Non-Refoulement and Extraterritorial Jurisdiction: State Sovereignty and Migration Controls at Sea in the European Context

Abstract

The principle of non-refoulement found in the UN Convention on the Status of Refugees has been widely regarded as the core element of the international refugee protection regime. However, in the recent era of restrictive external migration controls, its significance and ambit diminished to the extent that states began to regard it as a general moral principle that imposed only narrowly defined legal constraints. In particular, interception or interdiction of refugees on the high seas came to be regarded as activities falling outside the legal ambit of the non-refoulement obligation. However, in Europe, this has begun to change. The non-refoulement obligation found in Article 3 of the European Convention on Human Rights (ECHR) has been recognized as a legal constraint on state sovereignty in relation to migration controls on the high seas. This article scrutinizes how the developing concept of jurisdiction in human rights law, particularly as found in the ECHR, has expanded the scope of application of the principle of non-refoulement, and presents some important implications. The concept of state sovereignty has begun to undergo a paradigm shift that places extraterritorial human rights concerns relating to external migration controls squarely within a legal rather than merely a moral framework.

1. Introduction

No Contracting State shall expel or return (‘refouler’) a refugee in any manner whatsoever to the frontiers of territories where his life or freedom would be threatened . . . ¹

The principle of non-refoulement has long been recognized as the central principle of international refugee law. Regrettably, major refugee-intake countries have denied its extraterritorial applicability in the conduct of external migration controls such as interdiction or interception of refugees on the high seas. ² However, the European Court of Human Rights (ECtHR) in Hirsi has successfully challenged such state practices by expanding the scope of application of the non-refoulement obligation beyond state territory. ³

The principle of non-refoulement is enshrined in various European instruments. Typical examples are Article 19(2) of the Charter of Fundamental Rights of the European Union and Article 78(1) of the Treaty on the Functioning of the European Union. ⁴ In the context of joint-maritime operations at sea co-ordinated by Frontex, EU Regulation 656/2014 also ensures respect for the principle of non-refoulement. ⁵ Moreover, the principle of non-refoulement is regarded as binding in the course of EU military operations against human smuggling or trafficking in the southern central Mediterranean (EUNAVFOR MED), launched in June 2015. ⁶ This article exclusively focuses on the principle of non-refoulement under the European Convention on Human Rights (ECHR). ⁷

It examines how the shift in judicial analysis of the principle of non-refoulement under the ECHR has occurred. The shift may be attributable to the recent development of the concept of jurisdiction in human rights law. The point of departure is Article 1 of the ECHR, which reads as follows: ‘The High Contracting Parties shall secure to everyone within their jurisdiction the rights and freedoms defined in Section I of this Convention.’ ⁸ In the relevant cases, the ECtHR has duly recognized that jurisdiction within the meaning of Article 1 of the ECHR is established even beyond state territory in cases of physical custody of persons by state agencies. Since physical custody is implicated in interdiction operations at sea, the jurisprudence of the ECtHR has not only challenged the traditional concept of jurisdiction and state sovereignty, but it has also made a significant breakthrough in the protection of refugees intercepted by European states on the high seas.

In order to fully appreciate the developing concept of jurisdiction and its relevance in expanding the scope of application of the non-refoulement obligation, it is necessary to canvass the case law of the ECtHR that has recognized a wide range of accounts of jurisdiction beyond state territory. In particular, the Hirsi case is scrutinized as a case that, more than any other, has expanded the scope of application of the non-refoulement obligation even to the high seas in the context of the developing concept of jurisdiction in the ECHR. Furthermore, new potential to further expand the scope of application of the non-refoulement obligation is considered through an examination of the concept of state responsibility in public international law and the principles derived from the case law of the ECtHR. Before analyzing the case law of the ECtHR, it is important to acknowledge academic debates on the meaning of jurisdiction in human rights law. Because these debates provide a necessary context in which to read the case law of the ECtHR, the article begins there.

2. The meaning of jurisdiction under public international law and human rights law

In public international law, jurisdiction, as a core element of state sovereignty, has in general been regarded as being ‘closely related to the national territory’. ⁹ Simply stated, the concept of jurisdiction has traditionally been regarded as territorial in nature. ¹⁰ Two components of jurisdiction have been recognized: prescriptive and enforcement jurisdiction. ¹¹ Malcolm N. Shaw explains the two components as follows:

It is particularly necessary to distinguish between the capacity to make law, whether by legislative or executive or judicial action (prescriptive jurisdiction or the jurisdiction to prescribe) and the capacity to ensure compliance with such law whether by executive action or through the courts (enforcement jurisdiction or the jurisdiction to enforce). ¹²

In human rights law, including the ECHR, the meaning of jurisdiction has been the subject of considerable debate among scholars. In particular, scholars diverge in their opinions as to whether the notion of jurisdiction in human rights law is essentially territorial as found in general international law. On the one hand, it has been argued that the concept of jurisdiction found in human rights law should be distinguished from that found in general international law. For instance, Anja Klug and Tim Howe point out two different objectives of jurisdiction in general international law and human rights law:

The objective of the traditional notion of State jurisdiction . . . is to delineate the spheres of different sovereign States in a way that it respects the sovereignty of each State . . . Jurisdiction in the context of human rights law, however . . . defines the applicability of human rights obligations, and thus opens the possibility to assess State responsibility under human rights law. ¹³

