Abstract
This article critically examines the extraterritorial asylum processing model established by the Italy–Albania deal, arguing that it departs from traditional EU externalization strategies and presents new legal challenges. By analysing the criteria used by the EU Court of Justice to determine whether situations not explicitly regulated by EU law fall within its scope, the article shows that the examination of asylum claims in Albania could trigger the applicability of Union law, despite the territorial limitations of much of EU secondary legislation in the area of asylum. It further explores how this system may undermine the effectiveness and integrity of the CEAS and violate certain provisions of the Charter. The article concludes by advocating for CJEU intervention to clarify the relationship between extraterritorial asylum procedures and EU law, ensuring legal certainty and the consistent application of Union law.
Keywords
Introduction
On 6 November 2023, Italy and Albania signed a Protocol between the Government of the Republic of Italy and the Council of Ministers of the Republic of Albania on Strengthening Cooperation in Migration Matters (hereinafter, ‘the Protocol’). 1 The agreement aims at ‘reinforcing the bilateral cooperation in the management of migration flows’ and binds Albania to admit migrants on its territory ‘with the sole purpose of carrying out the [asylum] border and return procedures regulated by Italian and European law’, 2 which take place in two ‘areas’ on Albanian territory. The Protocol was ratified by the Albanian 3 and the Italian 4 Parliaments on 22 February and 23 February, respectively. The centres became operational on 14 October 2024. 5
The Italy–Albania deal has been praised as a ‘truly European’ and ‘historic agreement for the entire European Union’ by the Italian Prime Minister, 6 and as an example of ‘out-of-the-box thinking, based on fair sharing of responsibilities with third countries in line with obligations under EU and international law’ by the President of the European Commission. 7 This bilateral initiative aligns with a trend toward increased reliance on cooperation with third countries to manage migration inflows and prevent entry into EU territory. However, it presents innovative elements that depart from previously employed externalization tools in the European Union (EU) and introduces several legal complexities.
The offshore asylum processing model established by the Protocol involves assessing asylum claims through an accelerated border procedure on Albanian territory but under Italian jurisdiction – meaning under the control and supervision of Italian personnel and ‘in accordance with Italian and European law’. 8 These novelties raise significant legal and conceptual challenges. While the Protocol nominally replicates EU asylum standards, its extraterritorial implementation falls outside the direct scope of much of the secondary legislation comprising the Common European Asylum System (CEAS), creating uncertainty about the applicability and enforceability of EU asylum law in a third country. 9 This raises several questions about whether this policy of extraterritorial migration control can be satisfactorily conceptualized within classical EU externalization strategies, as well as about the relationship between national and EU asylum law.
Potential discrepancies between the asylum procedure in Albania and the one conducted within the EU territory might jeopardize the effectiveness of EU law in an area, that of asylum, of shared competence in which the Union has worked to ensure the uniform treatment of asylum seekers throughout Europe. 10 These harmonization efforts have been further enhanced by the adoption of the Pact on Asylum and Migration in May 2024. 11
Against this backdrop, the article begins by situating the Italy–Albania model within the broader phenomenon of externalization of EU migration controls, highlighting how it departs from previous practices of cooperation with third countries and the novel legal challenges it introduces. It then examines whether the deal aligns with EU law. First, we argue that certain aspects of the extraterritorial asylum procedure in Albania are likely to trigger the applicability of EU law, both with regard to primary law, and in particular the EU Charter of Fundamental Rights (hereinafter, also ‘the Charter’) 12 , as well as with regard to secondary legislation in the field of asylum – notably the Dublin Regulation, 13 the Asylum Procedures Directive (APD) 14 and the Reception Conditions Directive (RCD). 15 Next, we assess potential risks of incompatibility with these legal frameworks. Finally, the article explores various scenarios to qualify the relationship between extraterritorial asylum processing in Albania and the aforementioned legislation. 16
Rethinking externalization through the lens of the Italy–Albania Protocol
The concept of externalization, which emerged in the early 2000s, has today developed into an umbrella notion, covering a wide range of migration control practices affecting the rights of refugees. 17 While its precise definition is still debated, the United Nations High Commissioner for Refugees (UNHCR) broadly describes the phenomenon as encompassing ‘measures taken by States – unilaterally or in cooperation with other States – which are implemented or have effects outside their own territories, and which directly or indirectly prevent asylum-seekers and refugees from reaching a particular “destination” country or region, and/or from being able to claim or enjoy protection there’. 18 In the EU context, this strategy, also referred to as shifting borders, 19 has been widely advocated and implemented in several of its Member States, with Italy being one of its main promoters. It has taken two main approaches. The first is aimed at preventing arrivals into the territory; the second at removing people from the territory as swiftly as possible, by shifting the responsibility to third countries. The next section will briefly examine these two strategies, to then explain why the model created by the Italy–Albania deal is unique and cannot be satisfactorily conceptualized within the concept of externalization, while still borrowing features from it. 20
Policies of non-arrival
The first category of externalization relates to measures aimed at preventing Third Country Nationals (TCNs), including potential asylum seekers, from reaching the territory to request protection. 21 Relevant examples include visa regimes and carrier sanctions, 22 the posting of international liaison officers in third countries, 23 maritime interceptions and returns to third countries, 24 as well as the funding and training of border and migration management authorities of third countries.
