Rechercher
Menu
Je me connecte

Institut pour la recherche interdisciplinaire en sciences juridiques et criminologiques

Centre Charles De Visscher pour le droit international et européen

Rechercher

Institut pour la recherche interdisciplinaire en sciences juridiques et criminologiques

Centre Charles De Visscher pour le droit international et européen

CJEU, Comeri, Joined Cases C-706/25 and C-707/25, Opinion of Advocate General Medina Delivered on 11 June 2026

cedie | Louvain-la-Neuve

cedie
14 September 2026 , modifié le 17 September 2026

From Compatibility to Exclusive External Competence: Advocate General Medina’s Opinion in Comeri and Sidilli

Italy – Albania Protocol – Exclusive External Competence – Common European Asylum System – Extraterritorial Detention – Territorial Lease – Migration Externalisation.

Advocate General Medina’s Opinion in the joined Comeri and Sidilli cases advances the litigation generated by the Italy–Albania Protocol of 6 November 2023 into a new phase. In Sedrata (C-414/25), Advocate General Emiliou considered whether the extraterritorial transfer and detention of third-country nationals subject to return procedures were compatible with the Return Directive (Directive 2008/115). He also examined the legal framework applicable where a person detained in that context lodges an application for international protection, in particular Article 9(1) of the Asylum Procedures Directive (Directive 2013/32), read in conjunction with Article 2(p) thereof. The questions referred by the Corte d’Appello di Roma in Comeri and Sidilli raise a logically prior issue: whether Italy was competent to conclude an international agreement and adopt the measures necessary to implement it in a field substantially governed by EU law. The Opinion therefore addresses not only the compatibility of a particular detention regime with EU asylum law, but also the division of external competences between the European Union and the Member States. Its broader significance lies in the relationship it establishes between the externalisation of migration control and the Union’s exclusive external competence under Article 3(2) TFEU.

The most distinctive feature of Medina’s reasoning is methodological and sequential. The Opinion proceeds in two steps: it first determines whether the Union has exclusive external competence to conclude the Protocol, and only then assesses how the agreement affects EU asylum law. Within this second step, it distinguishes between the mere geographical relocation of detention facilities (which remains within Member State discretion) and the practical conditions governing the exercise of rights under EU asylum law, which may impinge on harmonised guarantees.

Fatima Zahra El harch

A. Facts of the Case and Conclusions

1. Facts of the Case

The reference arose in proceedings concerning two Moroccan nationals who, subject to removal decisions and initially detained in Italy, were transferred to facilities in Albania under the Protocol and the relevant Italian implementing legislation. Once in Albania, they lodged applications for international protection. The applications were considered to have been made with the purpose of delaying or frustrating removal, and the Italian authorities subsequently adopted new detention measures under the legislation implementing the Common European Asylum System (CEAS).

The characterisation of those measures matters. In the Advocate General’s view, the measures submitted for judicial validation were not a continuation of the applicants’ earlier detention for removal purposes; they were new and autonomous. Once the applicants had applied for international protection, their position came within the CEAS: the application suspended the return procedure and brought the applicants within the scope of the guarantees applicable to asylum seekers. It is this characterisation that engages Directives 2013/32 and 2013/33 and provides the basis for the subsequent competence analysis.

2. The Questions Referred and the Advocate General’s Conclusions

The referring court raised two questions. The first concerned whether the conclusion of the Protocol was compatible with the division of competences between the Union and the Member States, in particular in light of Articles 3(2) and 216(1) TFEU and the principle of sincere cooperation. The second, raised in the alternative, concerned the compatibility of the regime operated in the Albanian facilities with the guarantees laid down by the relevant EU asylum directives, including those concerning detention and effective judicial protection. The Advocate General also addressed questions relating to family visits and healthcare. These aspects, however, were considered hypothetical or addressed only for completeness and fall outside the scope of this comment. 