By the same token, Conall Mallory has remarked that ‘[h]uman rights law jurisdiction does not deal with a State's rights, but with its responsibilities and obligations to which it has committed through accession to an international treaty’. ¹⁴ Marko Milanovic also argues that:

. . . the notion of jurisdiction in human rights treaties relates essentially to a question of fact, of actual authority and control that a state has over a given territory or persons. ‘Jurisdiction’, in this context, simply means actual power, whether exercised lawfully or not – nothing more, and nothing less. ¹⁵

Significantly, Klug and Howe suggest that in international human rights law, jurisdiction may be established by ‘factual control (over territory or person), de jure jurisdiction, or “a personal link”’. ¹⁶ From this perspective, the notion of jurisdiction in human rights law is not primarily territorial, but it is established by factual evidence such as effective control over persons even outside states’ territories.

On the other hand, other scholars argue for the necessity of retaining a territorial notion of jurisdiction in human rights law. Dominic McGoldrick argues that, ‘[t]he meaning(s) of extraterritorial application have to be within a general framework of jurisdictional analysis in public international law. They are questions of law, not of philosophy or ethics, although those disciplines may have affected the relevant law.’ ¹⁷ The primacy of the territorial notion of jurisdiction in the context of human rights law has often been proposed against the background of ‘potential clashes with foreign territorial jurisdictions’. ¹⁸ This perspective does not deny the notion of extraterritorial jurisdiction per se; however, it accepts extraterritorial jurisdiction only in exceptional cases.

3. Case law of the ECtHR

Banković ¹⁹

In Banković, the applicants were citizens of the Federal Republic of Yugoslavia (FRY) who brought an action against NATO states on behalf of themselves and their deceased family members, arguing that NATO's air strike on Radio Televizije Srbije (RTS) and the concomitant deaths of their family members during the Kosovo crisis were, inter alia, breaches of the right to life (Article 2 of the ECHR). ²⁰ The critical issue in this case was whether the extraterritorial activities of air strikes by NATO could trigger the application of the ECHR for NATO countries. Regarding jurisdiction, the Grand Chamber stated:

In keeping with the essentially territorial notion of jurisdiction, the Court has accepted only in exceptional cases that acts of the Contracting States performed, or producing effects, outside their territories can constitute an exercise of jurisdiction by them within the meaning of Article 1 of the Convention. ²¹

The Court flatly rejected a ‘cause-and-effect’ concept of jurisdiction as proposed by the applicants. ²² The Court recognized four exceptional cases to territorial jurisdiction: ²³ (i) extradition or expulsion; (ii) ‘effective control’ over a territory by military action; ²⁴ (iii) activities of ‘diplomatic or consular agents abroad and on board craft and vessels registered in, or flying the flag of, that State’; ²⁵ and (iv) effects produced outside by an action inside the territories. ²⁶ Having said that, the Court held that the case was inadmissible due to the failure to establish jurisdiction.

In the ensuing academic debates, Milanovic has argued that the Court's reasoning that the notion of jurisdiction found in Article 1 of the ECHR is essentially territorial is not only ‘unsupported by anything produced by the Court’ but also is in contradiction with ‘the Court's own established jurisprudence’. ²⁷ Milanovic grapples with models of extraterritorial jurisdiction found in the jurisprudence of the ECtHR, along with other human rights treaties, which are not essentially territorial in nature:

First, there is what I will call the spatial model of jurisdiction – a state possesses jurisdiction whenever it has effective overall control of an area. . . . Secondly, there is the personal model of jurisdiction (or ‘state agent authority’) – a state has jurisdiction whenever it exercises authority or control over an individual. ²⁸

Milanovic's critique on the decision of the Banković court is persuasive, as case law before and after Banković appears to have been inconsistent with the territorial principle set out by the Banković court, or at least it gives such an impression. ²⁹ The Issa case below stands at the heart of such an impression. ³⁰

On the other hand, Sarah Miller remarks that the Banković court introduced the primarily territorial notion of jurisdiction as a general thread found in both international law and the jurisprudence of the ECtHR, while fully recognizing exceptional cases to such a territorial notion of jurisdiction whether under public international law, e.g., the flag state jurisdiction, or in case law of the ECtHR. ³¹ With respect to critique of seemingly inconsistent case law relating to the concept of jurisdiction before and after Banković, Miller argues as follows:

The European Court's seemingly inconsistent treatment of exceptions to territorial jurisdiction becomes a coherent body of law when these cases are viewed as manifestations of a territorially centred rule . . . By extending extraterritorial jurisdiction only to cases where a signatory state is essentially exercising functional sovereignty abroad, the Court strikes a balance between the twin, competing purposes of the Convention as a regional, European instrument and as a universalist charter for human rights. ³²

In this regard, Miller highlights the requirement of ‘a strong nexus to state territory’ in establishing jurisdiction of the ECHR. ³³

