Thursday 31 March 2016

What’s wrong with the UK immigration rules governing the rights of dual-British and EU citizens? Comment on the Lounes reference.

Dr Alina Tryfonidou, Associate Professor in EU Law, University of Reading

Last month, the High Court of England and Wales decided to make a reference to the ECJ for a preliminary ruling in the case of Lounes. The case – which is currently stayed awaiting the ECJ judgment on the reference – involves a couple comprised of a dual-British and Spanish national and a third-country national, who have been living together in the UK since 2013 and where they now seek to rely on the EU free movement rights of the former in order to claim family reunification rights for the latter. The question that is central to the resolution of the case is whether a Member State can provide in its legislation that from the moment that a national of a Member State acquires also the nationality of the Member State to which (s)he moved in exercise of EU free movement rights, (s)he ceases to be a beneficiary of EU free movement rights and of the secondary rights attached to them, such as family reunification rights.

Factual Background

Mr Lounes (the Claimant) is an Algerian national who entered the UK for the first time in 2010 and overstayed illegally. In 2013, whilst (still illegally) in the UK, he met and formed a relationship with Mrs Ormazabal, and in 2014 the couple married in London. Mrs Ormazabal was, originally, only a Spanish national. She moved to the UK in 1996 to study and since 2004 she has been employed on a full-time basis. In 2009, she acquired British citizenship whilst maintaining, also, her Spanish nationality and, thus, since then she is a dual-British and Spanish national.

Shortly after the marriage between the Claimant and Mrs Ormazabal took place, the Claimant applied to the Secretary of State for the Home Department (the Defendant) for an EEA residence card as the spouse of an EEA national who had made use of her free movement rights. This, however, brought the Claimant to the attention of the immigration authorities, which led to him being served with a ‘Notice to a person liable to removal’ on the ground that he had overstayed his visa in breach of immigration legislation, as well as a notice of a decision to remove him from the UK. In addition, the Defendant decided to refuse to issue the Claimant with the EEA residence card he had applied for on the ground that the Claimant’s ‘sponsor’ (i.e. Mrs Ormazabal) ‘does not meet the definition of an EEA national defined under Regulation 2 of the Immigration (EEA) Regulations 2006 (as amended on 16 July 2012) as she is also a British citizen’. In response, the Claimant applied for judicial review of the Defendant’s decision.

Legal Background

Under EU law, Union citizens can move to and/or reside in another Member State, usually this being a Member State other than that of their nationality. This right derives from the free movement provisions of the Treaty (Arts. 21, 45, 49, 56 TFEU), which are supplemented by pieces of secondary legislation which, inter alia, provide a number of secondary rights attached to the primary free movement rights stemming from the Treaty provisions. One such set of secondary rights are family reunification rights. Family reunification rights for all Union citizens who exercise free movement rights are laid down in Article 3.1 of Directive 2004/38 (also known as the 'citizens' Directive'), which provides that ‘This Directive shall apply to all Union citizens who move to or reside in a Member State other than that of which they are a national, and to their family members as defined in point 2 of Article 2 who accompany or join them’ (emphasis added). Point 2 of Article 2 includes ‘the spouse’ in the list of family members mentioned in Article 3.1, whilst Recital 5 of the Directive notes that ‘The right of all Union citizens to move and reside freely within the territory of the Member States should, if it is to be exercised under objective conditions of freedom and dignity, be also granted to their family members, irrespective of nationality’ (emphasis added). Accordingly, on the facts in Lounes, the Claimant argued that he can rely on Directive 2004/38 to claim that he can join his wife – who is a Spanish national who has moved to and settled in a Member State (UK) other than that of her (originally only) nationality – in the UK, where she resides.

The national rules applicable in this case are the Immigration (European Economic Area) Regulations 2006 (SI 2006/1003), as these have been amended in 2012, by SI 2012/1547 and SI 2012/2560. The definition of the various terms used in them is provided in section 2(1) of the Regulations. Instead of ‘Union citizen’, the Regulations use the (broader) term ‘EEA national’ when referring to the beneficiaries of free movement rights which are derived from EU law and can be exercised in the UK. The original version of the Regulations provided that ‘“EEA national” means a national of an EEA State’. The 2012 amendments amended this definition, which – following the amendment made by SI 2012/2560 – now reads ‘“EEA national” means a national of an EEA State who is not also a British citizen’ (emphasis added).

