POST 8 OF THE SERIES OF THE ODYSSEUS BLOG ON THE PACT ON MIGRATION & ASYLUM
By Daniel Thym, Director of the Research Center Immigration & Asylum Law, University of Konstanz
The spectre of asylum procedures in third states has been haunting EU asylum policy since its inception. The London Resolutions of 1992 promoted the idea. Ten years later, the British government under Tony Blair (Labour) sparked a lively political debate. So far the idea has never been put into practice on a larger scale, with the exception of the EU-Turkey-Statement to which we shall come back. The scarcity of practical experience is one factor explaining the startling mixture of enthusiasm and distrust any debate about the topic inevitably raises. The idea is bound to gather momentum during term of the next European Commission.
To prevent confusion, we should distinguish asylum procedures abroad (‘external processing’) from ‘safe third country’ schemes. The former (‘external processing’) concerns regular European asylum procedures in a country outside the EU. The Italian ‘Albania model’ follows that rationale. Two centres are currently being established in Albania where Italian officials will undertake regular asylum procedures, presumably via videoconferencing. Beneficiaries of international protection will be relocated to Italy, as will returnees whenever (voluntary) return fails in practice.
By contrast, the British ‘Rwanda plan’ was supposed to send asylum applicants who had entered the UK irregularly to Rwanda on the basis of an inadmissibility decision by the British authorities. They will receive an asylum procedure by the Rwandan authorities in accordance with domestic laws, without the option of eventual legal entry into the UK. The ‘safe third country’ provision in the new Asylum Procedures Regulation (EU) 2024/1348 follows the basic contours of that project.
In its 2024 manifesto, the European People’s Party (EPP), ‘advocate(s) a fundamental change in European asylum law’ on the basis of the safe third country concept (here, p. 6). The outgoing Commission President supports the plan half-heartedly. She explicitly lent her support to ‘smart’ policies in a letter to the European Council, whose strategic vision for 2024–27 has recently called upon the EU institutions to ‘consider new ways to prevent and counter irregular migration’ (here, Annex). Such references to ‘smart’ and ‘new ways’ are shorthand for the involvement of third states—an idea that has been pushed by Italy, Denmark, the Czech Republic, Austria, the Netherlands, and several other governments for some time.
Any debate about safe third country schemes must distinguish four questions. First, the significance of the ‘small print’ for the practical realisation and the legal assessment. Secondly, factual uncertainties as to whether such project will have a ‘deterrent effect’ which proponents promise and in the absence of which states will not usually be able to invest piles of cash into the realisation of safe third country schemes.
Thirdly, the new Asylum Procedures Regulation (EU) 2024/1348 requires Member States to perform an inadmissibility check, including the option of legal remedies, before sending asylum applicants to a safe third country. These admissibility procedures might prove a critical legal and administrative bottleneck, thus possibly triggering a swift revision of the newly adopted legislation as early as next year. The incoming Commission will be under an obligation to table a report about ‘targeted amendments’ on safe third countries in the summer of 2025 (here, Article 77(4)). That timing was deliberate given that majorities in the Council and the European Parliament might have shifted by the end of 2025. The German Christian Democrats, which are likely to retake the chancellery in the autumn of 2025, are fervent supporters of safe third country schemes (here, p. 23-24).
Fourthly, full respect for the principle of non-refoulement by the third state is a precondition for any safe third country arrangement, as is the provision of ‘effective protection’ in terms of living conditions there. Living conditions are an area where the new legislation introduces major changes in accordance with the plea, by the European Council of June 2017, to align protection standards with international minimum requirements, thus abandoning higher European standards (here, No. 23).
‘Yes but’: Relevance of the ‘Small-Print’
Our rough comparison of external processing in the form of the Italian ‘Albania model’ and safe third country schemes like the British ‘Rwanda plan’ deliberately ignored political and legal controversies surrounding both projects. Space precludes their detailed analysis in a contribution focusing on the new EU legislation. Having said this, it is precisely the ‘small-print’ which defines the legal assessment of any safe third country arrangement. Both the European Court of Human Rights (ECtHR) and UNHCR recognise that the prohibition of refoulement permits safe third country schemes, provided they fulfil several conditions (here, paras 128-138; here, No. 2).
