From Solidarity to Exclusion: The ‘Safe Country’ Concept in UK Asylum Law and the Irony of Borders

Abstract

This article argues that the asylum policy and legislative changes introduced by the UK government in the years 2022–2024 altered the original meaning of the concept ‘safe country’ as understood in international and EU law. The UK modified this concept, which from a solidarity concept became a means of exclusion, and which negatively affects the lives and rights of people seeking asylum in the UK. Using a doctrinal approach, the first part of this article sets the legal and historical context of the concept ‘safe country’. Departing from the analysis of the Refugee Convention, the article discusses how this mechanism was used by the EU legislation. From an idea of solidarity among EU Member States, it shifted from responsibility-sharing to burden-sharing while still allowing some guarantees to people seeking asylum. Using content analysis, the second part of this article evaluates the legal requirements set by the UK legislation together with implications of applying the ‘safe country’ concept to the asylum claims. It argues that, in recent years, the UK Government used the term ‘safe country’ as synonym of two (possibly three) different concepts, such as ‘first safe country’ and ‘safe third country’. It also shifted and pushed its meaning beyond the current commonly agreed interpretation of the term because it eroded the requirement of a link between the person seeking asylum and the ‘safe country’. Thus, the UK legislation deviated even further from the rationale underlying the Refugee Convention, international human rights standards and EU legislation because it passed the obligation to assess asylum claims to states with no link to people seeking asylum and without adequate risk assessment. The final part of this article discusses the limit to this policy and analyses the legal battle between the UK Parliament, the Government’s executive power, the UK Supreme Court and the Belfast High Court, which barred the UK Government from deporting people seeking asylum to a third country. This article concludes that there is some irony in the fact the term ‘safe country’ has been weaponised as a bordering tool by the UK Government, but ‘a border’ between the Republic of Ireland and Northern Ireland is limiting the negative effect of the concept ‘safe country’ on the very same people that is attempting to exclude from protection.

1. Introduction

This article examines the controversial interpretation and application of the concept ‘safe country’, which has become central in the UK asylum law following the proliferation of restrictive legislation enacted by the Conservative government across the years 2022 to 2024. From a broader perspective, this article illustrates the use of conceptual ambiguity as a legal strategy to allow domestic authorities to move people seeking asylum between jurisdictions, borders and different legal regimes. This approach often results in outcomes that advantage states. However, it fails to extend the same beneficial principles and protection to individuals seeking asylum.

The concept of a ‘safe country’ indicates “a procedural mechanism for shuttling asylum seekers to other States said to have primary responsibility for them, thereby avoiding the necessity to make a decision on the merits because another country is deemed or imagined to be secure” (Goodwin-Gill and McAdam 2010, p. 392; Foster 2007, pp. 223–24; Piotrowicz 2002, p. 132). This means that the UK could reject asylum claims without the examination of substantive asylum grounds and individuals’ circumstances. Following that, the UK could return individuals seeking asylum to a different country. Here, the term ‘safe country’ comes into play. It has two main connotations: either as first safe country of asylum or as safe third country. The first term indicates that individuals can be returned either to countries where applicants have already been recognised as a refugee or where they could have availed themselves of sufficient protection, such as non-refoulement or, according to the second connotation, individuals could be returned to any other country countries where in the view of the competent authorities, people seeking asylum have a connection and are not subjected to direct or indirect refoulement.

This article expands the findings of previous scholarship on the implementation of EU law in the UK statutes before Brexit (John-Hopkins 2009), on the externalisation of asylum after Brexit (Powell 2024; Saenz Perez 2023) and, more recently, on the recognition of Rwanda as a safe third country (Pulvirenti et al. 2024; Robinson 2023) because it shows that the UK asylum scheme borrowed the expression ‘safe country’ (intended as both safe country of origin and safe third country) from EU law, incorporated it in its own statutes but it did not distinguish between the two terms. Conversely, it mixed both concepts, took advantage from the ambiguity of this term with the excuse of maintaining internal social cohesion and eroded international protection standards (Moreno-Lax 2015, p. 720). As will be explained at the end of this article, the UK Government ultimately created a third category of ‘safe third country’ which does not require any links between that specific state and the individuals claiming asylum. This generates uncertainty and inconsistencies in the interpretation and application of asylum law in the UK and facilitates the transfer of individuals seeking asylum across multiple jurisdictions. While this practice serves the interests of the UK in safeguarding its own borders, it does not adequately safeguard people seeking asylum, thereby undermining the principles of fairness and individual assessment enshrined in international refugee law.

The first part of this article (Section 2) focuses on the concept of ‘safe country’ using the vast body of the literature, which discusses the safe country term from different angles, from lawfulness of the concept itself against the Refugee Convention (Gil-Bazo 2015; John-Hopkins 2009; Marx 2019; Moreno-Lax 2015), to the specific requirements for a third country to be considered safe (Frelick et al. 2016; Goodwin-Gill and McAdam 2010; Hathaway and Neve 1997; Lambert 2012; Legomsky 2003; Marx 2019; Phuong 2005). Scholarship has widely applied those latter principles to specific case studies, such the EU legislation (including the agreement with Turkey) (Allain 2002; Costello 2005; Legomsky 2001), Tunisia (Giuffré et al. 2002), Israel and Australia (Bar-Tuvia 2018; McAdam and Purcell 2008), including several comparative studies between different systems (Abell 1997; Caron 2017; Gil-Bazo 2015; Frelick et al. 2016; Hyndman and Mountz 2008) and readmission treaties between the EU and Turkey (Arıner and Kader 2022; Ovacik 2020; Yavuz 2019) and Spain and Morocco (Caron 2017). Building on this scholarship, the aim of this section is to analyse how the term ‘safe country’ has changed over time and, for the first time, to analyse how States interpreted it from a solidarity concept (Marx 2019) to a means of exclusion and interdiction (Moreno-Lax 2015, p. 667).

Using content analysis of the term ‘safe country’ and its synonyms in the UK asylum legal provisions (Salehijam 2018), the second part of this paper (Section 3) demonstrates that the UK had been interpreting and applying the terms ‘safe country’ and ‘safe third country’ in its legislation as synonymous of first safe country to deflect its responsibility and push people seeking asylum back to other EU countries. Between 2022–2024, the UK Government rushed in Parliament a few statutes to solve the alleged migrant crisis affecting the UK shores and to guarantee safe legal routes (Home Office 2024). However, the Nationality and Borders Act (NABA) 2022 (UK Parliament 2022) and the Illegal Migration Act (IMA) 2023 (UK Parliament 2023) expanded the concept of inadmissible asylum applications. They also shifted the meaning of ‘safe country’ and pushed it beyond the current commonly agreed interpretation of the term, which requires a link between the person seeking asylum and the ‘safe country’. Thus, the UK legislation deviated even further from the rationale underlying the Refugee Convention and international human rights standards because it passed the obligation to another state without adequate risk assessment (John-Hopkins 2009, p. 237).