The effect of the above amendments is that a British national cannot rely in the UK on the rights stemming from the EU free movement provisions (including family reunification rights), and this is so even when that person is a dual British-EU national who has made use of his free movement rights. On the facts in the Lounes case this meant – as the referring court noted that it was submitted by the Defendant – that ‘after naturalisation as a British citizen, Mrs Ormazabal could no longer exercise any EU right to free movement within the UK despite her Spanish nationality and Union citizenship. In that respect, she was in the same position as other British citizens living in the UK who did not have dual nationality. She could, of course, exercise her EU right to free movement in other EU countries. It followed that the Claimant’s application to reside in the UK had to be determined according to UK domestic immigration law only; he could not rely upon the Directive and the EEA Regulations 2006’.

The Reference for a Preliminary Ruling

The question referred to the ECJ is the following:

‘Where a Spanish national and Union citizen:
i) moves to the United Kingdom, in the exercise of her right to free movement under Directive 2004/38/EC; and
ii) resides in the United Kingdom in the exercise of her right under Article 7 or Article 16 of Directive 2004/38/EC; and
iii) subsequently acquires British citizenship, which she holds in addition to her Spanish nationality, as a dual national; and
iv) several years after acquiring British citizenship, marries a third country national with whom she resides in the United Kingdom;
are she and her spouse both beneficiaries of Directive 2004/38, within the meaning of Article 3(1), whilst she is residing in the United Kingdom, and holding both Spanish nationality and British citizenship?’

The national court, therefore, asks, in essence, whether the amendments made in the 2006 Regulations which preclude Union citizens who, also, acquire British citizenship, from continuing to benefit from the rights stemming from the EU free movement provisions, are contrary to Article 21 TFEU and Directive 2004/38.

The referring court in its judgment observed that this is a test case and that similar claims have been stayed awaiting the outcome of this case.

Comments: What must the ECJ rule?

The Explanatory Notes to the amending instruments point out that the amendments made to the 2006 Regulations described earlier, were made in order to give effect to the ECJ judgment in Case C-434/09 McCarthy. However – and this is the important question – are the Regulations (as amended) a correct implementation of the ECJ’s judgment in McCarthy?

McCarthy involved a claim for family reunification rights in the UK by a dual UK and Irish national who was born – and had always lived – in that Member State. The only ‘external’ element in the case was the fact that Mrs McCarthy was a national of two Member States. The Court found that the situation was purely internal to the UK and, thus, was not covered by Article 21 TFEU, since it involved a Union citizen who had never exercised free movement rights and who had always resided in a Member State of which she is a national, where she was claiming family reunification rights. This indicates that the mere fact that someone holds the nationality of two Member States does not suffice for bringing that person within the scope of the free movement provisions of the Treaty, in the absence of the exercise of free movement rights. The Court in McCarthy distinguished the facts in that case from those in Garcia Avello which, according to it, involved dual Member State nationals residing in one of their States of nationality who had never exercised free movement rights but whose situation was not purely internal to a Member State because the contested measure could potentially deter the exercise of such rights.

The Court’s approach in the above cases demonstrates that what makes a situation purely internal to a Member State is not the mere fact that a Union citizen pursues an economic activity and/or resides in a Member States the nationality of which he possesses but rather the fact that the rights stemming from the free movement provisions have not been exercised and/or are not going to be impeded as a result of the contested measure. This is, in fact, also exemplified in an older case (Scholz) where the Court held that a German national who – having exercised free movement rights – settled in Italy, was still covered by (what is now) Article 45 TFEU and could rely on that provision against Italy, even after becoming naturalised as an Italian national.

Accordingly, in all cases, the important question is not whether a Union citizen seeks to exercise rights deriving from the EU free movement provisions in her Member State of nationality, but rather, whether there is a sufficient cross-border element. And although the Court’s approach in cases such as Angonese, Carpenter, Schempp and Ruiz Zambrano (all of which involved Union citizens seeking to enforce their EU rights in their Member State of (sole) nationality) demonstrates that it is increasingly easy to establish such a cross-border element on the facts of a case, there are still instances where such a link fails to be established, obvious, recent, examples of this being Iida (which involved a merely hypothetical obstacle to free movement[1]) as well as McCarthy itself.

