

The protocol brings the practical meaning of “safety” under EU law back into focus: it matters who decides – and which benchmarks they use.

This legal dilemma is central to the joined cases LP (C-758/24) and LC (C-759/24). On 16 October 2024, two Bangladeshi nationals, rescued at sea and detained in Gjadër, Albania, applied for international protection. Just weeks earlier, Bangladesh had been designated as a SCO under Legislative Decree No. 158/2024. As a result, the Territorial Commission of Rome subjected them to the accelerated border procedure (para. 16-17, Opinion).

Unlike its predecessor, the new decree no longer discloses the sources used to assess Bangladesh’s safety, raising concerns, particularly in light of the ECJ’s ruling in CV, C-406/22, just ten days earlier, on 4 October 2024. In that decision, the Court held that EU law does not permit Member States to designate a country only partially as a SCO. The required safety criteria must be fulfilled across the state’s entire territory. Italy responded by revising its list, removing countries with regional protection gaps but retaining those deemed unsafe for specific groups (paras. 26-27, Opinion).

The day after the applicants submitted their request, the Territorial Commission of Rome rejected both claims, citing the applicants’ failure to rebut the presumption of safety attached to Bangladesh. In response, the Tribunale Ordinario di Roma (Rome District Court) referred four preliminary questions to the ECJ on how Member States should interpret and apply the SCO provisions of the APD (para. 18, Opinion).

Written observations were submitted by 17 Member States and the European Commission, with the first three questions focusing on Articles 36, 37, and 46 APD, as well as Article 47 CFR (paras. 4–5; 22, Opinion).

Can information on safe country designations be kept a secret?

One key question raised by the referring court is whether a Member State can designate a third country as a SCO through a legislative act without disclosing its sources.

The court stresses that even where designations are issued by legislative act, this must not shield them from judicial scrutiny. Applicants and national courts must be able to challenge the presumption of safety effectively. In particular, judicial authorities must be in a position to examine the sources on which the designation is based, especially where these are not disclosed, ensuring an up-to-date evaluation of the individual case (para. 4, Opinion).

AG de la Tour agrees: legislative acts are permissible, but they must remain subject to judicial review if contested. Shielding them (Samba Diouf, C‑69/10) would undermine the effectiveness of the Directive (Article 288 (3) Treaty on the Functioning of the European Union), particularly the right to an effective remedy under Article 46(3) APD. Since such designations implement EU law, they must adhere to the fundamental principles and safeguards set out in Chapter II of APD 2013/32, as well as ensure compliance with the right to an effective judicial remedy (Article 46 APD, Article 47 CFR and paras. 37-39; 47 Opinion).

Even where sources are not disclosed, national courts may still rely on Article 37(3) APD, which allows them to gather independent information in accordance with the criteria in Annex I (para. 64, Opinion).

Can safe countries of origin only be partially safe?

The fourth preliminary question referred to in Alace and Canpelli addresses whether a Member State may designate a third country as a “SCO” under the APD even when certain groups within its population face persecution or serious harm. This calls for a closer reading of Articles 36 and 37(1) APD and Annex I, especially for countries like Bangladesh, where structural risks persist for specific communities (para. 66, Opinion).

AG de la Tour proposes that such partial designations are compatible with the APD. Member States, he suggests, may designate countries as SCOs while explicitly excluding certain groups from the presumption of safety. He identifies two cumulative conditions. First, the country’s legal and political situation must broadly reflect a functioning democracy in which the general population enjoys sustained protection from persecution or serious harm. Second, the national legislation must explicitly state that the SCO designation does not apply to groups known to be at risk (para. 95, Opinion).

Yet, this raises significant legal and practical concerns. Using “democracy” as a proxy for protection overlooks the systemic vulnerability of minorities, political dissidents, or survivors of gender-based violence – groups who often face targeted harm even in states that appear stable on the surface. In CV, the Court rejected such generalisations: safety must be assured throughout the territory. Where this is not the case, the presumption fails.