In the EU context, this approach primarily involves outsourcing border and migration management responsibilities to third countries, in exchange for financial and technical assistance. A prominent example is Italy's cooperation with the Libyan Coast Guard under the controversial 2017 Memorandum of Understanding (MoU) between Italy and Libya, funded in part by the EU. 25 Notably, the MoU was signed in the aftermath of the 2012 landmark ruling of Hirsi Jamaa and Others v. Italy. 26 In that case, the European Court of Human Rights (ECtHR) condemned Italy for directly intercepting migrants on the high seas and returning them to Libya without assessing their individual protection needs, in violation of the principle of non-refoulement and the prohibition of collective expulsions. Under the 2017 deal, the responsibility for intercepting migrants shifted from Italy to the Libyan Coast Guard, effectively distancing Italy from direct involvement while still hindering access to its territory through its partnership with Libyan authorities.
The main legal issue arising from this externalization tool concerns the difficulty of attributing responsibility for fundamental rights violations resulting from this practice to the EU or its Member States, whose involvement is limited to the provision of funding and technical assistance, while the fundamental rights violations are committed by and happen within the jurisdiction of non-EU authorities. 27
Policies of non-admission
The second category of externalization involves measures entailing transferring arrangements to third countries after asylum seekers have reached the territory of states where they seek protection. 28 When it comes to the EU context, this happens primarily by declaring asylum applications inadmissible, through the use of safe third country (STC) or first country of asylum concepts on the basis of Article 33 of the APD. The EU–Turkey Statement is widely considered the blueprint of this approach. 29 It relies on the STC concept to allow for the transfer of asylum seekers arriving in Greece via Turkey back to Turkey, provided that applicants are firstly given an opportunity to rebut the presumption that Turkey qualifies as a STC in their specific case. Under the current APD, as well as the forthcoming Asylum Procedure Regulation (APR), transfers based on the STC concept are only permissible if there is a connection between the applicant and the designated STC. 30 However, the new APR includes a provision for revising the STC concept in 2025, potentially paving the way for further externalization attempts. Such arrangements, strongly advocated by Italy during the APR negotiations, 31 would permit the transfer of asylum applicants to third countries with which they have no prior connection, akin to the now-repealed 2022 UK–Rwanda deal. 32
In addition to the STC concept, readmission clauses in international agreements and the Safe Country of Origin (SCoO) concept are also frequently mentioned as significant practices of externalization. 33 The SCoO allows an accelerated examination of the merit of an asylum application under Article 31(8)(b) of the APD, as to ensure quick readmissions. While measures aimed at facilitating readmission could, in the absence of adequate procedural guarantees, effectively prevent TCNs from accessing protection in practice, they do not conceptually constitute an outsourcing of responsibility. This is because the cooperating third country does not undertake a status determination procedure or assume a border control function on behalf of the EU. Instead, it merely agrees to readmit its own nationals. The Memorandum of Understanding (MoU) with Tunisia, through which the EU secured Tunisia's commitment to readmit its own nationals, 34 and which is aimed at ensuring the practical implementation of Tunisia's designation as a SCoO serves as an example of such a strategy. 35 Italy stands out as one of the EU Member States making extensive use of this policy, maintaining one of the longest national lists of safe countries of origin. 36
In asylum procedures, the concepts of STC and SCoO are often used in border procedures, taking place in border or transit zones, on the basis of Article 43 APD. Border procedures often rely, either explicitly or implicitly, on the fiction of non-entry, which is a legal concept used to treat asylum applicants physically present in a country's territory as if they have not yet legally entered it, to justify restriction of movement or deprivation of liberty during the procedure. 37 The underlying rationale is that applications deemed to have a low likelihood of success, with regard to either admissibility or merit, should be processed swiftly, enabling the prompt return of individuals who do not qualify for protection. However, the presumption of ineligibility for protection, and the application of the non-entry fiction, significantly lowers procedural guarantees. This includes shortened deadlines, a reversal of the burden of proof and the removal of the suspensive effect of appeals, thereby limiting the ability of individuals to fully articulate their protection needs, ultimately increasing the risk of refoulement. 38
The Italy–Albania Protocol: A novel approach to extraterritorial migration control in the EU
Against this backdrop, we argue that the Italy–Albania Protocol constitutes a new model of border-shifting initiatives based on cooperation with third countries. It entails offshore asylum processing, that is, the transfer of asylum seekers to a location outside the territory of the sending country or to a third country to undergo refugee status determination procedures there, which occurs within a non-entry fiction based on an accelerated border procedure applied to asylum applicants coming from SCoO. This represents a novelty in the EU panorama, albeit offshore processing has been systematically invoked as a possible solution to reduce irregular immigration during spikes in migration flows to Europe. 39
The content of the Italy–Albania agreement
The Italian ratification law provides that only migrants who are intercepted by Italian authorities outside the territorial sea of Italy or other EU Member States, including following a search and rescue (SAR) operation, can be transferred to Albania. 40 This is based on the premise that the Asylum Procedure Directive obliges Member States to assess applications for international protection in their territory, or at the border or in transit zones, interpreted as areas geographically located within the territory, only when the application is made ‘in their territory, at the border or transit zones, or in their territorial waters’. 41
As for its material scope, the Protocol only mentions ‘border and return procedures’, conducted in accordance with Italian and European law. 42 The ratification law specifies that three types of procedures can be carried out in the two areas. The first one takes place in the port of Shengjin, where a facility accommodates migrants to conduct initial health checks and a preliminary assessment (or, to use the words of a newly adopted regulation, a screening) 43 to identify vulnerable applicants who cannot be subjected to the asylum border procedure. Neither the Protocol nor the ratification law excludes particular categories of persons, such as vulnerable groups, from the scope of the agreement, but their exclusion results from the application of the Italian asylum legislation to the processing. 44