In McCarthy, it was clear that free movement rights were neither exercised nor were they going to be impeded as a result of the contested measure. The situation in Lounes, however, is fundamentally different, because it does involve the exercise of free movement rights (the original movement of Mrs Ormazabal from Spain to the UK) and the contested national measure (i.e. the removal of the Claimant from the UK) is capable of impeding the exercise of those rights.[2]

However, which right exactly which is stemming from the free movement provisions can be impeded as a result of the refusal of family reunification rights in a situation such as that pertaining in Lounes?
The aim of the free movement provisions is, primarily, to protect the free movement rights of Union citizens. These have, always, been read as including the right to move to another Member State to take-up an economic activity and/or residence there; in other words, it is clear that these provisions cover the initial right of a Union citizen to move to, and settle in, the territory of another Member State. However, in more recent case-law, the Court has made it clear that the free movement provisions, also, require Member States not to act in a way which discourages Union citizens from continuing to exercise the free movement (see S & G case) and/or residence (see Metock case) rights they derive from the EU free movement provisions: in other words, Member State measures amount to a restriction caught by the EU free movement provisions not only when they prevent a Union citizen from exercising his/her initial right to move to another Member State and/or to take-up residence there, but also, if they discourage a Union citizen from continuing to reside in the territory of the host State or, more broadly, from continuing to exercise the rights stemming from the free movement provisions (e.g. to continue to pursue an economic activity in a cross-border context).

Hence, in a situation like that in Lounes, where the initial rights to free movement and taking-up of residence in the host State had already been exercised by Mrs Ormazabal and would not be impeded as a result of the contested measure, a breach of Article 21 TFEU can, still, be established, since the contested measure can discourage her from continuing to reside in the UK: as noted in Metock, Union citizens are liable to be discouraged from continuing to reside in the territory of the host Member State if they are refused the right to be joined or accompanied by their close family members in its territory.

Legislation such as the 2006 Regulations which refuse family reunification rights to all Union citizens who hold (or, also, hold) British nationality – irrespective of whether there is a cross-border element on the facts of the case – amounts, therefore, to a restriction on the exercise of the rights stemming from the free movement provisions. Such a restriction is, clearly, unjustifiable under the public policy derogation from the free movement provisions, as it is an en bloc refusal of free movement rights to Union citizens and their family members, and is, thus, not based on the personal conduct of the persons involved in a specific case, which is required by Article 27(2) of Directive 2004/38. 

Accordingly, the ECJ must rule that legislation such as the 2006 Regulations is in breach of the EU free movement provisions (and of Directive 2004/38). The UK will only correctly implement McCarthy if it amends its legislation to provide that persons who hold dual nationality of the UK and another Member State are not beneficiaries of the rights stemming from the EU free movement provisions (and the secondary legislation implementing them) in the UK only if they have never exercised free movement rights (as established in McCarthy) and/or if the rights stemming from the free movement provisions will not be impeded as a result of the contested measure (as established in Garcia Avello).

Barnard & Peers: chapter 13
JHA4: chapter I:6

[1]           For a discussion of the distinction between potential obstacles and purely hypothetical obstacles to the exercise of free movement rights for the purpose of establishing a cross-border element see A. Tryfonidou, ‘(Further) Signs of a Turn of the Tide in the CJEU’s Citizenship Jurisprudence, Case C-40/11 Iida, Judgment of 8 November 2012, not yet reported’ (2013) 20(2) MJ 302, pp. 307-313.
[2]           As Advocate General Tesauro noted in his Opinion in Singh, para. 5: ‘the simple exercise of the right of free movement within the Community is not in itself sufficient to bring a particular set of circumstances within the scope of Community law; there must be some connecting factor between the exercise of the right of free movement and the right relied on by the individual’. In other words, in order for a breach of the free movement provisions to be established, it does not suffice that free movement rights have been exercised, but it must be proved, in addition, that such rights are impeded as a result of the contested measure. This is also evident by the fact that in some cases (e.g. Cowan) the Court read the free movement provisions together with the prohibition of discrimination laid down in Article 18 TFEU and (only) found a breach of the latter, where a person who had exercised free movement rights was discriminated against in the host State but the discrimination complained of was not capable of impeding the exercise of those rights.

Tuesday 29 March 2016

Preliminary references and investment tribunals: is the Luxembourg Court extending a helping hand?

Hannes Lenk, PhD Candidate at the University of Gothenburg

The relationship of arbitral tribunals with the Court of Justice of the European Union (CJEU) has been the subject of a long-lasting juridical struggle. The current position is as simple and pragmatic as it is controversial. Commercial arbitration tribunals are not considered to be a ‘court and tribunal of a Member State’ within the meaning of the Article 267 TFEU and, thus, unable to refer questions to the CJEU on matters of interpretation of EU law.  At the same time, it is an open secret that questions of EU law do arise during arbitration proceedings, and there is an inherent risk that tribunals get it wrong—at least sometimes. In commercial arbitration these shortcomings might be addressed through the indirect involvement of domestic courts and the CJEU at the recognition and enforcement stage of arbitral awards. A similar possibility might not exist in investment arbitration and for some time now the question of whether or not investment tribunals are entitled to request preliminary references from the CJEU has been simmering under the surface of a deeply politicised debate on investor-state dispute settlement provisions in currently ongoing negotiations for deep and comprehensive trade and investment agreements with, inter alia, Canada and the US. A recent opinion of Advocate General Wathelet might break new ground in this debate and prepare the field for future judicial dialogue.