In practice, moreover, group-based exceptions are difficult to operationalise. Accelerated procedures often lack the administrative capacity and legal safeguards to make nuanced assessments, particularly in cases involving intersectional harms. If individuals are wrongly subjected to the presumption of safety, they may face limited access to legal aid, restricted appeals, and inadequate interviews, heightening the risk of refoulement and undermining Articles 33 of the Refugee Convention and 47 CFR.

As the Court is examining the joined cases under the accelerated preliminary ruling procedure (Article 105 of the Rules of Procedure), a decision can be expected soon.

When safety becomes strategy

Partial designation was explicitly allowed under the 2005 APD (Article 30) but was deliberately removed from the recast Directive. As the Court has made clear that this deletion was intentional. It reflects a conscious legislative decision to strengthen legal certainty and ensure that protection needs are assessed in a holistic and individualised manner. Any return to partial designation, the Court emphasised, must be reintroduced by the EU legislator (CV, paras. 73, 81).

Nevertheless, AG de la Tour identifies room for flexibility. He argues that the concept of “generally” safe countries in Annex I permits partial designation. He also points to discrepancies between different language versions of the Directive to support a more open reading (para. 75-79, Opinion). His argument, however, is less grounded in legal necessity than policy logic. Partial designation, he suggests, could help national authorities manage asylum caseloads more efficiently by filtering out applicants deemed unlikely to require protection and reallocating resources to others (para. 85, Opinion).

This reframes the SCO mechanism into a tool of administrative expediency by reframing safety not as a condition for protection, but as a filter for efficiency. It shifts the focus from protection to efficiency – an approach the Court firmly rejected in CV. The presumption of safety, it held, must apply across the entire territory. Otherwise, the Directive’s purpose is eroded, and the risk of wrongful rejections increases (CV, para. 81).

When safety is redefined: legal tension on the horizon

The AG’s Opinion offers more than a legal reading. It gestures toward a future policy shift. The forthcoming  Asylum Procedures Regulation (APR), a cornerstone of the New Pact on Migration and Asylum, explicitly reintroduces the possibility of partial designation (para. 94, Opinion). This signals a shift in EU asylum policy: away from an individual rights-based model toward a more management-oriented approach.

Yet under current law, such flexibility remains impermissible. Until the APR enters into force, Member States are bound by stricter standards of the APD 2013/32. Partial designation remains incompatible with EU law.

Whether the Court will reaffirm the existing framework or lean into the policy direction of the Pact remains open. But for asylum seekers, the implications are immediate: efficiency may be rising on the EU’s agenda, but the cost may be borne by those most in need of protection.

When safety is shipped abroad: the EU’s external turn

The APR entered into force on 11 June 2024 and is scheduled to apply from June 2026. However, the Commission’s recent push to fast-track key provisions, especially the EU-wide list of SCOs, signals a shift in priorities.

This trajectory mirrors the AG’s reasoning in Alace and Canpelli, which departs from the ECJ’s protection-focused stance in CV. While the proposal claims to affirm fundamental rights, its operational logic tells a different story: Bangladesh, for instance, is expected to appear on the EU SCO list despite reports of torture and repression. That the applicants in Alace and Canpelli are Bangladeshi underscores what is at stake.

These developments must be read in the context of the EU’s broader turn towards externalisation. Asylum responsibilities are being relocated, but legal responsibility remains within the EU’s jurisdiction, as illustrated by the Italy–Albania protocol. One might ask whether the Commission’s proposal is intended to legitimise such arrangements. If key provisions of the APR  are applied ahead of schedule, this could render the fourth preliminary question in Alace and Canpelli. At the same time, the very need to fast-track Article 61(2) APR suggests a contradiction of the AG’s position: if early implementation is necessary, it implies that the current APD does not allow for partial SCO designations.