The second and third procedures will instead take place in the area of Gjader, which hosts two facilities: one, with a current capacity of 880 places, is intended to conduct an asylum border procedure based on the SCoO concept 45 which should last maximum 28 days, and the second, with a current capacity of 144 places, is intended to conduct a return procedure for up to 18 months. 46
With regard to the law applicable to the processing, the ratification law refers to the national law transposing the Qualification Directive (QD), 47 the APD and the RCD, 48 together with an unspecified ‘body of Italian and European legislation’ concerning the requirements and procedures relating to the admission and stay of foreigners in the national territory. 49 The law further states that the jurisdiction is Italian in relation to all the procedures envisaged by the Protocol, and identifies the administrative and judicial authorities in Rome as the competent ones to examine the applications for international protection, as well as to deal with disputes concerning asylum, detention, repatriation and criminal conduct. 50
In the light of what has been described so far, the system, which commits to treating applicants ‘as if’ they were in Italy, could be defined as an ‘Italian enclave in Albanian territory’ 51 where Italian law applies in full, while the Albanian authorities are only responsible for ensuring the security of the ‘external perimeter’ of the areas. 52 The wording of both the Protocol and the ratification law makes it clear that the Italian Government's intention is not to make use of a parallel, and alternative, procedure to that in force in Italy, but rather to apply the same body of Italian law that transposes the relevant EU directives, to both applicants in Italy and those in Albania.
Re-thinking the concepts of externalization, territory and border in the application of the EU-regulated border procedure in Albania
The Italy–Albania model, which effectively relocates the EU border procedure to a third country, embodies two core features of the externalization models traditionally employed in the EU. It restricts physical access to the territory (non-entry), while also conducting asylum border processing under the concept of a Safe Country of Origin (SCoO) in a non-entry fiction scenario (non-admission). However, Italy retains responsibility for relocating beneficiaries of international protection and for returning those who do not qualify for asylum. This means that Italy does not externalize its responsibility for international protection or border management to Albania, which serves merely as a supportive state. Paradoxically, this approach diverges from the trend among several states – pursued through outsourcing agreements – to absolve themselves of responsibility for actions occurring outside their territory (or even in ‘border’ areas). 53 This shift raises novel legal questions.
As noted, both the Protocol and its ratification law stipulate the applicability of the relevant EU asylum directives to the extraterritorial asylum procedures conducted in Albania. However, this raises ambiguities regarding the possibility of applying EU asylum law in a third country. Notably, the QD, which governs the criteria for granting international protection, does not explicitly impose a territorial limitation, as it applies to ‘anyone who has submitted an asylum application pending a final decision’. 54 Regarding primary law, Article 78 of the Treaty on the Functioning of the European Union (TFEU), which grants the EU competence to regulate a uniform status for third-country nationals in need of international protection, similarly lacks an explicit territorial restriction. By contrast, the Asylum Procedures Directive (APD) and the Reception Conditions Directive (RCD) explicitly confine their scope to Member States’ territories, borders, transit zones and territorial waters.
The Italian ratification law does not resolve the tension between the territorial limitations of EU directives and their extraterritorial application in Albania. Instead, it vaguely asserts that EU law applies ‘as it is compatible’. 55 This interpretative ambiguity leads to the question of whether EU law permits the issuance of a permit to stay under a procedure that, while formally replicating EU standards, is conducted outside the EU geographical territory. It seems that Italian authorities have tackled the issue with an extensive interpretation of the notion of ‘border’, which encompasses areas geographically located outside the perimeter of the Italian territory. In this respect, one could indeed question whether Member States retain the discretion to define which geographical areas they consider as their border or territory for the purpose of applying EU (asylum) law, 56 provided that such designations comply with international law, as neither the Treaties 57 nor the APD contain a clear definition of these notions.
The possibility of issuing asylum extraterritorially has been, however, met with scepticism by EU institutions. Since the outset of discussions on ‘offshore processing’ of applications for international protection, the institutions have emphasized the strict territorial application of asylum rules. 58 In 2018, following the proposal of the European Council to create ‘regional disembarkation platforms’ outside the territory of the Union, the European Commission explicitly rejected the possibility of granting asylum outside the EU, stating that ‘the only way to obtain refugee status outside the EU is [….] resettlement. Allowing people to “apply” for asylum outside the EU would require an extraterritorial application of EU law that is currently neither possible nor desirable.’ 59 Instead, the transfer of migrants rescued in non-EU territorial waters or international waters – including by vessels flying the flag of a Member State – to third countries was an option envisaged as permissible only when consistent with the principle of non-refoulement. This meant that EU Member States would be allowed to transfer migrants to third countries when their lives would not be at risk, but without retaining responsibility for examining asylum seekers’ applications post-disembarkation. The Legal Service of the European Parliament adhered to the Commission's conclusions, stating that while migrants rescued by vessels flying an EU State's flag in international waters might be under that State’s jurisdiction, this does not necessarily mean that the vessel is considered part of EU territory under Article 3 of the APD. 60 The position of the Institution seems to imply that, while Member States are free to define what they consider territory and borders, including for instance vessels in international waters, 61 the concept of territory and borders for the purpose of applying EU asylum law has an autonomous meaning, which, in their view, is limited to the geographical territory of the Union.