Commercial arbitration: from Nordsee to Eco Swiss

'Article 267 TFEU is an important instrument for cohesion and coherence in the judicial system of the European Union (EU), including domestic courts as ‘ordinary courts of the EU legal order’ (Opinion 1/09, para. 80). By way of establishing a judicial dialogue, the preliminary reference mechanism guarantees that individuals have their rights under EU law enforced in domestic courts, and assures a uniform interpretation and application of EU law in all Member States. Notably, the decision to request a preliminary reference is generally within the discretion of domestic courts, which are obligated to refer questions only in limited circumstances, i.e. in instances where the case is pending before a domestic court of last instance.

Article 267
1. The Court of Justice of the European Union shall have jurisdiction to give preliminary rulings concerning:
(a) the interpretation of the Treaties;
(b) the validity and interpretation of acts of the institutions, bodies, offices or agencies of the Union;
2. Where such a question is raised before any court or tribunal of a Member State, that court or tribunal may, if it considers that a decision on the question is necessary to enable it to give judgment, request the Court to give a ruling thereon.
3. Where any such question is raised in a case pending before a court or tribunal of a Member State against whose decisions there is no judicial remedy under national law, that court or tribunal shall bring the matter before the Court.
4. If such a question is raised in a case pending before a court or tribunal of a Member State with regard to a person in custody, the Court of Justice of the European Union shall act with the minimum of delay.'

However, Article 267 TFEU includes a significant procedural limitation. In order to to request a reference from the CJEU the judicial body must be covered by the concept of ‘any court or tribunal of a Member State’. The CJEU has historically interpreted this concept restrictively. In Dorsch Consult the CJEU clarified the characteristics that need to be taken into account. Accordingly, a ‘court of tribunal’ is any judicial body that exercises judicial functions, i.e. that is (a) established by law, (b) a permanent institution, (c) with compulsory jurisdiction, (d) whose procedure is inter partes, (e) applying rules of law, and (f) acts independent of other branches of government. Applying these criteria to a commercial arbitration tribunal, the CJEU subsequently declared in Nordsee that despite “certain similarities between the activities of the arbitration tribunal … and those of an ordinary court”, the tribunal in question was not a ‘court or tribunal’ within the meaning of the preliminary reference procedure.

Arbitral tribunals are, therefore, left without guidance on the interpretation of EU law where this becomes relevant during the arbitration proceedings. From an EU law perspective, the adverse effect of incorrect interpretation and application of EU law in commercial arbitration is mitigated by the indirect involvement of domestic courts, and by association the CJEU. Indeed, domestic courts play an important role in supporting the arbitral tribunal upon request, as well as in the recognition and enforcement of arbitral awards. In Eco Swiss the CJEU emphasized that domestic courts are generally required to assess the compatibility of arbitral awards with EU public policy and may request a preliminary reference from the CJEU to that end. The award in Eco Swiss was considered a violation of EU competition rules (now Article 101 TFEU), which, according to the CJEU, constitutes a ‘fundamental provision which is essential for the accomplishment of tasks entrusted to the [Union]’ (para. 36). The CJEU furthermore clarified that it is to be considered part of public policy in the meaning of Article V(1)(c) and (e), and II(b) of the New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards.  Consequently, arbitral awards that are irreconcilable with EU public policy are unenforceable within the territory of the Member States under domestic and international law.

Investment arbitration: commercial arbitration in disguise or something else entirely?

This approach is open for much criticism, not least because the concept of EU public policy remains notoriously undefined. “In light of a constant referral of additional tasks upon the European Union over the last twenty years, it has become increasingly difficult to identify those provisions which may be regarded as fundamental for those tasks.” (Basedow, p. 373). Nonetheless, from an EU law perspective it appears to provide a pragmatic procedural solution that reserves the involvement of the CJEU and assures the application of EU law in domestic courts. Particularly in the area of investment arbitration it has been suggested that invoking EU public policy at the enforcement stage might prevent some of the most controversial awards from gaining legal effect within the EU legal order. But investment arbitration is in many ways different from commercial arbitration. The vast majority of proceedings is governed by the rules of the International Centre for Settlement of Investment Disputes (ICSID). Article 54 ICSID provides for the automatic recognition and enforcement of awards, excluding domestic courts from any involvement in the review of ICSID awards vis-à-vis public policy. To make matters worse, non-ICSID awards are not seldom enforced outside the territory of the respondent state. Enforcement of the controversial Micula award, for instance, is currently sought in the US.