Finally, the last part of this paper (Section 4) analyses the legal battle between the UK Government, which enforced its programmatic plan through the Westminster Parliament, the Supreme Court and the Belfast High Court (BHC). While the UK government attempted to remove people seeking asylum in UK to a third safe country, like Rwanda, the Supreme Court and the BHC halted its plan. Specifically, the BHC concluded that the IMA violated EU legislation and, therefore, Article 2 of the Windsor Framework. Following this, the Court disapplied those provisions in Northern Ireland and made impossible for the UK Government to send people seeking asylum to a third safe country. States generally shift fictional borders around refugees to exclude them from protection (Arbel 2013; Osso 2023; Reynolds 2020) and use the term ‘safe country’ as a bordering tool (Osso 2023). However, this article concludes that there is some irony in the fact ‘a border’ between the Republic of Ireland and Northern Ireland is protecting the very same people that it is intended to exclude. There is additionally irony in the fact that the EU initially contributed to the erosion of their rights through the creation, interpretation and application of the ‘safe country’ concept. Yet, the EU has indirectly safeguarded people seeking asylum in the UK.

2. Safe Country: Origins and Evolution of the Concept

Before analysing how the UK government has been using the concept of safe country, this article analyses the changes in this term over the years as a product of various legal provisions and international policies and practices (Osso 2023). Although contentious (Coleman 2009), the text of the Refugee Convention (Foster 2007, p. 223) does not expressly prohibit the use of safe countries (Goodwin-Gill and McAdam 2010, pp. 88–91; Morgades-Gil 2020). Similarly, no restrictions can be identified in its drafting history (Goodwin-Gill and McAdam 2010, pp. 88–91). This missing reference is generally used by states as a loophole to send people seeking asylum to different countries. While the UN High Commissioner for Refugees (UNHCR 2002) has clarified that labelling countries of origin as ‘safe’ should be only permitted to share the burden of asylum claims, other countries prefer to completely shift this burden to third countries (Refword 1992; Noll 1997, 2000).

The following sections will demonstrate that when the concept of ‘safe country’ was created, it reflected a migrant-centric approach (Gammeltoft-Hansen 2011, p. 133). However, its widespread use around the world and in Europe changed its core meaning towards a more State-centric interpretation aimed at limiting the autonomy of migrants (Osso 2023, p. 282). This comprehensive history of the term will build the foundation for the analysis on the UK asylum system in Section 3 of this paper.

2.1. Origins of the Concept: A Delicate Balance Between States’ Concerns and Refugees’ Personal Circumstances

As will be explained in this section, in its original phase, the concept of ‘safe country’ was designed to protect asylum seekers while also addressing states’ concerns about managing asylum claims. This section argues that it was aimed at ensuring that decisions were fair and considered the personal circumstances of asylum seekers, rather than being used solely as tools for exclusion.

Although never explicitly mentioned by the UNHCR’s Executive Committee, the foundations of the concept ‘safe country’ intended as first safe country can be found in the 1979 Conclusion No 15 (XXX) on ‘Refugees Without an Asylum Country’. This document created some non-binding guidelines for States to address situations involving a large-scale influx of asylum-seekers and identifying the country responsible for an asylum request. Conclusion No 15 was seeking to reconcile different interests at stake. On the one hand, it was aware of the interest of certain states to deny asylum to asylum seekers and their intention to send them to a third state and, on the other hand, it valued the right of each asylum seeker to apply for asylum in the state where they wish to do so.

Although states were left to determine the grounds on which they could return asylum seekers to another country to avoid any possible disagreement among states, the non-binding directions included in Conclusion No 15 were designed to protect asylum seekers. Indeed, they indicated that states should have taken into account the duration and nature of any travel of the asylum-seeker in other countries, the intentions of the asylum-seeker as regards the country in which they wish to request asylum, any connection or close links with another State only if it appears fair and reasonable (although asylum should not be refused solely on the ground that it could be sought from another State). Therefore, those principles were supportive of the personal circumstances of the people seeking asylum and not thought as a tool of exclusion.

Similarly, a few years after, in 1989, in EXCOM Conclusion 58 (XL) the UNHCR implicitly referred to the concept of ‘country of first asylum’ when suggested that individuals who had found protection in a particular country, should normally not move from that country in an irregular manner in order to find durable solutions elsewhere. In this document, the UNHCR was addressing the growing concern of irregular migrants who had already found protection in a different state but who were not satisfied with their educational and employment prospects. Thus, they were illegally moving to another country to seek more favourable living conditions rather than to escape persecution.

This document did not infringe the spirit of the Refugee Convention because the latter, read in the light of its object and purpose, avails the use of first safe country in certain specific situations. For instance, the Refugee Convention includes some scenarios in which individuals lose entitlement to protection. One of these options is contained in Article 1E of the Refugee Convention, which excludes the possibility to obtain the refugees’ status if the asylum applicants had already been recognised as refugees in the first safe country (Moreno-Lax 2015, pp. 670–71). However, the meaning of the concept safe country was further stretched in Europe.

2.2. Origin of the Term in Europe: A Compromise to ‘Cure’ Refugees in Orbit

In Europe, the concept of safe country was developed as country of first asylum in 1980s to help ‘refugees in orbit’ (Melander 1978; Morgades-Gil 2020, p. 87; Moreno-Lax 2015, p. 666; Freier et al. 2021; Costello 2005, p. 40), indicating those individuals, who were seeking asylum, but who were unable to apply for it or were denied it. Thus, they kept moving from a country to another to seek protection (Moreno-Lax 2015, p. 664). More specifically, Denmark was the first country to use it but then, its use spread quickly in Europe. By the end of 90s every Western country has a similar clause (Byrne et al. 2004, p. 360; Lassen and Hughes 1997).

The problem of ‘refugees in orbit’ originated from the increase in the number of migrants arriving in Europe in the 1970s, where certain countries tried to restrict migrants’ access to their territory (Moreno-Lax 2015, p. 665). Countries started to adopt more stringent rules to block those individuals for different reasons. Some countries considered people seeking asylum to be economic migrants from unduly taking advantage of asylum procedures (de Vries 2007, p. 83; De Genova 2017, p. 7; Legomsky 2001, pp. 132–33). Other countries experienced substantial levels of family reunification (Melissourgos et al. 2023) or were significantly affected by the presence of large migrant communities within their territories (Van Liempt 2011). Additionally, some states, like Scandinavian countries, experienced a rise in immigration as a consequence of more restrictive policies implemented by neighbouring countries (Brekke 2004). The notion of ‘safe country’ was used as possible solution to ‘cure’ this gap considering that asylum seekers should apply for asylum in the first safe country they reach (Osso 2023, p. 279).

At that time, states did not share a concerted definition of the term but, as already clarified in the previous section of this article, its aim seemed to fall within the allowed scenarios of the Refugee Convention. Additionally, this interpretation was not necessarily having negative repercussions on people seeking asylum, but it was intended as a solution of compromise. It could have still been understood as both a rule of admissibility and an exclusion clause (Moreno-Lax 2015, p. 667). However, its adoption within the EU legislation opened to its use as an interdiction tool (Moreno-Lax 2015, p. 667), a bordering tool to exclude people seeking asylum from the European protection, as will be explained in the next sections.