Before turning to competence, the Opinion addresses admissibility. The prescribed period for validation had expired without a decision, and the applicants had already been released before the hearing. The Advocate General nevertheless accepted that the reference remained admissible, on the basis of the referring court’s assertion that it was still required to rule on the lawfulness of the detention. The reasoning is notable because the Opinion otherwise rejects broader arguments based on the short duration of the measures or their potential repetition. It thus leaves a tension between a strict approach to the conditions for preliminary references and the effective judicial protection of individuals subjected to short-term deprivation of liberty. If detention measures systematically expire before judicial scrutiny can occur, strict application of the admissibility requirements may place an entire category of measures beyond effective preliminary review. The point is especially relevant where the practical effectiveness of judicial protection forms part of the substantive analysis.

On the external competence, the Advocate General grounds exclusivity in Article 3(2) TFEU, which confers exclusive Union competence where the conclusion of an international agreement may affect common rules or alter their scope. The provision codifies the ERTA doctrine and, in the Opinion’s view, supplies the relevant test for exclusivity. Article 216(1) TFEU concerns the existence of external competence; Article 3(2) determines when that competence is exclusive. The reasoning follows the logic of pre-emption (see Arena). Exclusivity may arise where the Union has fully occupied a field, or where a national measure may affect common rules or alter their scope. The extent and degree of harmonisation therefore become decisive: minimum harmonisation leaves Member States a margin of action, whereas full harmonisation may preclude independent external action in the harmonised area.

Applying this approach to detention, the Advocate General distinguishes three elements: the geographical location of detention, the grounds justifying detention, and the material and procedural conditions in which detention is implemented. The first two do not, in her view, encroach upon exclusive Union competence. EU secondary law does not prescribe where detention facilities must be located, so the choice of location remains within Member State discretion. The grounds for detention in the relevant directive are exhaustively harmonised, but the grounds relied upon in the Italian framework correspond to those permitted by EU law. The Protocol therefore supplies the geographical and organisational framework without creating new grounds for detention. By contrast, the Advocate General concludes that deficiencies in the material and procedural conditions of detention, particularly as regards effective judicial protection and the guarantee of immediate release, may affect common EU rules and thus fall within the Union’s exclusive external competence. The resulting proposition carries considerable weight: externalisation is not, in itself, prohibited by EU law. Member States may relocate the administrative setting in which they exercise their powers, but they cannot use relocation to modify the legal regime governing that exercise.

It is in relation to these material and procedural conditions that the Advocate General identifies a possible effect on common rules. Her starting point is that harmonised guarantees cannot be reduced to the formal existence of a legal basis; their practical exercise must also be secured. This becomes particularly important where rights operate in a context structurally different from the national territory. Several shortcomings are identified in relation to effective judicial protection: the absence of an express guarantee of confidentiality in communications between applicants and their lawyers in the audiovisual setting, the predominance of remote participation over physical access by counsel, the ceiling placed on reimbursement of lawyers’ travel expenses, and the absence of an express provision concerning interpretation. A further shortcoming, distinct from those affecting judicial protection, concerns release from detention itself. Once the legal basis for detention has fallen away, the individual may nevertheless remain unable to leave Albania independently or return to Italy, instead being subjected to a predetermined transfer route under escort. Drawing on the Guzzardi v. Italy logic, the Advocate General characterises this situation as de facto detention for the purposes of EU asylum law and the Charter. The absence of arrangements ensuring effective and immediate release therefore deprives the relevant guarantee of practical effect. The point exposes a difficulty inherent in offshoring: the geographical relocation of detention can itself create conditions in which a formally recognised right becomes practically ineffective.

B. Discussion

The discussion unfolds in two stages. First, it examines the territorial premise underpinning the entire analysis: whether, and on what doctrinal basis, EU asylum law applies to detention carried out in Albania, a question the Advocate General resolves through the concept of a territorial lease. Second, it addresses the methodological difficulty at the heart of the Opinion: when the deficient practical effectiveness of harmonised guarantees shifts from being a matter of compatibility or implementation to one of exclusive external competence under Article 3(2) TFEU.