Gaffney and Basedow have recently advocated the view that investment tribunals should be able to request references under Article 267 TFEU. It is in this context noteworthy that the CJEU in Nordsee conceded to the possibility that an arbitral tribunal might fall within the scope of Article 267 TFEU, provided that the tribunal derives its jurisdiction not exclusively from party autonomy of the disputing parties, but instead involves the exercise of state authority to the extent that it can be considered an institution of the state. This view was later confirmed in Ascendi, a request from the Tribunal Arbitral Tributário in Portugal. The CJEU observed that Portuguese law provides for the resolution of tax disputes through arbitration, which also regulates the functioning and constitution of the tribunal. “[The Tribunal’s] jurisdiction stems directly from the provisions of Decree-Law No 10/2011 and is not, as a result, subject to the prior expression of the parties’ will to submit their dispute to arbitration”, the CJEU concluded (para. 29). Not unlike the Tribunal Arbitral Tributário, investment tribunals are an alternative dispute settlement system provided for in law, i.e. the the underlying investment agreement, which constitute a “non-transient element of [the domestic] judicial system” (Basedow, p. 379-380).

The idea to construe investment tribunals as ‘court or tribunal’ for the purpose of Article 267 TFEU is not merely a scholarly endeavor to square the circle. There are signs from within the CJEU that this might present an acceptable solution to the problem of integrating investment tribunals in the EU legal order. In his recent Opinion in Genentech, a preliminary ruling from the Cour d’appel de Paris concerning the notion of EU public policy in the recognition and enforcement of arbitral awards, Advocate General Wathelet presented his well-balanced and carefully drafted view on the relationship of arbitral tribunals with the CJEU. On the outset the AG simply confirms well established case law along the lines of Nordsee and Eco Swiss.

'Referring to the system for reviewing the compatibility of international arbitral awards with EU law through the public policy reservation […] the Court has held that arbitral tribunals ‘constituted pursuant to an agreement’ are not courts of the Member States within the meaning of Article 267 TFEU. Consequently, they cannot refer questions for a preliminary ruling. It is therefore for the courts of the Member States, within the meaning of Article 267 TFEU, to examine, if necessary by referring a question for a preliminary ruling, the compatibility of (international or domestic) arbitral awards with EU law where an action is brought before them for annulment or enforcement, or where any other form of action or review is sought under the relevant national legislation.'

Much more powerful considerations are hidden in the footnotes. Here the AG addresses the situation of investment tribunals explicitly to which, in his view, different considerations should apply.

Footnote 34

'Based on this case-law, the arbitral tribunals hearing cases within the framework of the Convention on the Settlement of Investment Disputes Between States and Nationals of Other States (ICSID) could be regarded as being able to refer questions to the Court for a preliminary ruling. See, to that effect, [Basedow], p. 376 to 381. Since the number and size of investment arbitrations raising questions on the application of EU law are increasing, particularly in the field of State aid, the possibility for arbitral tribunals to refer questions for a preliminary ruling could help to ensure the correct and effective implementation of EU law.'

Indeed, these remarks are merely obiter dicta and the CJEU is unlikely to address any of this in the final judgment, other than—probably—following the AG on substance. The opinion, nonetheless, sends a strong signal to investment tribunals, i.e. an explicit invitation to request preliminary references from the CJEU on matters of interpretation of EU law.

Helping hand or last straw

In the best of cases AG Wathelet’s opinion would be construed as a helping hand from Luxembourg to find an amicable solution to the current conflict, based on judicial dialogue and mutual comity. However, investment tribunals have thus far refused to engage with the CJEU on questions of EU law. Or put differently, it does not appear as if investment tribunals consider EU law to be of actual relevance to arbitration. In Oostergetel and Laurentius, for instance, the tribunal acknowledged that there is “absence of any conclusive position of the [CJEU]” on the relevant issues of EU law, but subsequently rejected the respondent’s request to refer a question to the CJEU with the help of a domestic court (para. 109). The investment tribunal in Micula rejected concerns raised by the Commission to the effect that the award, if rendered, were unenforceable under EU state aid law; plainly ignoring the resulting conflict.