These conditions lay beneath the judgment of the UK Supreme Court of November last year. Judges were not convinced that the Rwandan asylum system offered sufficient security, as required by the European Convention on Human Rights (ECHR). In case the British Labour party wins the election on 4 July and abandons the ‘Rwanda plan’, we may never learn whether the revised Treaty, negotiated by the British government at short notice, would have been by British courts and the ECtHR. For our purposes, these debates offer limited guidance anyway. EU institutions cannot simply emulate the British ‘example’ of instructing national courts not to assess the safety of third states, as the British government did in legislation adopted after the Supreme Court ruling. Domestic courts in the Member States will enforce the requirements under the Charter of Fundamental Rights, which effectively incorporates the ECHR into EU law.
Factual Uncertainties: ‘Deterrent Effect’ in Practice
In another respect, however, any implementation of both the British ‘Rwanda plan’ and the Italian ‘Albania model’ might prove instructive. Safe third country models habitually assume that it is sufficient to return a limited number of people, whereupon the number of arrivals would fall automatically. Australia is the only country which has ‘achieved’ such a deterrent effect in practice, albeit at the price of massive humanitarian costs. Australia transferred 4,100 people to reception centres on remote Pacific islands – a transfer rate of almost 100% if we exclude those being ‘pushed back’ after a rudimentary screening on the high seas. After having implemented that scheme systematically for about one year during 2013/14, the number of maritime arrivals decreased sustainably.
Against this backdrop, a decisive question for European countries will be whether a similar deterrence can be achieved with much lower transfer rates. Italy plans to accommodate a maximum of 3,000 people in Albania at any given moment—much less than the number of arrivals per month during the summer (to increase the likelihood of ‘success’ may be one reason why the government has postponed the implementation of the scheme until the autumn when the number of arrivals decreases). The situation would have been similar under the British ‘Rwanda plan’. The government has never said publicly how many asylum seekers it plans to transfer to Central Africa; the media reported about a few hundred – out of many thousand applicants via the Channel during the first months of this year.
In the case of the EU-Turkey Statement, several factors explained the substantial reduction in arrivals to Greece following the announcement of the safe third country scheme on 16 March 2016: closure of the Western Balkans route; better living conditions in Turkey; difficult situation on the Greek islands; end of the ‘battle for Aleppo’; Turkish border controls; etc. The significance of contextual factors was reaffirmed by continuous low numbers of arrivals, once it became apparent that the Greek authorities were unable to transfer a substantial number of Syrians to Turkey for legal and practical reasons; the overwhelming majority remained in Greece. We shall learn what the practical effects of the British and Italian policies will be if they are implemented in the coming months. These factual considerations come on top of the political and legal conditions.
Admissibility Procedures on European Soil
Whenever someone reaches the Schengen area and applies for asylum, Member State must respect the procedural safeguards laid down in the present Asylum Procedures Directive and new Asylum Procedures Regulation, which will apply from 12 June 2026 onwards. Their prescriptions authorise transfers to safe third countries only after an admissibility check on European soil, including the option of legal remedies. These procedural safeguards are a critical legal and administrative bottleneck for any safe third country scheme – as exemplified by the EU-Turkey Statement. The long duration of admissibility checks and legal oversight was one of several reasons why few Syrians were transferred back to Turkey. The Greek authorities took months to decide, as did the Greek courts which, moreover, annulled many transfer decisions on legal grounds.