1. Territorial Lease and the Extraterritorial Application of EU Law

This conclusion, however, presupposes that EU asylum law applies to the detention of applicants for international protection in the areas concerned, which are situated in Albania. The Governments of Italy and the Netherlands and the Commission argued that applications for international protection made there fell outside the territorial scope of Directives 2013/32 and 2013/33, since the facilities were comparable to diplomatic representations of the Member States, which are expressly excluded from the scope of those directives. The Advocate General rejects that argument through the concept of a ‘territorial lease’ drawing on public international law (see Strauss). Under the Protocol, Albania recognises Italy’s effective and exclusive exercise of state prerogatives over the relevant areas while retaining residual sovereignty. The Opinion identifies three elements supporting this characterisation: a fixed and renewable term, Italy’s financial contribution to the construction and operation of the facilities, and the exclusive exercise of public powers by Italy. 

On this basis, the reference to the ‘territory of a Member State’ in the directives is read functionally, encompassing geographical areas over which a Member State effectively exercises jurisdiction by virtue of a title recognised in international law. The underlying rationale is one of anti-circumvention: a Member State cannot avoid EU obligations merely by transferring the exercise of its powers to a third country through an international agreement. The reasoning recalls the Strasbourg Court’s approach to extraterritorial jurisdiction, in particular the principle that the exercise of public authority outside national territory does not, by itself, release a State from its human-rights obligations (see Hirsi). The central question becomes not simply where the individual is physically located, but who exercises public power over that individual. There is, however, a slight tension in the Opinion’s formulation. At one point, the Advocate General considers it unnecessary to classify the Albanian facilities as Italian ‘territory’, while the operative conclusion refers to the ‘territory of a Member State’ (pp. 71–72). The precise basis of applicability therefore remains somewhat unclear: does EU law apply because the facilities are, functionally, part of the Member State’s territory, or because the Member State exercises jurisdiction there despite the absence of territorial sovereignty?

The distinction also emerges from a comparison with Advocate General Emiliou’s approach in Sedrata (C-414/25). Both Opinions accept that EU law continues to govern the exercise of Member State powers in facilities located in Albania under Italian jurisdiction, but they reach that conclusion through different doctrinal routes. In Sedrata, Emiliou adopted a functional interpretation of the concepts of ‘border’ and ‘transit zone’, treating them as capable of encompassing areas under a Member State’s jurisdiction irrespective of their physical location (see De Leo, Frasca & Bornemman). By contrast, Medina expressly rejects classifying the Albanian facilities as border or transit zones because they are situated within the territory of a third State. She relies instead on the territorial-lease construction. If the Court were to follow Advocate General Medina’s approach, it would have to determine not only whether EU law follows the exercise of public power abroad, but also which doctrinal basis should support that conclusion.

2. From Compatibility to Competence: A Methodological Problem

The central question raised by the Opinion is therefore not whether these guarantees matter; they clearly do. The more difficult issue is why their deficient implementation should necessarily be characterised as an infringement of the Union’s exclusive external competence. Article 3(2) TFEU has a structural function. It determines who may act externally, ensuring that Member State commitments do not undermine the unity and consistency of common EU rules. The defects identified by the Advocate General, however, do not all concern a conflict between the Protocol and those rules. Some concern the effectiveness of guarantees within the system created by the Protocol and, more specifically, the way in which Italian authorities have implemented the relevant EU requirements. This distinction matters. A defect resulting directly from the international agreement itself, most clearly the reference to Albanian law governing access by lawyers, can plausibly be understood as a treaty-based interference with common rules: it is the international commitment itself that alters the practical conditions under which an EU-law right is exercised.

The argument becomes less straightforward where the alleged deficiency results primarily from national implementing choices. If every defective implementation of a harmonised guarantee can be characterised as an alteration of the scope of common rules and therefore as an encroachment on exclusive external competence, Article 3(2) risks becoming a general mechanism for reviewing the compatibility of national implementation with EU law. That would blur an important distinction between two questions: who has the power to act externally, and whether the actor has exercised that power consistently with EU law. The Opinion is therefore most convincing where the defect is inherent in the international agreement itself. It becomes harder to sustain where the problem lies principally in the subsequent implementation of that agreement.