Issa v. Turkey ³⁴

In many respects, this case is as controversial as the Banković case. The fact that it was decided after Banković makes the concept of jurisdiction all the more confused. In 2004, the Chamber in Issa dealt with the alleged killings of Iraqi shepherds by Turkish soldiers. There are three distinctive features in this judgment that are recognized and highlighted by Lord Brown in Al-Skeini. ³⁵ Firstly, the Court in Issa adopts a more or less flexible concept of control. ³⁶ Secondly, and importantly, the Court relies on the decisions of international bodies such as the Inter-American Commission of Human Rights and the Human Rights Committee (HRC), whose jurisprudence has developed a conception that is closely related to a ‘personal model’ of jurisdiction in Milanovic's terms. ³⁷ Thirdly, the Court focuses on ‘the activity of the contracting state, rather than on the requirement that the victim should be within its jurisdiction’, ³⁸ which gives the impression that the Court in Issa may have confused jurisdiction with state responsibility, a distinction that is discussed below. ³⁹

Although Milanovic has argued that the Issa court ‘endorsed the personal model of jurisdiction in addition to the spatial one’, ⁴⁰ Miller vehemently opposes this view of the Issa Court, stating that:

Under the logic of Issa, jurisdiction is not primarily territorial; a state is bound by the Convention wherever it acts, and its obligations abroad are no different from its obligations at home. This premise is diametrically opposed to the Court's conclusions in Banković . . . ⁴¹

The stark discrepancy in opinions regarding the concept of jurisdiction between Milanovic and Miller is hardly surprising, having regard to a wide range of decisions of the ECtHR.

3.1 The continuing debate on the meaning of jurisdiction in case law after Banković

In 2005, in Öcalan v. Turkey, ⁴² the Grand Chamber held that:

[i]t is common ground that, directly after being handed over to the Turkish officials by the Kenyan officials, the applicant was effectively under Turkish authority and therefore within the ‘jurisdiction’ of that State . . .even though in this instance Turkey exercised its authority outside its territory. ⁴³

It appears that the Court endorsed the ‘personal model’ adopted by the Issa court above. ⁴⁴ In 2009, the ECtHR in Al-Saadoon held that the UK government should prohibit the transfer of the applicants in Iraq, over whom it exercised ‘exclusive control’, to the Iraqi authorities. ⁴⁵ In this case, the Court found a jurisdictional linkage as follows:

The Court considers that, given the total and exclusive de facto, and subsequently also de jure, control exercised by the United Kingdom authorities over the premises in question, the individuals detained there, including the applicants, were within the United Kingdom's jurisdiction . . . ⁴⁶

The main issue in another important case, Medvedyev, is closely related to that of maritime interception. ⁴⁷ In 2010, a Cambodia-registered ship, the Winner, was intercepted on the high seas by a French frigate for the purpose of implementing anti-drug measures under the agreement with the Cambodian government. Later, crew members brought an action against France, arguing that they suffered the deprivation of liberty while being detained on the Winner by French authorities. The Grand Chamber in Medvedyev held that:

. . . as this was a case of France having exercised full and exclusive control over the Winner and its crew, at least de facto, from the time of its interception, in a continuous and uninterrupted manner until they were tried in France, the applicants were effectively within France's jurisdiction for the purposes of Article 1 of the Convention . . . ⁴⁸

In other words, de facto effective control over persons and the vessel is sufficient to establish a jurisdictional link without de jure jurisdiction, e.g., the flag state jurisdiction. ⁴⁹

In light of these cases, it does appear that, after Banković, the ECtHR has not emphasized the territorial nature of jurisdiction as much as it did previously. In this regard, Lawson proposes a ‘gradual approach to the notion of jurisdiction’ under which the obligation to uphold the rights of the ECHR depends on the degree of effective control over territory or persons. ⁵⁰ In contrast, Miller has argued that, ‘[t]he European Court has never found jurisdiction in cases involving a state's extraterritorial actions absent some preceding or subsequent nexus to the state's physical territory’. ⁵¹ Miller points out that the applicants in Öcalan were forced to return to the territory of Turkey and this fact gives rise to territorial nexus. ⁵² In fact, in Medvedyev, the concerned crew members also had been taken to the territory of France where they were convicted for drug-related charges. ⁵³

3.2 Al-Skeini: ⁵⁴ The end of dispute?

In 2011, the Grand Chamber of the ECtHR in Al-Skeini attempted to clarify the issue of jurisdiction. In brief, the Court acknowledged all the case law that is seemingly inconsistent. It confirmed that jurisdiction under Article 1 of the ECHR is ‘primarily territorial’. ⁵⁵ The Court also recognized ‘a number of exceptional circumstances’ outside of territorial boundaries, which could give rise to the establishment of jurisdiction. ⁵⁶ The Court acknowledged two critical exceptional categories: ‘state agent authority and control’ and ‘effective control over an area’. ⁵⁷

Relevant to the current focus on external migration controls, under the title of ‘State agent authority and control’, the Court, importantly, recognized:

. . . in certain circumstances, the use of force by a State's agents operating outside its territory may bring the individual thereby brought under the control of the State's authorities into the State's Article 1 jurisdiction. This principle has been applied where an individual is taken into the custody of State agents abroad. ⁵⁸

The Court explicitly endorses the Issa, Al-Saadoon and Medvedyev cases; all three cases share the fact that states exercised ‘total’, ‘full’ or ‘exclusive’ control over persons and places. ⁵⁹