Italy will have recourse to a legal ‘backdoor’ when implementing the ‘Albania model’, since the procedural safeguards in the EU’s asylum legislation become applicable whenever someone makes an application for asylum ‘at the border, in territorial waters or in transit zones’ (here, Article 3(1)). Migrants and refugees rescued on the high seas cannot invoke these guarantees enshrined in EU legislation, meaning that ‘only’ fundamental rights apply (here, paras 238-242). Italy will not have to implement an admissibility check for the simple reason that it remains fully responsible for the treatment of asylum applicants in line with Article 1 ECHR, since its officials will exercise the effective control over the reception centres on Albanian soil. Italy will not, in other words, surrender effective control over the transferees to a third country. That rationale cannot be extended to safe third country schemes following the basic rationale of the British ‘Rwanda plan’.
Connection Criterion: the ‘Anti-Rwanda’ Rule
When the Council reached its ‘historic’ compromise on the reform of asylum legislation in June 2023, the German government insisted on the ‘connection criterion’ according to which asylum seekers cannot be returned to a safe third country in the absence of a connection ‘on the basis of which it would be reasonable for [the asylum seeker] to go to that country’ (here, Article 59(5)(b)). At the bequest of the European Parliament, a non-binding opening clause for scenarios of transit was discarded at trilogue stage. The final wording is identical with the present Article 38(2)(a) of Directive 2013/32/EU which a CJEU had interpreted so that short-term transit was insufficient (here, paras 44-50). It is unclear when transit turns into a ‘connection’ that makes return ‘reasonable’. The Greek Council of State held a prior stay of 1.5 months in Turkey to be enough.
The connection criterion could be discontinued. UNHCR recommends it politically (here, No. 5), while recognising that it ‘is not mandatory under international law’ (here, No. 6). Its purpose has traditionally been to guarantee admission by the third state. This was not always the case in the early days of safe third country schemes when states refused asylum applications as being inadmissible on the basis of national laws without the guarantee of actual return. That risk does not exist under the new Asylum Procedures Regulation which, like the present Directive, only authorises the rejection of an application as inadmissible if a safe third country readmits the person (here, Articles 38(1)(b) and 59(9)). Whenever the transfer fails, a regular asylum procedure must be carried out on EU territory—as Advocate General Pikamäe has recently reaffirmed (here, points 59-64)
The alternative objective of the connection criterion, mentioned occasionally these days, to avoid secondary movements and to ensure local integration, is legitimate but cannot change the lack of binding legal force. The debate about whether it should be discontinued may soon resurface. The Commission report about ‘targeted amendments’ in the summer of 2025, mentioned previously, was the ‘price’ sceptical governments insisted upon to accept the German insistence on maintaining the connection criterion. In practice, it prevents any Rwanda-style cooperation without the prior amendment of EU legislation. The Italian ‘Albania model’ need not comply with the connection criterion, since it will only cover people rescued by state ships on the high seas where the Asylum Procedures Directive/Regulation does not apply, unlike in the territorial waters.
Legal Remedies in- and outside Border Procedures
Legal remedies guarantee that no transfer takes place without a domestic court having given its consent to an administrative decision rejecting an asylum application as inadmissible on the basis of a personal interview. A seemingly technical question concerns the nature of court intervention which can be either the preliminary injunction or the final judgment. This choice can have significant implications. Within Germany, a decision on interim remedies is taken within 40 days on statistical average, while the final judgment is being delivered after two years in asylum matters.