2.3. Origin of the Term Within the EU: When Mutual Trust and Solidarity Are Not Enough

The EU legislation introduced the concept as ‘first safe country’ in the early 1990s in two legally binding treaties, the Schengen Implementation Agreement and the 1990 Dublin Convention (Abell 1997, p. 571; Lavenex 2006, p. 334) under the pressure of an increased influx of people seeking asylum following the end of the cold war, the fall of the communist regimes and the Balkans war. Both instruments introduced a negative mutual recognition duty (Guild 2001, p. 206), according to which a state is exempted from its responsibilities to protect people seeking asylum if any of the EU countries could have provided a positive asylum determination (Achermann and Gattiker 1995; Hurwitz 1999; Marx 2001). This policy was based on the assumption that all EU countries were safe and had an equivalent standard of protection (Morgades-Gil 2020, p. 83). It required the explicit consent of the receiving state and it was inspired by the spirit of solidarity and burden-sharing among the signatory parties of the Refugee Convention and the EU (Moreno-Lax 2015, p. 672).

In 1992, the London Resolution helped incorporating of this concept into national legislation before Dublin Convention entered into force. The latter, although a non-binding instrument, established some common criteria for rejecting manifestly unfounded asylum claims and returning individuals who were not in genuine need of protection. The draft resolution of the latter document is one of the first documents, which reveals a cynical approach to the problem of asylum shopping by stating that certain individuals “come from countries in which levels of security, economic; opportunity or individual liberty are below those of the Member states”. Thus, member states could declare as manifestly unfounded the applications from individuals who could reasonably be expected to seek protection in a “first host country” or even from applicants who have not been in the safe country “since leaving the country in which [they] claim[…] to fear persecution but where [they] would clearly be admissible”. This document left to the discretion of each state the decision of whether to apply the safe country doctrine or not. Although the latter was conceived as an option or a mere exception to the obligations enshrined in the Refugee Convention, this concept soon became the rule in all the EU countries (Achermann and Gattiker 1995, p. 23).

While the EU was lifting the borders for its citizens within the EU member states between the 80s and 90s (Moreno-Lax 2015, p. 675), it was simultaneously creating new borders for people seeking asylum (Osso 2023, p. 285; Abell 1997; Frelick et al. 2016, p. 191; Joly 1994, p. 159). This was the beginning of a new trend. The EU started using the umbrella term of ‘safe country’ as a bordering tool. The rationale behind this seems to be the belief that people seeking asylum were not in genuine need for protection but, with their secondary movements, they wanted to improve their own economic status with a better job, education or healthcare (Edwards 2005, p. 304; Legomsky 2001, p. 597; Foster 2007, p. 232). Probably any characterisation of migration flows could prove itself inaccurate because they might include economic migrants, irregular migration, people seeking asylum and each person might travel for more than one reasons (Mattila 2000). In any case, states deemed that those individuals could be returned to the state(s) where they came from. This produced three results: one on people seeking asylum, one on EU states and the last one on the concept of safe country. First, it created a new category of people seeking asylum caught in a legal vacuum (Melander 1986, p. 221; Moreno-Lax 2015, p. 666; Freier et al. 2021, p. 518), unable to apply for asylum and pushed from a country to another like in sort of domino effect, an infinite chain of refoulement, or neo-refoulement, if we use the words of Hyndman and Mountz (Hyndman and Mountz 2008, p. 249; Caron 2017, p. 29). Second, countries shaped their migration policy with the intent to send asylum seekers back from where they came from (Lambert 2012, p. 318) and to erode refugee rights and international protection standards (Moreno-Lax 2015, p. 720). This, indeed, reflected the end of the burden-sharing approach in favour of a burden-shifting legal paradigm (Moreno-Lax 2015, p. 675; Hathaway and Neve 1997, p. 117). Finally, the concept of first safe country evolved into the notion of ‘safe third country’ even outside the EU, where the refugee is believed to have a prior connection. Thus, from an ad intra perspective (Morgades-Gil 2020, pp. 83–84), the term acquired an external dimension, which was developed by the following EU legislation, including the Dublin II Regulation (343/2003), the 2005 Asylum Procedures Directive (2005/85/EC), the recast Asylum Procedures Directive (APD) and the recast Dublin III Regulation, which will be analysed in the next section.

2.4. Codification and Harmonisation Within the EU: From Burden-Sharing to Burden-Shifting

In the last 20 years, the EU asylum law changed its approach towards people seeking asylum. The burden-shifting approach of the EU countries towards people seeking asylum become more evident through broader mechanisms, including carrier sanctions, which significantly restricted the ability of asylum seekers to travel directly to countries like the UK and thereby increased the relevance of “safe third country” doctrines (Directive 2001/51/EC supplementing Article 26 of the Convention implementing the Schengen Agreement).

This section demonstrates that the legitimacy of the concept ‘safe country’ was not questioned anymore but the EU was more concerned about creating workable legal definitions that would safeguard EU’s interests. The EU broadened states’ discretion while promoting little guarantees for people seeking asylum who, at least, needed a connection with the receiving safe state.

While the 2003 Dublin II Regulation assumed that all member states should be considered as safe countries for third-country nationals, the EU codified “safe third country” and “safe country of origin” as two distinct concepts in the 2005 Asylum Procedures Directive (2005/85/EC). More specifically, Article 25 confirmed the inadmissibility of certain asylum claims if another member state has already granted asylum or another non-EU country can be considered either a first safe country or a safe third country. The fact that a State declared an application ‘inadmissible’ changes the status of the applicants from person seeking asylum to irregular migrants (Coleman 2009, p. 308).

The interest of safeguarding states’ interest to the detriment of people seeking asylum is evident in the EU provisions, which codify the concepts safe third country” and “safe country of origin” as will explained below. Article 26 establishes that applicants can consider a state as first safe country if they have been recognised in that country as refugees and they can still avail themselves of that protection or, alternatively, if they enjoy sufficient protection including non-refoulement. Under both circumstances, the applicants must be re-admitted to that country. Thus, the readmission in the first safe country requires the consent of the receiving State but it is not necessary to receive the consent of the asylum applicant. The state-centric approach is clear.

With regard to safe third country, Article 27(1) reads that members states may apply the safe third country concept when the competent authorities are satisfied that a person seeking asylum will enjoy in that country their right to life and liberty without any discrimination, the principle of non-refoulement, prohibition of removal and the possibility to request refugee status. The second section of this Article leaves wide discretion to the national authorities to apply this concept in accordance with their national legislation, but it establishes that a connection between the person seeking asylum and the safe third country concerned must exist. At that stage, the EU countries applied this approach in a quasi-automatic way (Morgades-Gil 2020, p. 85) because they had a list of countries designed as safe.

Those criteria are not tailored on individual situations. Thus, in addition to the criteria established either in Article 29 (minimum common list of third countries regarded as safe countries of origin) or Article 30 (national designation of third countries as safe countries of origin), Article 31 provides a residual definition of the safe country of origin concept tailored to the individual applicant. It reads that states can consider it as safe only either if the applicant has the nationality of that country or the individual is a stateless person and was formerly habitually resident in that country. However, in both cases, the applicant must not have submitted any serious grounds for considering the country not to be a safe country of origin.