Gaffney suggested that a lack of guidance on questions of EU law would prompt a domestic court’s responsibility under Article 267 TFEU. However, even domestic courts might be cautious of involving requests for preliminary references in investment arbitration cases. When the award on jurisdiction in Achmea was challenged in May 2012 the Higher Regional Court of Frankfurt decided that, while EU law was raised during the arbitration, the dispute concerned in fact the interpretation of the arbitration clause in the investment agreement and as such fell outside the scope of interpretation of EU law. Ultimately, the final award was challenged before the same court in December 2014. The Frankfurt court recognized that the compatibility of arbitration clauses in intra-EU investment agreements with the Treaties is much debated but refused to refer the question to the CJEU. These cases reflect anything but excitement about the involvement of the CJEU in the arbitration process. Rather than jumping on the invitation from Luxembourg to refer questions, AG Wathelet’s opinion runs the risk of being perceived as a last straw for investment tribunals that are ultimately expected to accept the dominance of EU law and the jurisdiction of the CJEU.

Remaining challenges

However, even if investment tribunals refer questions to the CJEU in the future, a few questions still remain. First, courts or tribunals against whose decision there is no judicial remedy are not only entitled, but, in accordance with Article 267(3) TFEU, obligated to refer questions on the interpretation and the legality of EU law. Although domestic arbitration laws may provide for investment awards to be set aside, it does not prevent the award from being enforced under Article (1)(e) of the New York Convention in another state. Article 52 ICSID provides for an internal procedure for the annulment of ICSID awards on limited grounds, which effectively excludes the involvement of domestic courts. Considering, therefore, that an investment award cannot be appealed or permanently set aside on the basis of wrongful interpretation of EU law, investment tribunals might fit squarely into Article 267(3) TFEU.

The investment court, which was recently incorporated in the Comprehensive Economic and Trade Agreement with Canada (CETA) and the EU-Vietnam FTA, and which is proposed in Transatlantic Trade and Investment Partnership with the US (TTIP), raises similar concerns. Decisions of the Tribunal may be appealed before the Appeals Tribunal, inter alia, on grounds of the wrongful appreciation of domestic law (as a matter of fact). Albeit that the first instance Tribunal is relieved from any obligation under Article 267(3) TFEU, it ultimately shifts this burden onto the Appeals Tribunal. The more fundamental problem in this regard is that the EU Treaties cannot actually obligate investment tribunals to refer questions to the CJEU.

Secondly, decisions of the CJEU under the preliminary reference procedure are binding on the referring court. Without explicit safeguards in the investment agreement, however, investment tribunals are under no obligation to follow the interpretation of the CJEU (Gaffney, p. 13). There is no obvious reason why investment tribunals would refer a question to the CJEU just to subsequently ignore the answer provided. Be that as it may, these two above reservations are likely to affect the essential characteristics of Article 267 TFEU, and the powers conferred thereunder on the CJEU. According to well-established case law of the CJEU, this would adversely affect the autonomy of the EU legal order and consequently violate the Treaty (Opinion 1/09, para. 77-79). An interpretation of Article 267 TFEU that invites arbitral tribunals to refer questions but neither obligates them to do so under Article 267(3) TFEU nor renders answers of the CJEU binding on the referring investment tribunal would, thus, be incompatible with the Treaties.

Third, and perhaps most problematic, are denial of justice cases where the interpretation of domestic law might itself be the reason for an investment dispute. It would be bizarre scenario, indeed, for investment tribunals to request a preliminary ruling from the CJEU on a domestic court’s interpretation of EU law, particularly if the CJEU was involved during the domestic proceedings. Under the EU-Vietnam FTA and CETA, such a scenario could be captured by manifest arbitrariness (e.g. Article 8.10(2)(c) CETA). Gaffney points out a few other challenges such as the steadily growing influx of preliminary references that is already creating a backlog of cases, and which is likely to extend the arbitration process for several month, if not years (p. 14).


Although we are unlikely to see changes in the approach of the CJEU to commercial arbitration anytime soon, AG Wathelet’s opinion amounts to a strong endorsement of the view that investment tribunals are an entirely different story. Whether or not the preliminary reference procedure paves the way for much needed judicial comity between Luxembourg and investor-state tribunals is, however, still very much an open question. 

Barnard & Peers: chapter 10
Photo: ICSID headquarters, Washington DC
Photo credit:

Wednesday 23 March 2016

Looking for the Babel-fish? Language discrimination and EU law

Giovanni Zaccaroni, Ph.D (University of Bologna), Junior lawyer FratiniVergano 

Introduction: Language differentiation in the EU

There is no need to create language differentiation in Brussels as there was in the ancient Babel. There are already 24 official languages, according to Regulation n. 1/1958, and the EU Charter of Fundamental Rights protects the rights of EU citizens to address the EU Institutions in their own languages. The working languages mostly used within the Institutions are three: English, French and German; however, from a legal standpoint, all the documents concerning the Institutions need to be translated in all the official languages of the EU.

