The “Innovative” Italy-Albania Deal under Scrutiny
In 2023, Italy and Albania concluded a Protocol that allows Italy both to process asylum applications through an accelerated border procedure lasting up to 28 days and to carry out returns in Albania. Italy has built two centers for this purpose that opened on October 14. One, in Shengjin, is meant to register asylum applications and identify vulnerable applicants who cannot be subjected to the border procedure and should thus be returned to Italy. The second, in Gjader, is divided into three sections: one, with 880 places, for asylum seekers; another, with 144 places, for migrants awaiting repatriation; and a third, with 20 places, for migrants undergoing criminal proceedings. According to the Protocol the capacity could increase to up to 3000 places in total (Art. 4).
The agreement has been praised by the President of the European Commission as an “out-of-the-box thinking”, and several member states view the deal as a pilot for assessing the feasibility of similar “innovative” models. While the Italy-Albania deal differs from traditional EU externalization tools, legal challenges could undermine its success, as a recent ruling by Italian judicial authorities that has temporarily halted its implementation demonstrates.
The EU’s Externalisation Toolbox: what’s new?
While its exact definition is still debated, the UN High Commissioner for Refugees (UNHCR) defines externalisation as actions beyond the border which, directly or indirectly, prevent asylum seekers from reaching a specific destination or from claiming protection. They become unlawful when they lack proper safeguards, shifting responsibility to other states without ensuring effective access to protection.
Externalisation strategies can be grouped into three different categories.
The first includes transferring asylum seekers to third countries after they arrive in the territory. It relies on safe country notions to: a) declare asylum applications inadmissible, using the safe third country (STC) rule, as seen with returns to Turkey under the EU-Turkey Statement, or the first country of asylum concept; or b) speed up their processing when applicants come from countries generally considered safe, using the safe country of origin (SCoO) concept. These procedures usually occur at borders or in transit zones and use the legal fiction of non-entry to justify restrictions of movement or deprivation of liberty.
The second involves measures taken before potential applicants reach the country’s territory. It mainly relies on outsourcing border control to third countries, like the EU-funded cooperation with the Libyan Coast Guard.
The third strategy is offshore processing, whereby asylum seekers are sent outside the country for status determination. While new in the EU, it is often discussed during spikes in arrivals. This model is based on the U.S. practice in the 1990s at Guantanamo Bay, where Haitians intercepted at sea were processed. Recognized refugees were relocated to the U.S., while others were repatriated. This policy also inspired the controversial Australian Pacific Solution in 2001.
The novelty of the Italy-Albania Model
Against this backdrop, the Italy-Albania model is unique. It operates on the premise that the Asylum Procedure Directive (APD) requires member states to assess asylum applications in their territory, or at the border or in transit zones geographically located within the territory, only when they are made within their territory, at the border, in transit zones, or in territorial waters (Article 3 APD). Based on this reasoning, when applications are made in international waters, Italy can assess them in border areas that are under its jurisdiction but outside its territory (Article 4 Protocol). For this purpose, law n. 14/24 equates the two centers in Albania to Italy’s “border areas.” This means that migrants are treated “as if” they were in Italy, with national law implementing the relevant EU legislation regulating the processing and potential return. The administrative and judicial authorities in Rome are responsible for handling asylum, detention, and repatriation matters (Article 4).
In this context, the Italy-Albania model, by de-territorializing the EU asylum border procedure, draws on features from all three existing types of externalization methods to create something new. First, it implements an accelerated border procedure based on the SCoO concept (category 1). Second, it prevents physical access to the territory (category 2). Lastly, it involves external processing, with Italy remaining responsible for relocating those granted international protection and returning those who do not qualify for entry (category 3).
The temporary suspension of the Italy-Albania Protocol
On October 18, the Civil Court of Rome refused to validate the detention of 12 asylum seekers from Bangladesh and Egypt, ordering their return to Italy, in what has been referred to as a “heavy blow” to the deal. However, the judges did not evaluate the legality of the deal itself; instead, they focused on whether the applicants could be detained under the procedure. Relying on a recent decision of the European Court of Justice (ECJ), they held that the accelerated border procedure should not have been applied. As such, it was not necessary to assess the necessity and proportionality of their detention in Albania.
On October 4 , the ECJ had ruled that member states cannot designate third countries as an SCO with territorial limitations. The Court reached this conclusion through a literal and contextual interpretation of the APD. It determined that a third country can only be classified as a SCoO if it is generally safe and free from persecution, torture, or inhumane treatment, including serious harm for conflict-based violence, throughout its entire territory (paras. 35 and 66). To support this interpretation, the Court noted that an amendment to Article 30(1) of Directive 2005/85, which was in effect before the current version, explicitly removed the option to consider third countries as safe if there were exceptions. This indicates that lawmakers did not intend to allow such exceptions (para. 75).
Italy’s list of SCoO is accompanied by unpublished country reports obtained by the Italian Association for Judicial Studies on Immigration (ASGI) through a Freedom of Information Act request. They indicate that some countries, including Bangladesh and Egypt, are considered safe with territorial limitations and exceptions for specific at-risk groups, such as the LGBTIQ+ community, victims of female genital mutilation, and human rights defenders. The Court of Rome thus found that this designation as SCoO was incompatible with the ECJ’s ruling, leading to an order for the applicants to be transferred back to Italy.