This seems to be the tilting point which disregards refugees’ rights in favour of a more state-centric approach. It is also evident that the term ‘safe third country’ has become a border control mechanism. Both states’ wide margin of appreciation to identify a safe third country with loose parameters and their ability to unilaterally declare some states as ‘safe’ operate as a perfect tool to consecutively delegate responsibility from a state to another. The fact that the concept of ‘safe third country’ grounds itself on the idea that all the countries of Common European Asylum System (CEAS) (28 EU Member States before Brexit and four associate States—Norway, Liechtenstein, Iceland and Switzerland) can be considered safe because they offer equivalent standards of protection (Pascouau 2012). This equivalence is based on the principles of mutual trust and solidarity, cornerstones of the EU. In this context, the principle of solidarity indicates State’s ability adopt common rules to share certain duties, while the principle of mutual trust provide certainty that States will apply those rules effectively (Pascouau 2012). However, this is not a sufficient guarantee. The NS and ME case showed that this assumption is false especially when there are substantial grounds for believing that there are systemic flaws in the asylum procedure and reception conditions for asylum applicants in certain member state (N. S. (C-411/10) v Secretary of State for the Home Department and M. E. and Others (C-493/10) v Refugee Applications Commissioner and Minister for Justice, Equality and Law Reform).

Additionally, with the codification of the concept of a ‘safe third country’, this regulation added a further external dimension to the immigration and asylum policies of the EU, which discriminates against people seeking asylum. The EU projected the principle of mutual trust into transferring individuals to safe third countries (Morgades-Gil 2020, p. 92). Although the intent was to eliminate irregular migration, the application of the safe third country notion penalises the unauthorised entry of any individual who is not coming directly to any of the EU states (Osso 2023, pp. 285–86). In this way, the EU created a new category of inadmissible subjects (Osso 2023, p. 286). This system pushes people seeking asylum beyond the EU borders or, at best, to the Greek hotspots (Moreno-Lax 2015, p. 678; Coleman 2009, pp. 265–66; Triandafyllidou and Dimitriadi 2013, pp. 601–3). Countries believe that this is a necessary evil to reduce the pressure on EU countries (Marx 2019, p. 585) but this article contends that the EU created categories of inadmissible subjects to control the flow of immigration and to protect its EU internal markets and social welfare systems by restricting the entry of individuals who might become a financial burden. This mechanism does not entail sufficient safeguards for people seeking asylum who do not benefit from the principle of solidarity and from the support they might need.

For this reason, in 2013, the Asylum Procedures Directive (APD), which codified the NS jurisprudence, provided a more comprehensive definition of safe country of origin in Article 36 and safe third country in Article 38, while adding the concept of European safe third country in Article 39. It also enhances states’ discretion when Article 38 of the APD reads that Member States are not obliged to apply the concept of sae third country but they “may apply” it when the competent national authorities are satisfied that an individual seeking asylum will be protected according to certain standards. The agreed minimum threshold included the protection of life and liberty because of discrimination; safeguard against death penalty, torture or inhuman or degrading treatments; full respect of the principle of non-refoulement. Additionally, Article 38 includes some additional guarantees for people seeking asylum because it still requires a connection between the applicant and the third country.

At the time of Brexit, the relevant legislation for the UK was contained in the so-called 2013 Dublin III did not question the concept of safe country anymore. While the definition of first safe country and safe third countries remained unchanged, the Dublin III Regulation streamlined the rules establishing the responsibility for examining asylum applications. The EU had recognised the challenges that certain States faced in managing migratory flows and Dublin III asks people claiming asylum to apply in the first EU Member State they entered. Although the EU is built around the principle of solidarity and sharing responsibility around checks at the borders, Dublin III was never driven by principle of solidarity among states and fair sharing of responsibility to support countries that face situations of migratory pressure.

The conclusion on this mechanism is; however, that Dublin III does not extend the solidarity principle to people seeking asylum in need of international protection through these legal manoeuvres (Morano-Foadi 2017, pp. 223–54; Garlick 2016, p. 166). Conversely, the EU has endorsed a de-solidarity principle (Mann 2016, p. 3), rooted in the view that expressing solidarity to refugees and migrants is self-destructive for the same union (Okafor et al. 2025). In conclusion, the concept of ‘safe third country’ had assumed its full power as a migration management tool, used by the EU to exclude people seeking asylum from the protection afforded by the 1951 Refugee Convention.

2.5. Outside the EU

The concept of safe country has proliferated in the form of both first safe country and safe third country beyond the European borders as well. An example of codification of the former notion (first safe country) is contained in the 2016 Safe Third Country Agreement (STCA) between Canada and the United States. It applies to people who are seeking asylum entry to Canada from the US. The STCA is based on the notion of first safe country because it requires that individuals seeking asylum should apply for protection in the country of their first arrival if determined to be safe. The US is the only country that is designated as a safe third country by Canada under the Immigration and Refugee Protection Act. Thus, on that basis, the Canada border officials could immediately return people seeking asylum to the US and vice versa. A similar policy has been used in relation to migration flows from Mexico to the US under the first Trump’s administration (Andreas 2022).

Before Brexit, the evolution of the latter concept (safe third country) has been more controversial as will be explained in the remainder of this section. This notion was, instead, used by some central and eastern European states, together with some bilateral readmission agreements (Petersen 2002, p. 353) and by other countries like Australia and Israel (Bar-Tuvia 2018). These latter agreements are characterised by a crucial difference with the Dublin approach, which is the absence of the requirement of a link between the people seeking asylum and the country where they are going to be transferred. Thus, individuals can be transferred to a fourth state (Office of the High Commissioner of Human Rights 2025).

In 2005, the UNHCR published some recommendations which clarified that the ‘fair and reasonable’ link between the people seeking asylum and the state by stating that “requiring a connection […] is not mandatory under international law”. However, this interpretation seems to substantially depart from the obligations set in the Refugee Convention. It is true that the Refugee Convention does provide any safeguard against safe third country because it does not expressly ban states to return people seeking asylum to a third state where they could have requested protection (Foster 2007, p. 244; Moreno-Lax 2015). This rule is matched by the idea that the right to asylum does not exist (Marx 2019, p. 583. For a different view, see, Hathaway 2005, pp. 740–42) or, at least, not in the country where asylum seekers choose to apply (Legomsky 2001, pp. 612–13). Thus, it availed the idea that people seeking asylum could be returned to a country that they briefly passed en route (Office of the High Commissioner of Human Rights 2005). However, the transfer to a completely new country seems a considerable departure from the obligations enshrined in the Refugee Convention. This Post-Dublin regime which passes the obligation to assess asylum claims to a fourth country, which is in fact a special category of ‘safe third country’ to which the applicants have no link, undermines the spirit and the letter of the Refugee Convention.

In this legal and political scenario, the UK created its own asylum legislation.

3. UK Asylum Law and the Concept of ‘Safe Country’

As already mentioned above, States soon began using the concept of ‘safe country’ to support exclusionary policies towards people seeking asylum. The UK has been following this trend as well. The following sections will demonstrate that, before Brexit the UK politics did not distinguish between ‘first safe country’ and ‘safe third country’ because the UK’s intention was to always return those individuals to EU countries. This approach was supported by the geographically close vicinity with France and Germany, EU states through which people seeking asylum in the UK were traveling.