This was still manageable until the great enlargement of 2004, but after that it has become increasingly difficult for the Institutions to communicate in the plethora of different languages. This creates quite a complex linguistic regime which could sound odd to a person who is not confident with EU issues (surely not as much as the double seat of the European Parliament). This has rightly brought Advocate-General Kokott (in C-566/10-P) to quote the Babel-fish of the "Hitchhiker's guide to galaxy" and to refer to the aspiration coming from certain Member states of having within the EU Institutions the brand new version of this legendary beast.

So it was easily predictable that sooner or later, in the franglais-friendly environment (with some solid German stronghold) of the EU Institutions, one of the unlucky countries where these three languages are not easily spoken would have attempted to challenge the validity of an act of the Institutions on the ground that it was violating language discrimination. This was precisely what occurred when Italy and Spain appealed in front of the General Court, and then the Court of Justice, several EPSO (the inter-institutional body in charge with the selection and recruiting of a substantial part of the EU staff) competitions which appear to be waiving the linguistic regime established since the creation of the European Community by the Regulation n. 1/1958 (as successively amended), the Staff Regulations and the Charter of Fundamental Rights.

1. The facts of the cases at stake

The judgments of the Court arise from the appeal, brought by the Italian Republic, of the judgement of the General Court in Joined Cases T-166/07 and T-285/07, where the General Court dismissed an action for annulment against an EPSO competition based precisely on linguistic discrimination. The General Court held that there was no violation of the principle of non discrimination on the ground of language (as it is enshrined in the various EU legal instruments). In particular, the General Court maintained that the Commission assured an adequate publicity of the competition (while not translating the text in all official languages) and that there was no proof of the fact that the limitation of the selection of the second language to English, French and German was discriminatory.

The case was appealed in C 566/10-P on the ground that the decision of the General Court was wrong in law, as it was failing to consider that the duty not to discriminate on the ground of language was extended also to the hiring procedure of the EU Institutions. The Italian Republic won the appeal in the case, and that made easier for the General Court to rule on subsequent appeals of EPSO competitions brought by Italy and Spain in Joined Cases T-124/13 and T-191/13 (released in September 2015) and T-275/13 (released on 17 December 2015). However, (as it is testified by the reasoning of the Court in paragraphs 70 ss. in case T-166/07) the case law of the Court of Justice was quite vague, up to 2012, as regards the application of the principle of non-discrimination on the ground of language to the EU Institutions and to the specific case of EPSO competitions.

It was to a certain extent obvious that each Institution, in this case the European Commission, was free to impose some language requirements in order to assure that the members of the staff hired were able to take up immediately the duties connected with their civil servant status. However, since the entrance into force of the Amsterdam Treaty, which provided the legal basis (now Article 18 of the TFEU), inter alia, for the twinned directives 2000/43/EC (concerning race discrimination) and 2000/78/EC (concerning discrimination on four other grounds), and the entrance into force of the Charter of Fundamental Rights, the situation concerning the prohibition of discrimination has slightly changed. The fact that the Charter of Fundamental Rights is addressed primarily to the EU Institutions leaves very little discretion to the Court as to the application of the regime expressed in Regulation 1/1958 and by the Staff Regulations to the activity of the EPSO (an inter-Institutional body), and the strict rules about the burden of proof in cases of discrimination imposed on the Commission to justify each derogation from the general prohibition of discrimination of which Regulation 1/58 is an expression.

2. The equality scrutiny and the strict concept of equality

The reason why these cases are worthy of comment is that they are a very useful example of the changing attitude towards equality scrutiny in Court of Justice case law of the last few years. In particular, they are a sign of the deeper implementation at the level of judicial protection of a strict notion of equality (in which equality is presumed, unless the contrary is proved), which seems to leave apart the more traditional "relational concept" of equality (where, on the contrary, much more attention is devoted to the selection of the appropriate comparator). One would say, nothing new. The regime of anti-discrimination law in the EU after Amsterdam is built in this way: the discrimination is presumed, unless the contrary is proved. But the fact that as a general trend in anti-discrimination law there is no comparison as such and that the structure of the equality judgment switches straight to the justification regime implies that the proof become increasingly difficult.

As to the traditional understanding of the notion of equality, the structure of the judgment has always been clearly divided in two main different steps: comparison and justification. The comparison implies the choice of the appropriate comparator, and, if the situation of the people subjected to the equality scrutiny is different from the one of the comparator, then there is no discrimination. In these cases (T-124/13, T-191/13 and T-275/13, but also in C-566/10-P) there is no evidence of the fact that the Court of Justice has made the choice of selecting an appropriate comparator. It is only presumed, positively or negatively, that there is no interest of service justifying a derogatory linguistic regime for the European Institutions. On the other side, if one looks at paragraph 99 of the decision in joined cases T-166/07 and T-285/07, which was the basis of the appeal of the first case (C-566/10-P), one find the opposite approach: a more detailed analysis and an evaluation of the differences (a comparison) between the case of people speaking multiple languages and people speaking only one.