The most significant statement of the Court in Al-Skeini is that, ‘[w]hat is decisive in such cases is the exercise of physical power and control over the person in question’. ⁶⁰ In this statement, there is no nexus to the physical territory of a state. In addition, the Court emphasized that the exceptions ‘must be determined with reference to the particular facts’. ⁶¹ In other words, ECtHR jurisprudence regarding jurisdiction is primarily based on facts rather than on a generalizable principle. ⁶²

In Al-Skeini, the claimants were relatives or family members of six Iraqi people who were allegedly killed or died due to mistreatment by British forces personnel. When this case was heard in the House of Lords in the United Kingdom, five alleged victims were held not to have been under the jurisdiction of the UK. Lord Rodger stated that the UK troops did not exercise ‘effective control’ over those who were killed in the course of military operations of British forces, even in the sense of Issa. ⁶³ On the other hand, the Secretary of State conceded that the victim, Mr. Mousa, was within the ambit of jurisdiction of the ECHR as he was detained and severely beaten in a British military base in Iraq, which caused his death. ⁶⁴

However, the ECtHR overruled the decision of the House of Lords, even finding jurisdiction in relation to the applicants other than Mr. Mousa:

. . . the United Kingdom (together with the United States) assumed in Iraq the exercise of some of the public powers normally to be exercised by a sovereign government. In particular, the United Kingdom assumed authority and responsibility for the maintenance of security in South East Iraq. In these exceptional circumstances, the Court considers that the United Kingdom, through its soldiers engaged in security operations in Basrah during the period in question, exercised authority and control over individuals killed in the course of such security operations, so as to establish a jurisdictional link between the deceased and the United Kingdom for the purposes of Article 1 of the Convention. ⁶⁵

‘Authority and control’ over persons is qualified by the circumstances in which the UK exercised the functions of ‘the public powers normally to be exercised by a sovereign government’. ⁶⁶ Milanovic rightly remarks that:

. . . the Court applied a personal model of jurisdiction to the killing of all six applicants, but it did so only exceptionally, because the UK exercised public powers in Iraq. But, a contrario, had the UK not exercised such public powers, the personal model of jurisdiction would not have applied. ⁶⁷

The decision in Al-Skeini may be interpreted as nothing but confirming all the case law in the past. In fact, the Banković court fully recognized the extraterritorial jurisdiction in the case where a contracting state ‘exercises all or some of the public powers normally to be exercised by that Government’ through ‘the consent, invitation or acquiescence of the Government of that territory’. ⁶⁸ However, significantly, the Court at least clarifies two things in relation to the scenario of interdiction of refugees at sea: (i) jurisdiction is not necessarily associated with a territorial nexus in the case of custody of a person on the vessel by a contracting state; and (ii) additional exceptional cases to the territorial nature of jurisdiction can be established according to facts (e.g., the level of control or influence) rather than fixed principles (e.g., territorial nature), which opens the door for the possibility of establishing jurisdiction in the case of a contracting state's indirect involvement in interdiction of refugees within the territorial waters of a third state.

3.3 Hirsi Jamaa and Others v. Italy ⁶⁹

The Hirsi case provides a meaningful point of contact between the developing concept of jurisdiction in the ECHR and the extraterritorial reach of the principle of non-refoulement. It is the Hirsi court's analysis of the concept of jurisdiction that has made it possible to expand the scope of the obligation of non-refoulement found in Article 3 of the ECHR. As the ECHR applies extraterritorially, so too does the non-refoulement obligation, which is embedded in Article 3 of the ECHR.

In 2009, about 200 migrants, including alleged asylum seekers and refugees, attempted to leave Libya for Italy by boat. On the high seas, however, they were intercepted by Italian coastguard and police vessels, and then were transferred onto Italian military ships. In the end, they were returned to Tripoli, Libya, without any process of identification, nor any attempt to determine claims for refugee status that may have been forthcoming. Later, 24 people of African origin who were among the returned group brought claims against Italy in the ECtHR, arguing that Italy had breached the ECHR and its Protocol by failing to secure their rights and freedoms, even though they were within its jurisdiction. ⁷⁰

The Court held that Italy was liable for breaching Articles 3 and 13 of the ECHR and Article 4 of Protocol No.4, as a result of the interdiction and its subsequent pushback activities that occurred on the high seas. ⁷¹ With respect to the jurisdictional issue, the Court held that factual evidence established jurisdiction within the meaning of the ECHR in that ‘the applicants were under the continuous and exclusive de jure and de facto control of the Italian authorities’. ⁷² The Court considered the fact that the applicants were transferred onto vessels flying Italian flags, which established de jure jurisdiction on the high seas according to international maritime law and the relevant domestic Italian laws. ⁷³ Furthermore, the Court also recognized de facto jurisdiction based on factual evidence of exclusive control of the applicants by Italian military personnel. ⁷⁴ Significantly, in this case, the Court did not require a territorial nexus in establishing jurisdiction; the interdicted migrants were not brought to the territory of Italy, but were pushed back to Libya from the high seas. In fact, the purpose of interdicting migrants on the high seas is to ‘escape’ from such a territorial connection in order to circumvent domestic legal constraints. ⁷⁵