As mentioned above, Italian legislation does not explicitly clarify whether it is extending EU law to a situation not expressly regulated by it or directly applying it to asylum processing in Albania. However, the provision limiting the deal to migrants rescued in international waters, while excluding those rescued within Italian territorial waters, suggests that Italy implicitly acknowledges that EU asylum law cannot be applied in Albania by simply artificially designating part of Albanian territory as an Italian border zone. The model established by the Italy–Albania deal thus avoids the direct application of EU (procedural) asylum law in a third country, and unilaterally extends it.
While this manoeuvre raises important questions about the permissible reach of EU law beyond its established territorial boundaries, it is notable that Italian judges addressing the implementation of the deal have not directly focused on this issue. Specifically, when asked to validate the detention orders for the first two groups of asylum applicants rescued on the high seas and transferred to Albania, the Court of Rome initially disapplied national legislation transposing the APD-regulated border procedure 62 and subsequently referred the matter to the ECJ. 63 However, the Court focused exclusively on issues concerning the general implementation of the Italian border procedure based on the SCoO concept, without addressing the fact that the procedure is conducted in a third country. This approach implicitly assumes the applicability of EU law to the processing of asylum applications in Albania. Yet, by sidestepping the question of whether EU asylum law can legitimately be applied in a third country, the Italian judges’ approach risks leaving unresolved the broader implications of using EU law as a framework for extraterritorial asylum processing – an issue we believe warrants further exploration to ensure legal certainty.
To conclude, in implementing this new model of extraterritorial migration control, Italy is navigating uncharted legal territory by applying EU asylum rules in a third state. While this approach leverages gaps in EU legislation – such as the QD's undefined territorial scope and the mismatch between the scope of application of unilaterally extended directives – it creates legal uncertainties about the relationship between EU and national law in extraterritorial contexts. These issues, we argue, are suitable to be clarified by the European Court of Justice. In the next section, we will firstly outline why there are grounds to say that the Court would be competent to address these matters and, secondly, we will examine the compatibility issues that might arise.
The question of the interplay with Union law: The sufficient connection criteria as determining the applicability of EU law 64
Some commentators have argued that the ‘unilateral extension of procedures arising from EU law to a context (the extraterritorial one) not expressly regulated by the latter is in accordance with EU law’. 65 This view seems to reflect the Commission's initial stance on the matter. 66 Others have, instead, directly used EU Directives to assess the legitimacy of the scheme, identifying potential conflicts between them. 67 As explained, this interpretative uncertainty arises from Italy's unilateral extension of EU directives that do not ordinarily apply in a third country. However, this does not necessarily mean that the entire system falls outside the scope of EU law. Certain elements of the Protocol may still impact the effective application of Union law, potentially triggering the competence of the Court of Justice.
Generally speaking, EU law does not preclude its application to situations not expressly covered by it, unless there is a specific prohibition to do so. The APD itself allows for this possibility, stating that ‘Member States may decide to apply this Directive in proceedings concerning applications for any form of protection outside the scope of the Qualification Directive’, 68 and which have therefore a purely national validity. 69 However, the situation envisaged by the Protocol differs from the case addressed by this provision, as Italy intends to grant the same status of international protection recognized within its territory, governed by EU law, 70 based on procedures conducted in Albania. In other words, the intention is to extend the APD and the RCD to situations that, while outside their territorial scope, fall within the material scope of the QD. This option is currently neither contemplated nor expressly precluded by EU law.
Indeed, the APD does not explicitly prohibit the issuance of international protection statuses through extraterritorial procedures, except in cases involving applications made at embassies. 71 This omission likely reflects the fact that EU legislators had not considered scenarios such as the one proposed by the Italian government. Nonetheless, one could argue that the exclusion of asylum procedures within embassies from the APD's scope suggests an intention to exclude its extension extraterritorially. However, as argued elsewhere, the legal regime governing the areas referred to in the Protocol is significantly different from that applicable to diplomatic missions. While embassies are generally subject to the law of the host state, except in matters covered by diplomatic immunity, the areas designated under the Protocol are managed entirely in accordance with the law of the beneficiary state. Within these centres, Italian legislation governs all aspects, including access to international protection, return procedures and civil matters such as the registration of births and deaths. Crimes committed by migrants within the centres are prosecuted under Italian criminal law, rather than Albanian law. This arrangement creates a unique extraterritorial legal system that departs from traditional principles of international law governing diplomatic representation, and which is thus not prone to be analogically applied to it. 72 This confirms the need to explore the implications of the unilateral extension of an EU-regulated procedure to a situation not expressly covered by it, starting from the premise that this is not explicitly prohibited.