After Brexit, while the EU maintained its commitment to cohesion and responsiveness to ensure effective refugee and human rights protections within its borders (Legomsky 2001, p. 606; John-Hopkins 2009, p. 254), the UK was no longer bound by EU law and policies. Thus, it departed further from these principles. Not only did it finally include the concept of ‘safe third state’ in its legislation, effectively linking to the non-EU states, but it also eroded the solid criterion that the individual seeking asylum should have a connection to the safe country. Hence, it used the concept of ‘safe third country’ as a border regulation control mechanism, while also aligning itself with other non-EU countries that shift their responsibilities to ‘fourth countries’, a special category of ‘safe third country’ where the refugees had never transited.

Table 1. Content analysis of the term safe country in the UK asylum law.

Statute	Terminology in the Statute	Meaning	EU Country	Non-EU Countries	Connection Between the Individual and the Safe Third Country
S 71 Immigration and Asylum Act 1999	Safe third country	First safe country	YES	Not clear	YES
S 80 Nationality, Immigration and Asylum Act 2002	Third country	First safe country	YES	Not clear	YES
S 93 Nationality, Immigration and Asylum Act 2002 (S. 93 repealed (1.10.2004) by Asylum and Immigration (Treatment of Claimants, etc.) Act 2004 (c. 19), ss. 33(3)(b), 47, 48(1)–(3), Sch. 4; S.I. 2004/2523, art. 2, Sch.	Third country	First safe country + Safe third country	YES	YES	NO
S 94 Nationality and Borders Act 2022	Safe Countries of Origin	First safe country	YES	N/A	NO	S 94 A
S 33 Asylum and Immigration (Treatment of Claimants, etc.) Act 2004	Safe country	First safe country	YES	Not clear	YES
Schedule 3 Asylum and Immigration (Treatment of Claimants, etc.) Act 2004	Safe country Safe third country	First safe country	YES	Not clear	YES
S 80 Nationality and Borders Act 2022	Third country	Not clear	YES	Not clear	NO
S 80A. Nationality, Immigration and Asylum Act 2002 (as modified by Nationality and Borders Act 2022)	Safe States	First safe country	YES	NO	YES
S 80AA. Nationality, Immigration and Asylum Act 2002 (as modified by Nationality and Borders Act 2022)	Safe States	First safe country	YES	YES	Not clear
S 80B. Nationality, Immigration and Asylum Act 2002 (as modified by Nationality and Borders Act 2022)	Safe third country	First safe country	YES	YES	Not clear
S 80C. Nationality, Immigration and Asylum Act 2002 (as modified by Nationality and Borders Act 2022)	Safe third state	First safe country + Safe third country	YES	YES	NO
Schedule 1 Illegal Migration Act 2023	Countries or territories to which a person may be removed	First safe country + Safe third country	YES	YES	NO

presents a synopsis of the content analysis on the term ‘safe country’, ‘first safe country’ and ‘safe third country’ in the UK asylum statutes since the 90s, when those concepts emerged in Europe. The first column includes the relevant UK provisions, while the second clarifies terminology used by the Parliament. As anticipated in the introduction of this article and further explained in the following sections, the UK legislator ambiguously used these terms and created a legal fiction masterminded to serve its policy of internal cohesion against the external threat coming from refugees. For this reason, the third, fourth and fifth columns aim to explain, respectively, the effective meaning behind those terms and whether this interpretation coincides with EU or non-EU countries. The data contained in these columns are not always conclusive as they entail a wide degree of discretion by the UK authorities, given the close vicinity of the EU shores. The sixth column includes a reference to the requirement of a connection between the individual seeking asylum and the safe third country. It shows that this link was eroded by the post-Brexit legislation. The following sections will provide a more detailed explanation of this chart.

3.1. Safe Country Before Brexit

The political discourse on asylum and immigration in the UK has often been linked to national security issues. The UK immigration rules became more restrictive, complex and punitive in the wake of the 9/11 terrorist attacks in the US and the 7/7 London bombing (Saenz Perez 2023, p. 304; Léonard and Kaunert 2022; Jaskulowski 2018; Karamanidou 2015). While the intent to keep possible asylum claimants outside the UK borders was clear, the provisions on the concept of safe country, intended as both first safe country and safe third country, has been rather confusing. This section discusses the results of the content analysis of the provisions before Brexit contained in Table 1 and it demonstrates the UK legislator’s strategic use of conceptual ambiguity surrounding the concept of ‘safe country’. It will show that this approach secured outcomes favourable to the UK but it will starkly highlight that people seeking asylum are left vulnerable and unsupported.

S71 of the Immigration and Asylum Act 1999, entitled ‘Removal of asylum claimants to safe third countries’, introduced the concept of safe country into the UK legislation. Although the literal interpretation of this provision would suggest a reference to the concept of safe third country, the concept was intended as first safe country or as synonyms of any European country. This is clear from the Hansard when Mr Howard, commenting on the Bill, said that

“There is a well-recognised principle that those in danger of persecution apply for asylum in the first safe country in which they arrive. It is generally accepted internationally. The French and the Germans accept it and so do we. The difficulty is that we take so long to deal with such cases that the French and the Germans will not accept the people back. That is why we have to streamline our procedures”.

(Hansard 1995)

Additionally, his words demonstrate the false premises under which this concept was adopted. It seems that the MP believed that safe country was a generally accepted principle under international law when, at the time, it was a relatively new concept included within the EU legislation. Thus, he failed to acknowledge that the UK should have included this expression in its legislation because of the harmonising role of the EU legislation.

Another example of how the UK mixed the use of the concepts of ‘safe third country’ and ‘first country of asylum’ with a clear overlapping with EU countries is contained in the subsequent statute, the Nationality, Immigration and Asylum Act 2002. S 80, entitled ‘Removal of asylum-seekers to third country’ regulated the removal of a person in relation to whom a certificate has been issued. The provision contains some requirements to identify a safe country, including protection against direct and indirect refoulement. The literal interpretation of this provision and the legal requirements seemed to corroborate the view that the UK was incorporating the EU concept of ‘safe third country’ into its legislation. However, the text of the provision refers to the removal from the UK to ‘a member state’ in two different points. This seems, again, a clear reference to EU countries. However, additional confusion to the interpretation of this provision is added by the explanatory notes, which clarify that “[t]he definition of standing arrangements is amended to ensure that any bilateral agreements on asylum returns with Member States outside of the Dublin Convention to which section 11 already applied, also will fall within this provision”. Thus, in theory this provision seems to expand this concept to also include non-EU countries. However, in practice, the definition only includes EU countries. This view is further supported by the subsequent statutes and changes.

Indeed, this provision was repealed by the Asylum and Immigration (Treatment of Claimants, etc.) Act 2004, which also replaced the provisions contained in sections 11 and 12 of the Immigration and Asylum Act 1999. This statute introduces s 33 and Schedule 3, which contain a whitelist of safe countries where people seeking asylum can be transferred without substantive consideration of their asylum claim (UK Parliament 2004). Incidentally, the list only includes countries within the CEAS area. Therefore, once again, the UK legislation only referred to European countries. Despite this, the interpretation of this provision is further muddled by Part 5 of Schedule 3, which establishes that the Secretary of State may certify a non-listed country as safe for a particular individual under the Refugee Convention and the European Convention on Human Rights (ECHR). Therefore, it theoretically expands this definition to non-EU countries, but it is not clear whether this would have infringed immigration rules as set by EU standards.