Equally significant is the Court's analysis of the principle of non-refoulement, which it identified as an essential aspect of Article 3 of the ECHR. ⁷⁶ In relation to an alleged breach of Article 3 of the ECHR, the Court states that:

. . . the Court's task is . . . to ascertain whether there were sufficient guarantees that the parties concerned would not be arbitrarily returned to their countries of origin, where they had an arguable claim that their repatriation would breach Article 3 of the Convention. ⁷⁷

Judge Pinto De Albuquerque in the concurring opinion also recognized that:

. . . the non-refoulement obligation can be triggered by a breach or the risk of a breach of the essence of any European Convention right, such as the right to life, the right to physical integrity and the corresponding prohibition of torture and ill-treatment . . . ⁷⁸

In order to ascertain whether there was a breach of Article 3, the Court in Hirsi examined various documents, including those of the UNHCR and various human rights bodies, and held that:

. . . the Court considers that when the applicants were transferred to Libya, the Italian authorities knew or should have known that there were insufficient guarantees protecting the parties concerned from the risk of being arbitrarily returned to their countries of origin, having regard in particular to the lack of any asylum procedure and the impossibility of making the Libyan authorities recognise the refugee status granted by the UNHCR. ⁷⁹

In this regard, Italy was liable for breach of the principle of non-refoulement. The principle of non-refoulement was initially codified under Article 33 of the UN Convention on the Status of Refugees (The Refugee Convention), which provides that:

No Contracting State shall expel or return (‘refouler’) a refugee in any manner whatsoever to the frontiers of territories where his life or freedom would be threatened on account of his race, religion, nationality, membership of a particular social group or political opinion. ⁸⁰

Major refugee-intake countries such as the United States and Australia have denied that this provision has extraterritorial application to the scenario of interdiction on the high seas. ⁸¹

The decision in Hirsi offers a challenge to such state practice by removing the principle of non-refoulement from its original context ‘in the framework of international refugee law’, and placing it within the context of human rights law relating to the high seas. ⁸² In human rights law, the scope of application of the non-refoulement obligation is not limited to states’ territories. Since the principle is considered to be ‘the cornerstone of asylum and of international refugee law’, its extraterritorial application is of paramount significance in an era of restrictive external migration controls. ⁸³ Moreover, the doctrine of jurisdiction in human rights law may expand the scope of application of the non-refoulement obligation to the extent that it may cover an emerging trend of interdiction or interception practice within the territorial waters of a third country.

4. The emerging practice of interdiction at sea

So far, this article has shown that the ECtHR case law has developed the concept of jurisdiction to apply to cases where state agencies exert extraterritorial physical control over a person. As a result of Hirsi, a jurisdiction linkage within the meaning of Article 1 of the ECHR will exist where a state attempts to interdict or intercept refugees on the highs seas by using its own personnel and vessels. If it is determined that breach of the non-refoulement obligation occurred, the interdicting states will be held to be legally responsible for their conduct under Article 3 of the ECHR. Therefore, the enforceable rights under Article 3 of the ECHR, coupled with the developing concept of jurisdiction, has made it possible for the principle of non-refoulement to reach extraterritorially in a way that legally constrains interdiction practice of European states on the high seas.

However, the Grand Chamber's decision in Hirsi should not be understood as putting an end to interdiction policy altogether. Interdiction policy does not, in and of itself, breach the principle of non-refoulement. In fact, although the general rule is that no state may exercise its jurisdiction over a ship on the high seas except for the flag state, in exceptional cases, a state may implement interdiction operations on another flag vessel on the high seas in a legitimate manner. ⁸⁴ For example, a state may take appropriate measures in relation to people on board under Article 8 of the Smuggling Protocol, if it is believed that the vessel is implicated in human smuggling. ⁸⁵ The scope of the measures depends on the content of a flag state's authorization. ⁸⁶ In the case of stateless vessels, a warship may visit the vessels on the high seas under Article 110 of Convention on the Law of the Sea (UNCLOS), though it is not clear whether it authorizes the arrest of crew on board. ⁸⁷ Moreover, bilateral agreements may provide a legal basis for interception or interdiction. In fact, legal bases of Italy's interdiction operation on the high seas in Hirsi were bilateral agreements concluded between Italy and Libya in the period of 2007 to 2009. ⁸⁸

This being the case, two conditions must be met in order to hold a contracting state of the ECHR liable for its interdiction policy. First and foremost, it should be determined that jurisdiction under Article 1 of the ECHR is established. Even though there may be breach of the non-refoulement obligation under Article 3 of the ECHR, without the establishment of jurisdiction under Article 1, the ECtHR will declare a case inadmissible. Secondly, an actual breach of the non-refoulement obligation should be found in the process of interdiction. The absence of adequate identification procedure for refugees in the process of interdiction may give rise to a breach of the non-refoulement obligation. Goodwin-Gill remarks that relevant authorities should ‘identify all those intercepted, and keep records regarding nationality, age, personal circumstances and reasons for passage’. ⁸⁹ In a similar manner, Stephen H. Legomsky argues that, ‘[i]f interdiction must be used . . . adequate provision for full and fair refugee status determinations is critical.’ ⁹⁰