Against this backdrop, it is of great importance that the European Parliament rejected the Council negotiating position that interim legal protection was sufficient. The provisional right to remain until the final judgment remains fully intact for inadmissibility decisions based on the safe third country concept (here, Article 68(1), (3)(b), (4) does not cover Art. 38(1)(b)). The only exception are border procedures. That prevents countries within the Schengen area from sending anyone to safe third countries quickly, whereas the actual transfer can take place after interim legal protection for inadmissibility decisions in the border procedure (Article 68(3)(a)(ii)). It can be compatible with the Charter to allow transfers after interim legal protection examining ‘both facts and points of law’ (Article 68(4)). That is exactly what the ambiguous reference to ‘automatic suspensive effect’ in ECtHR and CJEU case law requires (here, §§ 198-199; here, paras 51-53)
Note that asylum applicants covered by a safe third country arrangement may always be subject to border procedures (Article 44(1)(a)), even if they come from a country with high statistical recognition rates, such as Syria (Articles 44(1)(b), 42(1)(j)). We can expect the debate about ‘targeted amendments’ to re-examine the nature of judicial intervention and, possibly, the application of the border procedures within the Schengen area. After all, one may expect ‘diversion effects’ if, by way of example, applicants ‘escaped’ to Belgium when the Netherlands started operating a safe third country scheme. Diversion to the Western Balkans may possibly even be a hidden objective, or at least a side-effect, of the Italian ‘Albania model’ in case applicants can leave the reception centres de facto.
Situation in the Third State
Any safe third country scheme will usually proceed in three steps. First, governments participating in the scheme will negotiate the practical arrangements with the third state, either in the form of an international treaty or of a non-binding memorandum of understanding. That can be done by a single Member State or a ‘coalition of the willing’, possibly with the support of the EU institutions, for instance through the deployment of Frontex or the EU Asylum Agency. Secondly, the third state must be categorised as safe at national or European level, with both solutions having analogous practical effects (here, Articles 63 and 64). Finally, each asylum application must be assessed in accordance with the procedural requirements mentioned previously. Throughout these different steps, the level of safety a third state must provide is of paramount importance.
Comprehensive Protection against Refoulement
The prohibition of refoulement is the gold standard of international refugee law. It prohibits any transfer to states where someone would be persecuted or subject to grave human rights violations. It is beyond doubt that safe third countries have to provide comprehensive protection against illegal refoulement, including indirect ‘chain refoulement’ to another country that is not safe. EU legislation confirms that the level of protection follows the wide European understanding of non-refoulement, including serious and individual harm during civil wars or, by way of example, gender-specific persecution by nonstate actors (here, Article 59(1)(a)–(c)).
Whenever a third state does not interpret the prohibition of refoulement accordingly, exceptions for specific categories of persons are warranted (Article 59(2)). A classic example is gay men in case of safe third countries on the African continent. A novelty of the new legislation is that there can be exceptions for ‘specific parts of its territory’, meaning that the insecurity of one region does not hinder the categorisation of the remainder of the country. By contrast, it appears to be insufficient, in light of the wording, that safety is guaranteed within a limited geographic area only, for instance in a reception centre or refugee camp.
The level of safety must be confirmed by reliable administrative practices, not the law in the books or the ratification of international treaties. The ‘Rwanda judgment’ of the UK Supreme Court is an exemplar demonstration of how such scrutiny should unfold. Judges prohibited all transfers to Rwanda in light of reliable international reports that the domestic asylum system was not effective: it had dealt with few cases, Syrians and Afghans had rarely received protection, and a safe third country deal with Israel had resulted in illegal refoulement (here, paras 74-97). It should be noted that status determination by the third state follows its domestic legislation, not the procedural safeguards in the Asylum Procedures Regulation. The only condition is that procedures are robust enough to ensure unconditional respect for the prohibition of refoulement by the third state.
In practice, these statutory safety requirements can prove challenging. Most countries in the global South, including on the Southern shore of the Mediterranean, have ratified the Refugee Convention and other international treaties but they have mostly not established robust administrative asylum systems. Status determination by UNHCR does not provide an unfettered guarantee of safety either. UNHCR delivers most valuable work under difficult conditions, but it cannot always and automatically guarantee the level of protection required under European human rights law (here, §§ 153-156; here, §§ 287-291). As a result, reliable asylum procedures in third states for a significant number of applicants presuppose active support by international actors. EUAA and Frontex may become active on the basis of their respective mandate (here, Articles 71-74; here, Article 35).