3.2. Safe Country After Brexit

Since Brexit, the tone of the UK politics on immigration has become harsher (Griffiths and Yeo 2021). The UK government promised its electorate to curb illegal immigration, secure smugglers to justice and prevent individuals from embarking in dangerous journey (Saenz Perez 2023, p. 307; Jiang and Erez 2017). The language of safe legal journeys became prominent, while paradoxically the UK Government did not establish any safe routes for people seeking asylum (Mayblin 2021). Conversely, the lack of safe routes pushed individuals seeking asylum to undertake illegal and dangerous journeys to the UK. Thus, between 2022–2024, the UK Government blatantly disregarded the very same migrants’ safety (Griffiths and Trebilcock 2022). Finally, although the UK government stressed the importance of taking back control of their own British borders, it conveniently failed to distinguish between EU migration, international movements and people seeking asylum.

The spiralling of securitisation of the UK asylum law (Bello 2017; Valverde 2011, p. 4) also passed thought the ambiguity of concepts like ‘safe country’, ‘first safe country’ and ‘third first country’. Once the UK government realised that they could not return people seeking asylum to EU countries (Women and Equalities Committee 2022), in 2022 the UK Government expanded the concept of inadmissible asylum claims through s 16 of the NABA 2022, which amends Part 4A of the Nationality, Immigration and Asylum Act 2002. Sections 15 and 16 of NABA create a two-tier asylum system and established that the UK could remove people seeking asylum without issuing any certificate considering that S16(1) gave the Secretary of State for the Home Department (SSHD) the discretionary power to “declare an asylum claim made by a person […] who has a connection to a safe third state inadmissible”. The two-tier system distinguishes people seeking asylum on the basis of their arrival in the UK. If they had arrived ‘directly’ from countries where their life or freedom may be at risk, they will be included in Category 1. Conversely, if they had transited through a ‘safe third countries’ before reaching the UK, they belong to Category 2. The latter category of people could be transferred to a ‘third safe country’. As noted by Powell and Rifath, although the NABA confers to the SSHD a discretionary power rather than an obligation, given the present political context, it is highly likely that this discretionary power will be widely used (Powell and Rifath 2023, p. 759). Additionally, this blanket approach to safe country raised the threshold of evidence for those individuals who want to challenge the definition of ‘safe country’ (Powell and Rifath 2023, p. 760). These aspects, together with the expansion of the inadmissibility category under NABA, make the definition of ‘safe third country’ crucial.

S16 (in modifying s80B) establishes the criteria to define a country as safe. These are that the claimant’s life and liberty are not threatened in that State by reason of their race, religion, nationality, membership of a particular social group or political opinion; there is no risk of indirect refoulement in violation of the Refugee Convention and Article 3 of the ECHR and customary international law, and that person may apply to be recognised as a refugee in that State. This provision includes a fictional definition of safe third country because it seems to be built around a close connection with the EU countries. Indeed, it is likely that asylum applicants would cross the Channel to reach the UK shores and would depart from one of states geographically close to the UK. These are all countries within the CEAS and, therefore, deemed as safe. It seems that, once again, the UK Government used the term ‘safe third country’ as synonym of EU country and, more importantly, as first safe country. This would consequently lead to the inclusion of people seeking asylum in Category 2 and, thus, make the majority of the asylum applications in the UK potentially inadmissible (Gullis 2021).

The negative consequences on people seeking asylum of this legal fiction are evident from the additional modification of s.80C of the Nationality, Immigration and Asylum Act 2002 by s.16 of the NABA 2022 on meaning of connection to a safe third States. This change makes clear the confusing application of the two different concepts and, more importantly, defines ‘connection’ between the person seeking asylum and the receiving safe state in very broad terms. This effectively erodes the last assurances for individuals seeking asylum, as provided by the EU legislation in the Dublin III system.

The provision identifies five different conditions under which the claimant can be recognised as having a link to a safe third country. The first four conditions seem to indicate circumstances very similar to the definition of first safe country because it mentions that the individuals have or could have applied for asylum or a different type of protection. The assumption under those circumstances is that the individual seeking asylum is present on the territory of the relevant country to make the application. Therefore, the two concepts seem to blur into one another. These conditions are: 1. that the claimant has been recognised as a refugee in the safe third State and remains able to access protection in accordance with the Refugee Convention in that State; 2. that the claimant has otherwise been granted protection in a safe third State as a result of which the claimant would not be subject to any indirect refoulement; 3. that the claimant has made a relevant claim to the safe third State and the claim has not yet been determined, or has been refused and 4. that the claimant was previously present in, and eligible to make a relevant claim to, the safe third State, it would have been reasonable to expect them to make such a claim, and they failed to do so.

However, the fifth condition departs from this assumption and mentions “that, in the claimant’s particular circumstances, it would have been reasonable to expect them to have made a relevant claim to the safe third State (instead of making a claim in the United Kingdom)”. This last condition expands further the concept of connection to the point that it seems that no nexus at all is needed. If the UK was still part of the EU, the latter legislation would have required a nexus, according to the Dublin system. However, with Brexit, the UK can interpret this term the way they want. More in general, neither scholarship nor case law acknowledge that a State through which the asylum seeker has not travelled or to which he or she has no personal or family links could be considered a ‘first country of asylum’ or ‘safe third country’ (Marx 2019, p. 591). The UNHCR endorses a similar position when declares that people seeking asylum should not be returned to third countries with which they lack sufficient connection.

In 2023, Parliament passed the Illegal Migration Act (IMA), which facilitated the enforcement of stricter policies and language directed at migrants. The IMA does not mention the term safe third country in its provisions, but it assumes that the UK should avail itself of EU countries and first safe country, as will be explained below. S2 places a duty upon the Secretary of State to make any arrangements for the removal of individuals who unlawfully entered the United Kingdom. More specifically, s2(4) dictates that one of the exclusionary conditions is “that, in entering or arriving as mentioned in subsection (2), the person did not come directly to the United Kingdom from a country in which the person’s life and liberty were threatened by reason of their race, religion, nationality, membership of a particular social group or political opinion”. Thus, it establishes that people to legally seek asylum in the UK should come directly from a country where their life and liberty are endangered. Once again, it seems likely that asylum applicants will reach the UK by traveling from one of the EU bordering countries, which are included in the whitelist of first safe countries. Thus, the 2023 IMA seems to make the majority of people seeking asylum coming to the UK as illegal migrants. This ‘belt and braces approach’, according to which the UK has created multiple overlapping legal or procedural rules to ensure that a government’s position on asylum is upheld, reflects a deliberate strategy of the UK Government to insulate controversial policies from legal challenge and expedite removals, even at the risk of undermining international refugee protections (Salter and Mutlu 2011; Clayton 2014; Crawley 2001).