At this juncture, it is important to examine a different type of interdiction practice, joint patrol, which may not involve physical custody of a person by a contracting state of the ECHR. It appears that this practice has become another trend of external migration controls at sea. Paula García Andrade succinctly summarizes the emerging practice of joint patrol conducted by both a European state and an African state:

As regards the exercise of specific powers in the framework of joint sea patrols, a coastal State could authorize a third State, Spain in this case, to perform surveillance and interception activities, either by allowing the presence of Spanish agents on board the coastal State's ships, or by permitting the deployment of surveillance operations undertaken by Spanish State ships. In any case the powers of the authorised State's agents depend on the scope of the authorisation given by the agreement or memorandum signed for that purpose, and in any event national agents of the African countries involved should be on board, since the latter are entitled to enforce the third country internal legislation on border control with regard to vessels intercepted inside its territorial waters. ⁹¹

According to the extract, Spain is indirectly involved in the interdiction or interception of asylum seekers within the territorial waters of an African state – in this case, it is the coastal state that enforces the law within its territorial waters, not Spain. In fact, several contracting states of the ECHR have concluded bilateral agreements with northern African countries to such an effect. ⁹² With respect to characteristics of the agreements, Evelien Brouwer argues:

These agreements must be considered the result of intense bargaining between the European states and the ‘countries of transfer’, exchanging economical and development aid for cooperation at the sea borders, and within the third state, activities preventing persons to leave the latter state. ⁹³

In these circumstances, can a contracting state of the ECHR be held responsible for a breach of the non-refoulement obligation under Article 3 of the ECHR? In other words, can the jurisdiction in Article 1 of the ECHR be established in relation to European states’ indirect involvement in the interdiction of refugees within the territorial waters of African states? As examined in the case law above, the establishment of jurisdiction in the context of interdiction demands physical custody. The simple argument that refugees are sent back to the country of origin by indirect support of a contracting state of the ECHR without proving physical custody may not be sufficient to establish a jurisdictional linkage. In this regard, Mariagiulia Giuffré may be right in stating that, ‘EU member states that cooperate with third countries in patrolling external maritime borders are not always responsible under human rights treaties’. ⁹⁴ Thus, even if a breach of the non-refoulement obligation is found, a European state that co-operates with a third country may not be held liable for the breach.

4.1 State responsibility and complicity

Having observed a gap in human rights protections in these cases, some scholars look to the help of the public international law concept of state responsibility. For example, Giuffré introduces the International Law Commission (ILC)’s Draft Article on State Responsibility (ARSIWA) ⁹⁵ in order to, ‘provide a remedy because of a lack of the “jurisdictional link” between the state and the individuals concerned’. ⁹⁶ Under the concept of state responsibility, a state engaged in joint patrolling within the territorial waters of another state may be responsible for the breach of the principle of non-refoulement on three accounts: (i) ‘as a co-author of refoulement’ under Article 47 of ARSIWA; (ii) for providing ‘aid or assistance in the commission of an internationally wrongful act’ under Article 16; and (iii) for a breach of a principle of international law, i.e., ‘positive due diligence obligations’. ⁹⁷ As joint maritime patrols encompass various forms of participation from European states (e.g., deployment of naval or air forces, and ship-riders agreement), the scope of state responsibility necessarily varies in specific cases. ⁹⁸ This article especially pays attention to two scenarios: (i) interception by European vessels under the authority of enforcement officers on board from the coastal state (‘significant’ contribution); and (ii) interception by the coastal state's vessels with the help of agents on board from European states (‘limited’ contribution).

4.1.1 As a co-author of refoulement

Article 47 of ARSIWA copes with a situation where several states are involved in the same internationally wrongful act whether independently or co-operatively, e.g., concerted military attacks. ⁹⁹ The ILC Commentary remarks that ‘. . . in such cases each State is separately responsible for the conduct attributable to it, and that responsibility is not diminished or reduced by the fact that one or more other States are also responsible for the same act’. ¹⁰⁰ Such a co-authorship may be found in the case of ‘direct’ participation in or ‘sufficiently significant’ contribution to an internationally wrongful act. ¹⁰¹ Scenario 1 (‘significant’ contribution) above may be the case.

However, it should be remembered that participating European states have no legal authority to interdict irregular migrants; ¹⁰² technically, they are ‘merely’ providing ‘aid or assistance’ to a coastal state's own patrolling activities. In this context, it may be difficult to demarcate co-authorship and complicity in relation to European participation in joint maritime patrols. Ian Brownlie has pointed out that the provision of ‘aid or assistance’ in the context of aggression may not give rise to joint-responsibility unless it is accompanied with ‘the specific purpose of assisting an aggressor’. ¹⁰³ However, as will be demonstrated below, such an ‘intent’ element is more problematic than helpful in finding state responsibility in relation to European states’ participation in the joint patrolling.