Living Conditions in Accordance with International Law
In June 2017, the European Council had called upon the legislature, as we have seen, to lower the standards for socio-economic living conditions in safe third countries to international minimum requirements. That plea was followed through with determination. The legal background is the phased guarantee of socio-economic rights in the Refugee Convention. More specifically, we may distinguish rights requiring physical ‘presence’ (e.g. education and non-refoulement under Articles 22 and 33), ‘lawful presence’ (e.g. self-employment and freedom of movement under Articles 18 and 26), and ‘lawful stay’ or ‘permanent residence’ (e.g. employment, welfare, and social security under Articles 17(1), 23 and 24). While the first category is available during the asylum procedure, the third stage presupposes formal refugee status which states are, according to the habitual reading, not obliged to deliver qua international refugee law.
The implications of this graded approach are legally contested. State practice and the judgments of domestic courts on safe third country schemes indicate that safe third countries do not have to provide full access to the third stage (here, pp. 8-9, 18-22), although some academics claim such outcome is contrary to the spirit of the Refugee Convention (here; here). EU institutions sided with the former view, also considering that the precise meaning of international refugee law can remain ambiguous in the absence of an international court interpreting it authoritatively.
On that basis, the legislature introduced the generic notion of ‘effective protection’ in Article 57(1) Asylum Procedures Regulation (EU) 2024/1348. It will serve as the reference point for the legal and factual assessment of living conditions in safe third countries. Three crucial changes have been implemented in comparison to Article 38(1) Asylum Procedures Directive 2013/32/EU (for a detailed assessment, see here, pp. 49-54):
(1) an ‘adequate standard of living with regard to the overall situation’ shall include access to food, clothing, [and] housing or shelter’ (Recital 51): this is more precise but also less demanding than the previous guarantee of equal treatment with nationals in terms of access to social benefits;
(2) access to healthcare and ‘essential treatment’ under the conditions generally provided for in that third country replaces the former guarantee of equal treatment with nationals of the third state;
(3) authorisation to remain on the territory, presumably in the form of a residence permit, for as long as there is a need for protection will supplant refugee status ‘in accordance with’ the Refugee Convention. That amendment answers the former dispute whether formal ratification of the Refugee Convention without geographical limitation is mandatory under the former Article 38(1)(e) Asylum Procedures Directive. UNHCR highlights that ratification is ‘desirable’ but not mandatory; what matters is that protection against refoulement is available in law and in practice.
The only area where the notion of ‘effective protection’ is more generous than international minimum requirements is labour market access, which, according to Recital 51, must be given under the same conditions as for any other foreigner residing in the safe third country. There is a political rationale behind this guarantee the European Parliament had insisted upon. Labour market access upon completion of status determination guarantees that transferees can build a life of their own in the safe third country, as foreseen by the Global Compact on Refugees. EU legislation opts against the dystopia of the Australian ‘Pacific Solution’ which had left returnees detained in refugee camps on remote Pacific islands for years. Any safe third country scheme in accordance with the Asylum Procedures Regulation will have to sponsor local integration into host states.
Conclusion
References to ‘smart’ and ‘new ways’ in policy documents circulating in Brussels before the summer break indicate that the adoption of the New Pact will not put an end to forward-thinking initiatives. Safe third country schemes are poised to play a critical role in these debates. They might even, as we have seen, involve ‘targeted amendments’ of the new legislation as early as next year. Political controversies may start soon when the Commission President assigns the justice and home affairs portfolio to the future Commissioner and Vice President and when these two appointees are being grilled in the European Parliament’s LIBE Committee before the approval of the college of Commissioners this autumn. These debates are likely to concentrate on whether there is a political desire to implement a safe third country scheme. If that was the case, the diplomatic, legal, administrative, and practical hurdles might prove challenging.
Disclaimer: The author has served as a witness in hearings of the German government on a potential realisation of the safe third country scheme in the run-up to the interim report presented in June 2024; in this broader context, he wrote an expert opinion on legal requirements for safe third countries in asylum law and practical implementation options.