The view that the UK Government was confusingly applying two different concepts of ‘first safe country’ and ‘safe third country’ is further corroborated by the list contained in Schedule 1 of the 2003 IMA. Until the beginning of 2024, the list only included CEAS counties. However, there have been some late additions of non-EU countries, such as Georgia, India and Albania. The UK asylum approach adopted after Brexit presumes a strong cooperation with EU members states and it is dependent on the full respect of EU law and ECHR (Mitsilegas and Guild 2024). While the UK always had a strong cooperation history with France, Brexit put the UK in a weaker position to return people seeking asylum to the EU countries. Indeed, at the time, the UK did not have other workable agreements with any states outside the EU.

4. The Legal Battle Between the UK Parliament, UK Government and the Supreme Court

The last part of this paper focuses on the legal battle between the UK Government, which enforced the concept of ‘safe third country’, even in cases where people seeking asylum have no connection to that county, the UK Parliament, the UK judiciaries in the AAA case the Belfast High Court. On one hand, the UK government tried to refoul people seeking asylum in UK to Rwanda; on the other hand, the UK Supreme Court and the Belfast High Court halted its plan.

4.1. The Rwanda Policy

On 14 April 2022, the UK decided to sign a Memorandum of Understanding (MoU) for the provision of an asylum partnership arrangement with Rwanda. Following to that, on 18 March 2023, they signed Addendum to the MoU and, on 28 November 2022, some Notes Verbales on the topic complemented the previous documents. Those documents constitute the so called ‘Rwanda Policy’, which does not contain any reference to the concept of ‘safe country’ but it is fashioned under this umbrella term. The MoU is not a formal treaty. It is a non-binding political agreement and, as such, it was not subject to sufficient parliamentary scrutiny under the Constitutional Reform and Governance Act 2010 (UK Parliament 2010a). Likewise, this Rwanda policy presented significant limitations under the Equality Impact Assessment, which led some scholars to think that it would constitute a violation of the rights of those with protected characteristics under the Equality Act 2010 (UK Parliament 2010b).

More specifically, Section 2.1 of the MoU clarifies that this includes “a mechanism for the relocation of people seeking asylum whose claims are not being considered by the United Kingdom, to Rwanda, which will process their claims and settle or remove (as appropriate) individuals after their claim is decided, in accordance with Rwanda domestic law, the Refugee Convention, current international standards, including in accordance with international human rights law and including the assurances given under this Arrangement” (Home Office 2022). The UK is responsible for carrying out the initial screening of people seeking asylum (Section 5.1), providing people seeking asylum’s details to Rwanda (Section 5.2), organising safe transport to Rwanda (6.1) and sharing the relevant documents with Rwanda (Section 6.3). On the other hand, Rwanda is responsible for providing access to people seeking asylum to Rwanda’s territory (Section 7.1), checking their documents (Section 7.2), providing accommodation (Section 8.1), assessing asylum claims (Section 9.1), granting them refugee status (Section 10.1), removing them from the country in case of rejection (Sections 10.2–10.4), granting rights to people seeking asylum (Section 10.5), taking the necessary steps to ensure that the needs of modern slavery victims are accommodated (Section 14) and relocating vulnerable people seeking asylum back to the UK (Section 16) (Home Office 2022). Thus, the Rwanda Policy establishes little responsibility for the UK, while Rwanda is responsible for assessing asylum claims and for providing access to people seeking asylum to Rwanda’s territory. Additionally, the Addendum to the MoU, signed on 18 March 2023, clarifies the rationale behind the MoU. The latter aims to deter dangerous and illegal journeys which are putting people’s lives at risk; disrupt the business model of people smugglers who are exploiting vulnerable people; provide an option for people who desire asylum or protection to make such a claim in Rwanda or otherwise make another type of immigration application in accordance with Rwandan domestic law, the Refugee Convention, current international standards, including in accordance with international human rights law.

Considering the absence of parliamentary scrutiny and the unclear interpretation of the UK’s human rights obligations, there was no doubts that these issues would have been ended up being examined by the UK courts. Eventually, this was the case.

4.2. The Legal Battle Before the UK Judges in the AAA Case

A heated legal battle followed the release of the MoU. Ten people seeking asylum coming from Syria, Iraq, Iran, Vietnam, Sudan and Albania, who reached the UK by crossing the English Channel from France in small boats, challenged the Rwanda policy through three tiers of the UK judicial system. The decisions looked at whether Rwanda could be considered ‘safe’ so that the asylum claimants could be safely refouled there but the discussion on the legality of the core concept of safe third country with which people seeking asylum had no previous connections lacked breath and meaningful analysis.

On 19 December 2022, in the AAA case, the High Court found that the Secretary of State for the Home Department (SSHD) had lawfully decided that Rwanda is a safe third country. This decision took into consideration a “thorough examination of all relevant generally available information” at the time, a threshold as required by the European Court of Human Rights in Ilias and Ahmed v Hungary. More specifically the High Court judges held that the claimants’ requests were merely speculative because they are not based “on any evidence of any presently-held opinion” given that no removals had been enforced at that stage. After having determined that, the judges considered whether sending asylum claimants to a safe third country could be infringing Articles 31 and 33 of the Refugee Convention, respectively, on prohibition to impose penalties upon people seeking asylum and on their right of not being refouled. They concluded that given that Rwanda is a safe country, removing asylum claimants cannot be considered contrary to Article 33 (principle of non-refoulement) and, consequently, a removal that is not violating the principle of non-refoulement is not a penalty for the purposes of Article 31.

The claimants subsequently appealed the High Court’s decision. Upholding the appellants’ claim by a majority of 3 to 1, the Court of Appeal concluded that Rwanda is not a safe country because of the risk of indirect refoulment, which meant that there was a real risk that persons sent to Rwanda will be returned to their home countries where they faced persecution or other inhumane treatment. The risk of indirect refoulment would cause the Government to act in violation of Article 3 of the ECHR. This rationale is based on the obligation contained in Section 6 of the 1998 Human Rights Act, according to which public authorities must act consistently with the ECHR. Therefore, if the SSHD sends asylum claimants to Rwanda, they will violate the claimants’ human rights and, consequently, act in violation of the ECHR. Similarly to what previously held by the High Court, the Court of Appeal clarified that sending people seeking asylum to a third country cannot be considered a penalty and, thus, it is not in violation of Article 31. As Palmer argued, the Court of Appeal lost a crucial opportunity to offer a purposive reading of the Refugee Convention that upholds the UK’s own obligations to offer refuge to people fleeing war and persecution (Palmer 2023) because it did not comment on the legality of practice of removing individuals to a safe third country with which they have no connection. However, the Court of Appeal went a step further than the Hight Court by stating that the link between people seeking asylum and the safe third country would have been a requirement if the UK was had remained a member of the EU. This distinction is significant because it highlights a contextual interpretation of the provisions rooted in the principles of EU law and stresses the evolving legal standards of the UK asylum law system in the post-Brexit context.

Ultimately, the case was heard before the Supreme Court, which again addressed the question on whether Rwanda could be considered as a safe country. On 15 November 2023, the Supreme Court reached the conclusion that Rwanda was not a safe country because individuals removed to Rwanda face a real risk of ill treatment in violation of the principle of non-refoulement not only under Article 3 of the ECHR but also under customary law. Thus, the Supreme Court concluded that the lack of safety of Rwanda issue meant that the Rwanda policy must be declared unlawful.