4.1.2 Complicity under Article 16 of ARSIWA

Article 16 of ARSIWA is often invoked in relation to states’ indirect involvement (complicity) in an internationally wrongful act. Article 16 provides:

A State which aids or assists another State in the commission of an internationally wrongful act by the latter is internationally responsible for doing so if:

1. (a) that State does so with knowledge of the circumstances of the internationally wrongful act; and

2. (b) the act would be internationally wrongful if committed by that State. ¹⁰⁴

Relying on this article, Giuffré argues that, ‘a state may be responsible for violation of the principle of non-refoulement where it knowingly assists another state to return refugees to a place where their life or liberty might be threatened’. ¹⁰⁵ However, a careful reading of Article 16 may not warrant such a view. Although there has been a growing acceptance that Article 16 is reflective of customary international law, ¹⁰⁶ Maarten den Heijer rightly observes that ‘[t]he international law concept of aiding and assisting, or complicity, is not without controversial elements’. ¹⁰⁷ In other words, it is not a settled area of law. In particular, since no definition of ‘aid or assistance’ is provided in ARSIWA, neither the forms of complicity (e.g., whether it should be active, thus excluding omission), nor the nexus elements (e.g., whether it requires substantial contribution or mere participation), nor the subjective requirements (e.g., whether intent of the accomplice matters) are clear-cut. ¹⁰⁸

Nevertheless, a bottom-line understanding of the meaning and scope of ‘aid or assistance’ in current state practice may be suggested as follows: (i) the act of complicity should be ‘significant’ contribution to outcomes, albeit not necessarily in the form of essential or ‘indispensable’ contribution; ¹⁰⁹ (ii) the act of complicity needs to be ‘in the form of a positive act’, thus excluding ‘active incitement’ or ‘mere omission’; ¹¹⁰ (iii) controversially, the act of complicity has to be accompanied with ‘intent’ as well as ‘actual knowledge’ of ‘the circumstances of the internationally wrongful act’; ¹¹¹ and (iv) a complicit state must be bound by the same primary obligation that another state has breached (Article 16(b), the Pacta Tertiis rule). ¹¹²

In the case of Scenario 1, it may be forcefully argued that there is a positive act on the part of a European state, which significantly contributes to maritime patrolling of a coastal state. On the other hand, the ‘aid or assistance’ in Scenario 2 may not meet the requirement of ‘significant contribution’. With Scenario 1, then, the subjective element, i.e., intent, is a critical issue. This requirement of intent, though controversial, appears to have been recognized at least in state practice. ¹¹³ In fact, commentary on ARSIWA states that, ‘Article 16 deals with the situation where one State provides aid or assistance to another with a view to facilitating the commission of an internationally wrongful act by the latter’. ¹¹⁴

Accordingly, for a European state to be held responsible, it arguably must have provided aids such as patrolling vessels ‘with a view to facilitating’ a breach of the principle of non-refoulement. However, it should be noted that joint patrolling programmes, in which the European border agency, Frontex, has been involved, have been operated ostensibly for the purpose of combating irregular migration. ¹¹⁵ Even if it is conceded that a lesser stringent standard of knowledge should be adopted in the place of the controversial requirement of ‘intent’, it is important to recognize that the International Court of Justice (ICJ) has connected such knowledge with ‘the specific intent of the principal perpetrator’. ¹¹⁶ In other words, a European state must be aware of the coastal state's intention to violate the principle of non-refoulement in the course of joint maritime patrols. Accordingly, it goes back to the issue of ‘intent’. This being the case, it may be argued that the ‘intent’ element, whether it is required from a complicit party or a primary perpetrator, has rendered it difficult to hold a European state complicit in the breach of the principle of non-refoulement under Article 16 of ARSIWA.

4.1.3 Positive due diligence obligations

Some may argue that a complicit European state may be held responsible for the breach of positive due diligence obligations. International law has duly recognized positive due diligence obligations in various fields of law, most prominently in international environmental law. ¹¹⁷ For example, Miles Jackson has observed that, ‘. . . many instances of state participation in the harms caused by non-state actors are swept up by broader positive obligations imposed on states to protect against harms to other states or individuals’. ¹¹⁸ On a European level, positive obligations have also been recognized in relation to rights set out in the ECHR, whether independently or in conjunction with Article 1 of the ECHR. ¹¹⁹

It is significant to note that positive due diligence obligations are normally contemplated as duties of states within their territories – in other words, due diligence obligations ‘retain their territorial character’. ¹²⁰ Accordingly, in the context of external migration controls, it is not that breach of positive obligations incurs state responsibility, but that extraterritorial jurisdiction should give effect to the positive obligations so as to trigger the issue of state responsibility. After all, it goes back to the issue of jurisdiction rather than state responsibility.

4.2 An approach based on the jurisprudence of the ECtHR

The state responsibility approach certainly has some merits in the case of complicity; however, ambiguous or controversial elements have rendered it much less applicable and effective in relation to holding a complicit European state responsible under ARSIWA, at least in the context of joint maritime patrols. More fundamentally, the ECtHR has been adamant that, in order for the ECHR to apply to a particular case, the first threshold is to establish jurisdiction under Article 1 of the ECHR, not attribution under state responsibility. ¹²¹ The ECtHR has not adopted the attribution concept as a legitimate means to establish ‘jurisdiction’. ¹²² Michael O'Boyle plausibly argues that, ‘[t]he [state responsibility] approach . . . only makes sense as regards a treaty which has no limiting “jurisdiction” clause . . .’. ¹²³ Thus, in the European context, the requirement of establishing jurisdiction under Article 1 of the ECHR cannot be circumvented by way of introducing the concept of state responsibility under ARSIWA.