As a preliminary remark, it should be noted that the scope of Union law is closely related to its competence to regulate a given subject-matter. Although interrelated, the notions of scope and competence are nonetheless not identical: the second has a dynamic and flexible character, being understood as the concrete ability to regulate legal situations pertaining to areas of substantive competence attributed by the Treaties. 73 The scope of EU law can vary based on several factors. These relate notably to the actual exercise of competence in a given area or, in cases where Member States retain residual competence, the presence of a ‘sufficient connection’ to an EU law provision. The Court of Justice of the European Union (CJEU) has used the ‘sufficient connection’ criterion in two strands of cases having features similar to those presented by the Protocol. The first involves scenarios that formally fall within national competence but have the potential to interfere with Union objectives, thereby necessitating regulation to prevent such interference. The second concerns extraterritorial actions that can be governed by EU law due to their substantive link to the EU territory and its legal framework. These criteria, as identified by the CJEU, may serve as a useful analogy when analysing the legal implications of the Italy–Albania Protocol, in terms of applicability of Union law to the extraterritorial processing, since EU law intersects with both national situations not expressly covered by it, and extraterritorial contexts.
The applicability of Union law to national situations not directly regulated by it
The first strand of case law reflects instances where the Court of Justice has asserted its competence to assess whether national measures comply with EU law, even in situations that technically fall outside its direct scope. One example is the extension of EU law-derived rights, such as the right to family reunification, to EU citizens who have not exercised their freedom of movement. In such cases, the Court has justified its competence by emphasizing the need to preserve the uniform interpretation of EU law, interpreting EU-derived concepts even when applied to situations not explicitly covered by it. This is necessary to ensure ‘identical treatment of those situations and of those falling within the scope of EU law’, which would be jeopardized if notions taken from EU law were subject to divergent interpretation. 74
A second example involves scenarios where national laws, while formally partly outside the EU's competence (e.g., in criminal law), may nonetheless fall within the scope of EU law if they affect the attainment of objectives of provided in Union law. In cases like El Dridi 75 and Achughbabian, 76 the Court ruled that EU law (specifically the Return Directive) precludes Member States from using criminal sanctions, such as imprisonment, to address illegal stays by third-country nationals, even in the absence of EU competence to introduce criminal sanctions on illegal stay in criminal matters. Such measures were deemed to undermine the Directive's objectives and thus render it ineffective. Although the Court did not interpret domestic criminal law concepts, it held that the national provisions were incompatible with EU law.
Lastly, the Court has recognized the applicability of EU law, and specifically the Charter of Fundamental Rights, to situations where Member States exercise discretionary powers granted by EU legislation. In NS, for instance, the Court found that, since the UK recognized itself responsible for examining an asylum application by virtue of Article 3(2) of the Dublin Regulation, 77 a discretionary power which is an integral part of the CEAS, such discretion should be considered an implementation of EU law within the meaning of Article 51(1) of the Charter. 78
These cases collectively underscore the Court's approach to ensuring that national actions align with the objectives and principles of EU law, even in areas of residual competence.
The territorial extension of EU law
The second strand of case law concerns the territorial extension of EU law, where the Court of Justice has recognized the Union's ability to regulate activities conducted outside its territory if necessary to preserve the effectiveness of EU law. 79 In these cases, the Court assesses whether a sufficient connection exists between the activity in question, which should fall within an area of EU competence, and the territory and legal system of the EU. This approach has been used in a number of cases concerning situations of a transnational nature or relating to EU workers temporarily abroad, in the fields of competition law, 80 the right to privacy, 81 the environment, 82 consumer 83 and animal welfare protection, 84 electronic commerce 85 and labour law. 86 They illustrate a ‘functional approach to jurisdiction’, 87 whereby activities occurring outside the EU but producing effects within it are brought under EU law to safeguard rights or interests governed by the Union legislation. While this approach requires a connection to the Union, it does not strictly adhere to territorial considerations. 88
This strand of case law further confirms that the scope of EU law is closely tied to the division of competences rather than notions of territoriality. Although the Treaties and Court rulings reference the Union's or Member States’ territories, these references serve to identify the actors bound by EU law rather than limit its reach. 89 As evidence of this, the Charter of Fundamental Rights of the European Union is applicable whenever a Member State implements Union law, 90 and does not make use of the notion of jurisdiction (either territorial or extraterritorial) to delimit its scope of application, unlike most human rights treaties. 91
For instance, in X & X v. Belgium, the Court denied the applicability of the Charter of Fundamental Rights to an application for a humanitarian visa, for reasons that did not relate to the extraterritoriality of the application. In the Court's reasoning, as well as those of the referring judge, 92 the fact that the applicants were outside the territory of the Union was not relevant. On the other hand, using a functionalist approach, 93 the Court remarked that allowing third-country nationals to apply for international protection at Member States’ representations located on the territory of a third country – a situation excluded from the scope of the APD – would have undermined the general framework established by Regulation 2013/604/EU regarding the competence of states to examine asylum applications (Dublin Regulation), insofar as it would in practice have allowed applicants to choose where to apply for international protection. 94 Even though the Court excluded the applicability of the Visa Code, it recognized its competence to rule over the case. This functionalist approach underscores the Court's focus on maintaining the coherence and effectiveness of EU law, extending its reach when necessary to achieve the objectives underpinning EU legislation.
Application of the ‘sufficient connection’ criteria to the Italy–Albania deal
The analysis of the Court of Justice's use of the ‘sufficient connection’ criterion supports the conclusion that the Italy–Albania Protocol could fall within the scope of application of EU law, as it contains several elements linked to the EU legal system. The agreement concerns a matter of shared competence, 95 which has been exercised by the Union through secondary legislation that has harmonized the criteria, both substantive and procedural, for the recognition of international protection. As mentioned, the procedures outlined in the Protocol and its ratification law fall within the scope of EU law ratione materiae, which is precisely what allowed Italy to extend it in Albania without adopting new legislation. The applicability of Union law is, in this respect, rendered uncertain solely by the above-mentioned predominantly territorial nature of the CEAS.