4.3. The UK Government Strikes Back

A few weeks after its defeat before the Supreme Court, the UK Government decided to sign a treaty with Rwanda, supported by an act of the Parliament, the Safety of Rwanda (Asylum and Immigration) Act 2024. Section 2(1) of the Safety of Rwanda Act (SRA) established that “[e]very decision-maker must conclusively treat the Republic of Rwanda as a safe country”. Thus, it enabled the removal of persons to the Republic of Rwanda through the creation of an ad hoc legal fiction on the safety of Rwanda (Pulvirenti and Lalor 2024) when it imposes a statutory obligation upon the Home Secretary, immigration officers, courts, and tribunals to consider Rwanda as a safe country. The new legislation not only overrode the factual determination of the Supreme Court but also ringfenced the executive’s designation of Rwanda as a safe third country from judicial scrutiny by using the constitutional principle of parliamentary sovereignty. The SRA de facto removes the ability of the judiciary to scrutinise the safety of Rwanda. This is further reinforced by Section 3 of the SRA, which disapplies the key provisions of the 1998 HRA and limits the ability of individuals to challenge removals on human rights grounds, and by Section 5, which restricts the domestic effect of interim measures issued by the European Court of Human Rights, unless a minister decides to comply. By precluding national judges to assess the safety of Rwanda and limiting the application of the ECHR, the government eroded the sophisticated system of check and balances within the UK Constitution and undermined the ability of the UK to protect and fulfil its international obligations.

4.4. Back to Court—The Belfast High Court on the Northern Ireland Human Rights Commission’s Application and JR295’s Application and in the Matter of The Illegal Migration Act 2023

The last act of the legal battle between the UK Government acting through Parliament and the judicial scrutiny of the third safe country policy happened before the BHC, which disapplied the IMA 2023 in Northern Ireland. The BHC was called to determine whether the IMA 2023 was incompatible, first, with Article 2 of the Ireland/Northern Ireland Protocol or Windsor Framework, as implemented by section 7A of the European Union (Withdrawal) Act 2018, and second, with Articles 3, 4, 5, 6 and/or 8 of the ECHR. The key provision here is the Windsor Framework, which is a post-Brexit binding treaty between the UK and the EU. It guarantees that the EU law provisions would continue to have the same effect in Northern Ireland post-Brexit. Article 2 establishes that the UK must ensure that the people of Northern Ireland do not suffer a “diminution of rights, safeguards, or equality of opportunity”.

On 13 of May 2024, the BHC held that the duty to remove people seeking asylum to a different country was deemed inconsistent with EU law and, more specifically, articles 25–27 of the Procedures Directive. Those EU provisions require a connection with a safe third country to qualify an asylum application as admissible. Thus, the assumption for members states, as clarified in Section 2.4 of this paper is that members states’ legislation must contain rules demanding a connection between the person seeking asylum and the third country on the basis of which it would be reasonable for that person to go to that country. With the IMA 2023, the UK eroded that link and, therefore, posed itself in full breach of EU provisions and, more broadly, international agreements. The BHC concluded that the IMA 2023 should be disapplied in Northern Ireland because it breaches Article 2 of the Windsor Framework and Articles 3 (prohibition of torture), 4 (prohibition of slavery and forced labour), and 8 (right to respect for private and family life) of the ECHR.

The BHC ventured beyond the issues before the UK Supreme Court because it also dug into the core meaning of the definition of ‘safe third country’ arguing that even if those states might meet the definition because included in a whitelist by the UK government, EU law still requires an assessment of the safety of the third country for a particular applicant to verify that the individuals would not be subjected to torture, cruel, inhuman or degrading treatment or punishment. However, the IMA does not allow for this additional verification. The Court’s decision raised significant questions on the viability of the Rwanda plan and was probably one of the factors that lead the former Prime Ministers, Rishi Sunak, to call general elections a couple of weeks after the release of this judgment.

There is a double layer of irony in these happenings. First, it is ironic the fact that, between 2022 and 2024, the UK Government masked its intention to exclude people seeking asylum from protection using the language of ‘safe legal journey’. However, this policy did not safeguard migrants’ safety at all, as it was exposed by the BHC. Additionally, the second layer of irony lies in the fact that the concept of safe country has traditionally been used as a bordering tool. Yet, in this case, the border between Northern Ireland and the Republic of Ireland marked the end of punitive and dissuasive function of the concept of ‘safe third country’ within UK asylum provisions. Despite the EU started interpreting the ‘safe country’ concept in a way that was less favourable to people seeking asylum, it was still the EU which indirectly halted this increasingly problematic ‘slippery slope’ application of the term within the UK asylum law.

5. Conclusions

In an ideal world, all states would assume a fair share of responsibility, in line with burden-sharing principles for protecting people seeking asylum within their own jurisdiction to uphold the obligations of the Refugee Convention. However, many countries have begun to transfer asylum applicants from their own countries to return them to first, third or even fourth ‘safe countries’. This phenomenon stems from the interpretation and application of the concept of ‘safe country’, whose conceptual ambiguity has been used as a legal tactic to transfer people seeking asylum to different legislative systems. This term frequently yields results advantageous to the state, but it does not apply the same principles of solidarity to asylum seekers, who should be entitled to an individual assessment of their asylum claims. In this context, the concept of ‘safe country’ has increasingly been used as a barrier to prevent migration.

This paper explores the development of the concept ‘safe country’ in Europe and globally by clarifying its transformation from a compromise solution between States and people seeking asylum to an instrument of exclusion and deterrence used by States to transfer their responsibility to a different state. Despite this, EU legislation introduced minimum safeguards to states’ discretionary powers when it established the full respect of the principle of non-refoulment, and the existence of a connection between the receiving states and the asylum applicant.

Against this background, this paper analyses the interpretation and the application of the concept of a safe country, intended as ‘first safe country’ and ‘safe third country’, through a content analysis of the UK asylum legislation. Before Brexit, the UK borrowed this terminology from the EU member states, but adapted it to indicate primarily EU countries operating under the assumption that people seeking asylum had always transited through EU countries and, therefore, they should be returned there. Although the terminology remained unchanged, after Brexit, the increasingly hostile rhetoric of the UK Conservative government on immigration between 2022 and 2024 led the interpretation of the concept of ‘safe country’ to a more extreme interpretation. The UK Government established that no link between the person seeking asylum and the ‘safe country’ was necessary and decided to externalise asylum claims to Rwanda. This interpretation departs significantly from both the spirit and the letter of the Refugee Convention, although this has not yet been successfully challenged in court.

The paper also investigates the legal tensions between the UK legislative, executive and judiciary branches over the asylum externalisation policy. Through the analysis of the most recent jurisprudence, this paper concluded that the border between the Republic of Ireland and Northern Ireland had mitigated the harmful repercussion of the concept ‘safe country’ on people seeking asylum. This is particularly ironic, considering that the term has itself become a bordering tool. Additionally, although the EU contributed the slippery ‘expansion of application of the term ‘safe country’, it has indirectly stopped the UK from dismantling the last remaining safeguards for people seeking asylum.