This being the case, instead of relying on the concept of state responsibility, the focus should be shifted to the ECtHR's own case law – de jure or de facto jurisdiction. In relation to Scenario 1 above, Andrade raises an interesting question in relation to joint patrolling by Spain and Senegal:

The situation envisaged would be more complicated if the person stopped had gone on board the Spanish ship which participated in the joint patrol. In that case, could we consider that, since the person would be under Spanish jurisdiction, the return to Senegalese territory would imply a violation of the ‘non-refoulement’ principle by Spain? ¹²⁴

It is probable that, in Scenario 1, European states still have de jure jurisdiction (the flag state jurisdiction) over matters on their vessels even within the territorial waters of a third country. Richard A Barnes states that, ‘[f]lag States enjoy prescriptive and enforcement jurisdiction over ships flying their flag wherever the vessel is located. When a ship is within internal waters, port, or the territorial sea, jurisdiction is concurrent with the port/coastal State’. ¹²⁵ Debates may arise concerning the characteristic of coastal state jurisdiction – whether the coastal state has ‘plenary jurisdiction’ or whether it can only exercise jurisdictional power over certain matters in a limited way. ¹²⁶ In any event, it appears that flag state jurisdiction is not forfeited simply by a ship's entry into the territorial waters of a third country.

Significantly, it is probable that European states (flag states) may enjoy immunity from the coastal state's enforcement jurisdiction under Article 32 of the UNCLOS which reads, ‘[w]ith such exceptions as are contained in subsection A and in articles 30 and 31, nothing in this Convention affects the immunities of warships and other government ships operated for non-commercial purposes’. ¹²⁷ The European vessels engaged in the interdiction operations are government-operated patrol ships, which certainly are within the ambit of ‘warships and other government ships operated for non-commercial purposes’. ¹²⁸

Furthermore, it also may be argued that European states can exercise effective control over intercepted migrants, thus de facto jurisdiction, on the grounds that they are held in custody on European vessels by joint crews that are made up of European crews and the coastal states’ crews. ¹²⁹ Therefore, in the light of jurisprudence of the ECtHR in relation to extraterritorial jurisdiction as examined above, such findings may arguably establish jurisdiction of a European state within the meaning of Article 1 of the ECHR.

Once jurisdiction is established, as shown above in Hirsi, the standard of knowledge required for holding a state responsible for the breach of the non-refoulement obligation under Article 3 of the ECHR is ‘actual’ or ‘constructive’ knowledge (‘should have known’). ¹³⁰ Here, the jurisprudence of the ECtHR clearly differs from the controversial requirement of Article 16 of ARSIWA, that is, ‘intent’ of a participating state.

In relation to Scenario 2, the analysis is more complicated. There is neither de jure jurisdiction by virtue of flag state jurisdiction nor de facto jurisdiction by way of effective control over persons. In this circumstance, it seems to be hard to establish jurisdiction of a European state in relation to the ECHR under the current jurisprudence of the ECtHR. That being said, having regard to the progressive characteristic of the jurisprudence of the ECtHR in relation to extraterritorial jurisdiction, ¹³¹ we do hope to see the Court finding a way to establish extraterritorial jurisdiction in the case of complicity so that it may uphold the principle expressed in Issa: ‘Article 1 of the Convention [ECHR] cannot be interpreted so as to allow a State party to perpetrate violations of the Convention on the territory of another State, which it could not perpetrate on its own territory’. ¹³²

5. Conclusion: Beyond Europe

Jurisdiction is a core element of state sovereignty that previously had been understood as essentially a territorial concept. ¹³³ Furthermore, it has generally been thought that the concept of state sovereignty as it relates to border control is founded on the ‘unconditional’ power of a state. ¹³⁴ However, this has begun to change in Europe. Case law of the ECtHR has in significant ways modified our understanding of the concepts of jurisdiction and state sovereignty.

State sovereignty has begun to reflect human rights concerns such as the principle of non-refoulement, even beyond states’ territories. In this regard, the concept of state sovereignty in relation to external migration controls has undergone a paradigm shift from ‘unconditional’ sovereignty to ‘accountable’ sovereignty, at least within the European context. ¹³⁵ The challenge is based on a liberal interpretation of the term, ‘jurisdiction’, found in Article 1 of the ECHR.

Importantly, on the international level, the non-refoulement principle is also found in many international treaties such as in Article 7 of the International Covenant on Civil and Political Rights (ICCPR) and Article 3 of the Convention against Torture and other Cruel, Inhuman or Degrading Treatment or Punishment. ¹³⁶ International human rights bodies, such as the Human Rights Committee (HRC) of ICCPR and the Committee against Torture, also have confirmed extraterritorial application of these human rights instruments. ¹³⁷ The rulings of international human rights bodies may not have been as effective as the decisions of the ECtHR in domestic courts; however, they certainly give momentum to furthering discourse on extraterritorial application of the non-refoulement principle in other jurisdictions. The ‘forgotten’ principle of non-refoulement in an era of restrictive external migration controls has revived in Europe. This change of state practice in Europe, coupled with decisions of international human rights bodies, may give rise to worldwide impact on refugee laws and policies.