With this in mind, one could ask whether examining asylum applications in view of granting or refusing a permit for international protection could be sufficient, in itself, to bring the situation within the scope of EU law, regardless of the type of procedure used for that purpose. Indeed, as mentioned, migrants transferred to and detained in Albania acquire the status of applicants for international protection and would potentially be granted the status of refugees or subsidiary protection beneficiaries, in line with the QD. Italian officials will not limit themselves to selecting those who may have access to a humanitarian visa to then apply for asylum in the EU, but they will carry out a full examination of the merits of the application, on the basis of procedures that are fully assimilated to those under EU law.
This has the potential to produce several effects within the territory and the legal system of the EU. Beneficiaries of international protection will eventually be transferred to Italy, and therefore to the EU, 96 gaining rights such as intra-EU travel for up to 90 days within a 180-day period. 97 Furthermore, the Italian administrative and judicial authorities responsible for asylum applications, detention decisions and appeals operate from Italy, applying EU law to applications processed both in Italy and Albania, as made even more evident by the recent preliminary ruling submitted by the Court of Rome.
In conclusion, the connections identified between the Italy–Albania Protocol and the EU legal system appear sufficient to bring the situation within the scope of Union law. This would thereby attract the competence of the Court of Justice, which would be fully competent to assess whether certain elements of the extraterritorial processing in Albania could undermine the objectives of the EU directives applied extraterritorially, as well as other key instruments of the CEAS, such as the Dublin Regulation, notably absent from the ratification law. As will be explored below, the use of an EU-regulated procedure in a third country might risk jeopardizing the CEAS’ objective of creating harmonized asylum rules. This harmonization is crucial in a border-free area like the EU, as it ensures uniform protection standards, full compliance with fundamental rights, and prevents secondary movements within the Union. 98
The compatibility of the Italy–Albanian deal with EU law
Having clarified that the subject falls within the scope of EU law, it remains to be assessed whether and how it could undermine the effectiveness of supranational legislation and its objectives, in particular with regard to the framework of the CEAS, taking into account specifically the Dublin Regulation, the Asylum Procedure Directive and the Reception Conditions Directive, as well as when it comes to the EU Charter of Fundamental Rights.
The compatibility with the objectives of the Dublin system
The Italy–Albania Protocol raises significant concerns about its potential impact on the functioning of the Dublin system, a cornerstone of the CEAS. A critical issue lies in the lack of clarity over whether migrants processed in Albania are registered in the EURODAC system. 99 The failure to register asylum applicants in EURODAC could effectively exclude them from the Dublin procedure if they later applied for protection in another EU Member State. For instance, if asylum seekers left the ‘Italian’ centres in Albania, traversed the Balkans and submitted asylum applications elsewhere, the absence of EURODAC registration would prevent the receiving Member State from submitting a ‘take back’ request to Italy. Additionally, asylum seekers who receive a negative decision but who cannot be returned and are brought back to Italy could then move freely to other EU states. These states could subsequently become the ‘first country of entry’ if applicants were never registered in EURODAC. This scenario risks disrupting the allocation of responsibility for asylum claims under the Dublin Regulation.
In this regard, in X & X, the Court has already pointed out that the examination of asylum applications in a third country could undermine the overall structure of the Dublin Regulation. In that case, conducting an extraterritorial examination of asylum applications in embassies was found to jeopardize the allocation of competence established by the Dublin Regulation, as it would allow applicants to choose where to apply. In the case of the Italy–Albania Protocol, Italy would carry out the entire procedure without registering applicants in EURODAC or applying Dublin criteria. While this does not create a system where applicants can freely decide where to apply, it could still undermine the objectives of the Dublin system. Specifically, the competent Member State – Italy, if applicants were allowed to reach its territorial waters – could avoid its responsibility by designing an examination process in which EURODAC, the Dublin Regulation and the APD do not directly apply.
Even if Italy registered migrants in EURODAC and assumed full responsibility for ‘take back’ requests, another issue might emerge. Since the Regulation is governed exclusively by EU law, any unilateral extension of its applicability beyond its territorial limits could have direct repercussions for other Member States. For instance, those Member States which might be the competent ones under Dublin criteria other than the country of first entry would be required to assume responsibility.
Both situations show how other Member States could be impacted by Italy's unilateral decision to extend the EU asylum procedure to a third country. This underscores broader tensions between national initiatives and the harmonized framework of the CEAS.
The compatibility with the objectives of Asylum Procedures and the Reception Conditions Directives
In addition to the concerns about the compatibility with the responsibility-sharing framework of the Dublin system, the unilateral extension of the APD and the RCD to a third country further risks undermining the very objectives of these directives. They aim to create a unified asylum procedure within the EU, with common reception standards designated to ensure a fair and efficient assessment of asylum applications. Such a system is essential to guarantee that those who fulfil the criteria in the QD have effective access to international protection status. 100 Therefore, to ensure that the granting of international protection under the Italy–Albania Protocol does not undermine the goal of having a unified asylum procedure and common reception standards across EU Member States, the processing conducted in Albania must fully align with the standards set out in the relevant EU directives. In other words, it might be argued that Italy is not discretionally extending EU-derived rights in a context where it is not obliged to do so, but that alignment with EU standards is required in order to preserve the effectiveness of EU law in this area.