Notably, the asylum seekers detained in Albania had their applications processed in record time, with rejections communicated in 24 hours, even before the end of the hearing to validate their detention, which occurred 48 hours after the order was issued. However, since the judges ruled that the border procedure was not applicable, their appeals must now be considered within Italy.
The government’s response
The government responded quickly by appealing the ruling and issuing a law decree published on October 24. The decree updated the list of SCoO, removing those that were previously listed as safe with territorial exceptions. However, countries with risk-based exceptions, including Egypt and Bangladesh, remain on the list. The new decree also allows the decision by first instance tribunals to be appealed on the merits to the appellate courts, instead of going directly to the Court of Cassation, which can only rule on points of law. The aim seems to be to increase the chances of overturning unfavorable rulings from lower courts.
The government’s strategy appears to be to claim that the ECJ ruling only applies to territorial exceptions. However, this argument is unconvincing. The removal of the option to designate third countries as SCoO in the amendment of the 2005 version of the APD did not only concern territorial exceptions, but also those based on specific groups of people (see Article 30(3)). Hence, the ECJ’s requirement that a country must be generally safe from persecution and serious harm appears valid even for exceptions based on at-risk categories, as stated by the Court of Rome.
The future of the Italy-Albania Deal
If Italian authorities continue to transfer asylum seekers from SCoOs with exceptions for certain groups, judges might respond in two ways. They could either (a) continue to reject detention orders, arguing that Italy’s designation of an SCoO still does not comply with the ECJ ruling, or (b) pause the case and refer a question to the ECJ. It is worth noting that a question on this issue, currently pending, was already submitted by the Court of Florence. While the Court of Rome felt that the ECJ’s ruling of October 4th was enough to clarify the matter, judges may in the future prefer to let the ECJ make the final decision, especially given government pressure.
Even so, if the only issue pertains to how Italy applies the SCoO concept in border procedures, the deal with Albania may not be as dead as it seems. The reasoning of the Court of Rome suggests that the judges would have rejected the validation of detention orders even if the applicants had been detained in border areas physically within Italy. In other words, the judges did not focus on the fact that the applicants were in Albania, but merely that they should have been redirected to the regular asylum procedure in the territory, regardless of Albania’s designation as a border area.
This means that applicants from SCoOs that were previously listed without exceptions—Albania itself, Bosnia, Kosovo, North Macedonia, Montenegro, Serbia, and Cape Verde—could still be transferred. Italy might also try to claim that the designated countries are completely safe, as the new decree provides that information concerning the reasons of the listing will only be provided in January. Additionally, the ruling could be bypassed in the future because the new Asylum Procedure Regulation explicitly provides that: 1) a SCoO can be listed with exceptions for specific parts of their territory or clearly identifiable categories of persons (Art. 61(2) APR), and 2) an accelerated border procedure is applicable not only to asylum seekers coming from SCoO, but more broadly to applicants coming from countries with a with a recognition rate below 20% (Art. 42 (j)).
Alternative legal responses
However, other legal issues may emerge. Italian authorities have committed to applying the same standards as if the applicants were in Italy, but it is questionable whether this can be ensured in practice. For instance, vulnerability assessments, which are crucial to identifying those who cannot be processed under border procedures, 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. Additionally, authorities might be reluctant to conduct thorough assessments if they could result in transferring applicants back to Italy, given the high costs involved. Moreover, the law limits in-person meetings between applicants and lawyers to situations where remote counseling is impossible (Article 4(5) of law n. 14/24). This raises concerns about the ability to guarantee the right to defense effectively. This, along with the challenges of independent monitoring in a third country, raises doubts about the fairness of the process. Italy’s past failures to meet reception standards in hotspots, sanctioned by the European Court of Human Rights, suggest that similar legal challenges regarding inadequate reception conditions in Albania could arise in the future.
Furthermore, since Italy has extended the applicability of a EU-regulated procedure to a situation not explicitly covered by it, the Court of Justice may be asked to clarify whether concepts derived from EU law are being properly applied in the context of the procedure in Albania. This is important to ensure a consistent interpretation of EU law and to avoid undermining common objectives, such as creating a uniform asylum procedure across EU member states. A similar situation occurred when the Netherlands expanded the right to family reunification under the Family Reunification Directive to EU citizens who had not exercised their right to freedom of movement, and the ECJ was called to interpret EU-based concepts. In the case of the Italy-Albania deal, the Court could be asked, for instance, whether the “non-entry fiction” can be interpreted broadly enough to designate an area located in a third state as a border area for conducting accelerated border procedures. This is particularly relevant given the challenges in guaranteeing the same standards in practice, as previously discussed.
Conclusion
While the Court of Rome temporarily halted the implementation of the deal, this might not be the end of it, especially with the expansion of the scope of border procedures in the forthcoming APR. With both the European Commission and the European Council closely monitoring the developments of the deal, and examining other innovative ways to cooperate with third countries – including through the removal of the connection requirement to apply the STC concept, à la UK-Rwanda, and the revision of the Return Directive to create “return hubs” for rejected asylum seekers – it is likely that the Meloni government will employ all possible tools to make the plan work. Yet, additional legal challenges linked to the implementation of policies of cooperation with third countries, which do not adequately pay attention to fundamental rights risks, might slow down or hinder this and other similar forward-thinking initiatives.