The issue is reinforced by the Advocate General’s disaggregation of the Protocol. By assessing location, grounds and conditions separately according to the degree of harmonisation, the Opinion avoids treating the agreement as a whole as falling within exclusive Union competence. Yet the same disaggregation may dilute the overall assessment traditionally required under Article 3(2): an agreement may produce effects on common rules through the interaction of its individual components even where no single provision appears sufficient in isolation. This concern is well grounded in the Court’s case law. In determining whether an international agreement ‘may affect common rules or alter their scope’, the Court has consistently required a comprehensive assessment of the relationship between the agreement and the EU rules in force, taking into account not only the extent of the area covered, but also the nature and content of those rules and their foreseeable development (Opinion 1/03). The decisive question is whether independent action by the Member States would be capable of undermining the uniform and consistent application of the common rules and the proper functioning of the system which they establish, an inquiry which the Court has increasingly framed in terms of effects, rather than mere material overlap (Green Network, C-66/13; Opinion 3/15). Applied to the Protocol as a whole, rather than to its individual components in isolation, such an assessment could well lead to a different conclusion: the cumulative interaction of the rules on location, grounds and conditions may affect the operation of the common system, even where each element, considered separately, appears compatible with it.

The same difficulty resurfaces in the premise underlying the distinction between location and conditions. The Advocate General infers from the directives’ silence on the place of detention that Member States retain discretion over where detention takes place. But another reading is possible. The legislature may have structured the common asylum system on the implicit assumption that the relevant procedures are conducted within the territory of Member States. If so, silence on location would not necessarily identify an area of autonomy; it could instead reflect a structural presupposition of the system.

This matters because the Opinion itself demonstrates that location can profoundly affect the exercise of rights. The practical difficulties surrounding lawyers’ access, judicial protection, release and return are not simply incidental consequences of detention; they arise precisely because detention has been relocated to a third country. The distinction between where detention occurs and how it is conducted may therefore be less clear-cut than the Opinion suggests.

Concluding Remarks

Beyond the immediate fate of the Italy–Albania Protocol, Medina’s Opinion acquires significance against the broader evolution of the external dimension of EU migration policy. The New Pact instruments will provide a new legal framework for border procedures and related forms of cooperation with third countries, while the new Return Regulation expressly envisages the possibility of transferring persons subject to a return decision to ‘return hubs’ in third countries. These developments point towards an increasingly structured externalisation of elements of migration management beyond EU territory.

Against this background, the Opinion does not establish a general prohibition on externalisation. Rather, it formulates a model of legally conditioned externalisation: Member States may relocate the setting in which they exercise their powers, but they remain bound by the guarantees of the common system wherever EU law continues to govern the exercise of those powers. The Opinion’s most important contribution may therefore lie not in deciding whether detention centres can be located outside the Union, but in establishing the principle that geography cannot be used to dilute legal responsibility. Functional jurisdiction explains why EU guarantees follow the exercise of public power; the more difficult question, left open by the Opinion, is when the failure to secure those guarantees becomes a question of exclusive Union competence. 

That distinction will matter increasingly as Member States and the Union develop new forms of cooperation with third countries. Externalisation may change the physical setting in which migration powers are exercised, but it cannot, without further justification, change the level of protection afforded by EU law.

Ultimately, the legitimacy of externalisation should be assessed not by the location of the facilities alone, but by the legal responsibility that accompanies the exercise of public power. Wherever a Member State exercises its powers over applicants for international protection, it remains answerable for the guarantees of the common system. The geography of that exercise cannot become a mechanism for reducing those guarantees. 

C. Suggested Reading 

To read the case: CJEU, Comeri, Joined Cases C-706/25 and C-707/25, Opinion of Advocate General Medina Delivered on 11 June 2026, EU:C:2026:474.

Case law: 

Doctrine: 

To cite this contribution: F. Zahra El harch, “From Compatibility to Exclusive External Competence: Advocate General Medina’s Opinion in Comeri and Sidilli”, Cahiers de l’EDEM, August 2026.