Against this backdrop, while the Protocol provides for formal adherence, it is questionable whether such alignment can be effectively ensured. This relates to two main issues. The first concerns the feasibility of guaranteeing the same standards of treatment in a third country in practice, while the second pertains to the qualification of a third country's territory as a ‘border zone’ for the purpose of applying the EU-regulated border procedure.
As for the first issue, and without claiming to be exhaustive, there are certain elements in the Protocol and its ratification law that raise doubts about compatibility with the APD. For instance, regarding the right of defence, there is uncertainty about whether asylum seekers will be informed of their rights in a language they understand or have physical access to legal advisors, as the ratification law only stipulates that these rights will be provided when remote communication proves impossible. 101 Furthermore, the law mandates that communications with lawyers will typically occur via certified email, which, combined with restricted access to communication devices (such as mobile phones), could hinder effective communication. 102 Additionally, the law foresees that personal interviews and hearings will, by default, be conducted remotely, 103 despite numerous sources highlighting the challenges posed by online interviews, particularly regarding confidentiality and the effective opportunity for applicants to present all the grounds for their application comprehensively. 104 The difficulties relating to remote asylum interviews suggest that in-person ones should be strongly preferred. 105 Collectively, these provisions could hinder the exercise of defence rights and the right to be heard, ultimately limiting asylum seekers’ ability to fully participate in their asylum procedures.
Furthermore, the implementation of the Protocol risks selectively applying only certain provisions of the RCD. For instance, there are no detailed rules regarding how to ensure proper accommodation conditions or guarantee prompt and effective access to healthcare. 106 Additionally, when it comes to vulnerability assessments, which are crucial for identifying individuals who cannot be processed under border procedures, these assessments are conducted both at sea and after disembarkation in Albania. The difficult conditions during long journeys may hinder the early identification of less obvious vulnerabilities, such as being victims of torture. 107 This might lead vulnerable people to be processed under the accelerated border procedure in Albania, instead of under the ordinary one in the territory. 108
When it comes to the second issue, we have explained how the protocol works on the basis of the designation of Albania as an Italian border zone for the purpose of applying the EU-regulated border procedure. While, as mentioned, the APD does not define the terms ‘border’ or ‘transit zone’, a systematic analysis of the Directive, as well as of other instruments of EU law, support a narrow interpretation of the concept of border. 109 This arguably preclude the designation of centres in Albania as a ‘border’ or ‘transit zone’ to conduct a border procedure in. There is, indeed, a significant conceptual difference between border zones that Member States designate within their own geographical boundaries and those located outside of them. Border zones within the EU are areas that migrants reach spontaneously, where they are then required to remain for the duration of the procedure if they choose to apply for protection. 110 In contrast, designating a border zone outside the EU territory and forcibly transferring asylum seekers there seems inconsistent with the purpose of border procedures under the APD, as interceptions before arrival at the border zones prevent a clear understanding of whether applicants would have reached them on their own. This automatically subjects them to a more restrictive procedure, and it might in turn render it arbitrary. 111 This concern is particularly serious, as asylum applicants in Albania are not only subjected to an accelerated border procedure, but they are also detained. While, in theory, the applicability of Italian law should mean that those who qualify for alternative measures to detention in Italy would not be transferred to Albania, Italy's legal framework for alternative measures poses significant challenges. Specifically, Italy only allows access to alternative measures for those who can provide a passport or agree to deposit a financial guarantee. 112 As a result, many applicants may fail to meet these requirements and would be deemed ineligible for alternative measures. Clarifying whether the use of the EU concept of ‘border’ in a third country aligns with the purpose of APD is essential. It would avoid that concepts borrowed by EU law are subject to divergent interpretation, and ultimately prevent the arbitrary use of the border procedure and, by extension, an arbitrary deprivation of liberty in cases where border procedures are routinely carried out in detention.
In conclusion, the failure to fully align the procedures and reception standards in Albania with those guaranteed by EU directives, along with the potential lowering of these standards – a concrete risk, in our view – could undermine the integrity of the CEAS. This would create a disparity in standards of treatment 113 and contradict the objective of having a harmonized and common procedure across the EU.
The compatibility with the Charter
The extraterritorial processing in Albania could also raise significant concerns regarding the respect for fundamental rights as enshrined in the Charter, as failure to adhere to the standards in the directives recalled above could also violate the Charter itself. In this regard, since Italy is clearly implementing the QD in Albania, the Charter would apply independently of the direct applicability of the APD and the RCD in a third country. In other words, if a restrictive interpretation of the objectives of the APD and the RCD could suggest that Italy is not required to adopt all the standards outlined in the two Directives, it would still be obligated to design its procedures and reception conditions in a manner that guarantees effective access to the rights guaranteed by the Charter. The situation in this case would be comparable to limitations that the Charter places on national procedural autonomy. 114 Indeed, secondary law establishing common procedures and standards for asylum applicants across Member States serves not only to achieve harmonization within the EU, but also to effectively operationalize access to asylum and the prohibition of refoulement, as guaranteed by Articles 18 and 19 of the Charter. From this perspective, conducting asylum procedures outside EU territory with a lower level of protection than that provided under EU law would undermine the right to asylum as a fundamental right under EU primary law.
In addition to the right to asylum, other fundamental rights, such as the right to an effective remedy and the right to liberty, are also at stake under extraterritorial processing. When it comes to the former, the Court of Justice has addressed this issue numerous times, particularly concerning asylum border procedures, under Articles 6 and 47 of the Charter. 115 For example, the Court has ruled that the right to a hearing is an ‘essential procedural requirement’ that cannot be waived due to time constraints. 116 As noted, the generally short deadlines in the border procedure, combined with the remote nature of the personal interview and the lack of physical access to legal advisers, could compromise the right to be heard and, consequently, the right to an effective remedy.
Regarding the right to liberty, the Court has emphasized that any restrictions must be strictly necessary, based on an individual assessment, and respect the principle of proportionality. 117 The risk of arbitrary use of border procedures in Albania, along with the concrete difficulties of accessing alternatives to detention, serves as a clear example of the risks linked to the Protocol's implementation, which could raise concerns about compliance with Article 6 of the Charter. 118
Prospective scenarios for defining the relationship between extraterritorial asylum processing and EU law and concluding reflections
In this article, we have aimed to show that, despite the absence of explicit EU rules for extraterritorial asylum procedures, a ‘functional link’ exists between the asylum process in Albania and the EU legal system, bringing it under scope of Union law. This link stems from the direct application of the QD given its lack of territorial scope, the potential transfer of individuals eligible for protection to the EU and the impact that these procedures might have on the objectives of the CEAS, particularly the risk of undermining the goal of having a unified asylum procedure with common reception standards and of altering the Dublin Regulation's responsibility-sharing mechanism.
If, on the contrary, it is accepted that asylum processing outside the EU's territory is not connected to EU asylum law, the two systems would function independently. In this case, the legitimacy of the Protocol and its effects on the rights of asylum applicants would be evaluated solely based on international human rights law, including the European Convention on Human Rights and the Italian Constitution. In our view, this would, however, lead to unfair discrimination between applicants whose claims are processed within the EU and those detained in Albania, even though the procedural rules in Albania are aligned with EU law, just as in Italy. The unequal treatment would be even more apparent since the ratification law gives Italian authorities the power to assess asylum applications and appeals in Italy. This means that Italian officials would apply the same laws in both Italy and Albania, but with different legal consequences. If national law conflicts with EU law, a judge in Italy deciding on applicants within Italy could disapply the conflicting national law based on the principle of primacy, but this would not be possible for applicants detained in Albania.
However, this scenario appears unlikely in light of the examined case law. We think, instead, that the Court of Justice would have full competence to assess whether extraterritorial asylum processing complies with EU law, and that its intervention would ensure legal certainty, as well as the uniform application and effectiveness of EU law. This could reasonably be done, for instance, through infringement proceedings or preliminary rulings in the context of procedures involving the detention of asylum applicants in Albania.
The Court's intervention could lead to several outcomes. Firstly, the Court might declare the irreconcilability tout court of the off-shore processing with the objectives of the CEAS, for example by finding the scheme in violation of the Dublin Regulation's responsibility-sharing system or by deeming the use of a third country's territory as a ‘border zone’ for the purpose of conducting a EU-regulated accelerated border procedure inconsistent with the concept of ‘border’ in EU asylum law. Alternatively, the Court could find that a Member State is not generally prohibited from extending EU-regulated asylum procedures to a third country, as long as the APD and RCD provisions on procedures and reception standards are met. Finally, the Court might conclude that extraterritorial processing does not inherently violate CEAS objectives, and that Italy is not required to comply with the standards provided in the APD and RCD, given their territorial scope, but that compliance with the Charter would still be required since Italy is implementing the QD.
While we contend that the analysis presented in this article supports the conclusion that the system created by the Italy–Albania Protocol cannot be reconciled with the objectives of the Dublin Regulation, that extending the notion of ‘border areas’ to third countries would lead to the arbitrary use of border procedures and that processing in a third country would lower the standards set by the APD and the RCD and risk violating the Charter, the ultimate decision will eventually rest with the Court of Justice. In the upcoming years, the Court will most likely face the delicate task of assessing whether EU asylum law aligns with the growing practice of cooperating with third countries to restrict effective access to territorial asylum – a practice which for the first time in the EU also involves extraterritorial processing. Despite the newly adopted EU migration and asylum pact, the new rules do not seem to resolve the unfair distribution of responsibility for refugees protection among Member States. This imbalance may, in the future, drive a stronger push to shift international protection and migration management responsibilities outside the Union, fostering ‘experimentation with legal innovations’ designed to ‘remove these arrangements from EU scrutiny’. 119
Footnotes
Acknowledgements
The authors extend their sincere gratitude to Mariolina Eliantonio and the anonymous reviewers for their insightful comments and valuable suggestions on an earlier draft of this paper. We also thank the participants for their helpful feedback during academic events organized by the Maastricht Centre for European Law (MCEL), the Bocconi Lab for European Studies (BLEST), and the Utrecht Centre for Regulation and Enforcement in Europe (RENFORCE), where the main arguments of this paper were initially discussed. Any remaining errors are entirely our own.
Declaration of conflicting interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The authors disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the H2020 Marie Skłodowska-Curie Actions (grant number 847596).