From this point of view, the switching of the balance of the equality scrutiny to the justification step means that the outcome of these cases turns upon the strict approach to justification. In this case it is undeniable that there is a Regulation and a linguistic regime applicable to the EU Institutions, which leaves very little room for derogations. However, are we really sure that equality cases should presume discrimination and go straight to the justification issue, without indulging in a comparison? This approach is surely extremely useful when it comes to a situation of discrimination in which it is difficult to unveil the discriminatory behaviour, but in term of costs-benefit approach, to use it in the inter-Institutional relationship is like using a tank to kill a bee.

3. The “interests of service” versus the presumption of discrimination and the justification regime

However, it cannot be denied that there is a problem, from the viewpoint of the application of anti-discrimination law, in the apparent inconsistency between the language regime of the EU Institutions, written in the Treaties, and the behaviour of the same Institutions. The EU Institutions apply in the recruitment procedure a golden rule, which implies (reasonably from the viewpoint of people who have ever worked inside the EU) that every single civil servant should be able to communicate effectively, other than in his or her native language, in one of the so-called “working languages” (English, German and French). However, in the EU Treaties as well as in the various pieces of secondary legislation there is no sign of these “working languages”, other than the possibility set out in Regulation 1/58 for the Institutions to approve their own linguistic regime (note that the only EU body that has approved a specific linguistic regime which has passed the Court’s scrutiny is OHIM, the EU trademark agency).

Not one of the Institutions has so far introduced such a regime, and the outcome is that the linguistic regime of the European Commission had, in the three cases examined, been scrutinised for justifications. During the justification scrutiny what should be proved is that, provided that the aim to be reached by the Institutions is legitimate, there is an evident need for the application of a specific derogatory linguistic regime, relying on several pieces of quantitative and qualitative evidence.

In fact, the presumption of discrimination which lies at the core of the actual EU anti-discrimination legislative framework, and which is applied in the three cases by the Court, leaves the proof of the absence of the discrimination entirely on the perpetrator of the discriminatory behaviour. And in case that the proof is not available, also the proof of the objective justification lies on the person, or on the Institution, which is upholding the discriminatory behaviour. From this perspective here is where the apparent weakness of the argument of the Commission in these judgments lies: the Commission thinks that the existence of an interest of service is sufficient to escape the proof of the absence of the discrimination. The Court replies that unless specific and detailed evidences are produced, which demonstrate the contrary, the “interest of service” does not allow per se the Commission to derogate from the linguistic regime. Interestingly, the Commission attempted, in particular in Case T-124/13, to provide evidence on the need for the selection of a certain number of languages within the interest of the service. The Court first, and the General Court later on, have however considered this argument insufficient, on the ground that the principle of non discrimination, of which the linguistic regime of the Institutions is an application, has a specific structural importance for the survival of the whole EU machinery.

4. Costs and benefits of the outcome of the judgments at stake

At the end of the day, what stands out of the crowd in these judgments is the fact that the EU Institutions, and the EPSO, should translate in all the official languages of the EU the text of the exams of the competitions, and that each single EU citizen can apply to the competition provided that he or she speaks at least two EU languages, no matter which languages they are. However, it is easy to doubt that an applicant who is able to find his way through the tricky EPSO competition will immediately be able to take up the duties connected to his service if he or she only speaks, for instance, Italian and Polish. And what can be said about the linguistic regime of the recruiting procedure applied by other Institutions outside the EPSO scheme?

For instance, the Court of Justice itself requires a very good knowledge of French for its positions available within the Research and Documentation DG, implying also the effective ability to write in this language. Can such a requirement resist the linguistic regime applicable to the EU Institutions and the equality scrutiny applied by the very same Court? These are all very interesting questions, but until the Institutions and the Member States effectively regulate the issue of the working languages, it is highly likely that we will see more and more judgments like these, having as a main effect to paralyze the already complex machine of the EPSO, and to render even slower the process of hiring new staff, delaying once more the crucial generational turnover within the EU Institutions.

Barnard and Peers: chapter 3
Photo credit:
Babelfish credit: Douglas Adams

Monday 21 March 2016

Libel law, EU law and the ECHR: A Comment on Arlewin v. Sweden

Athanassios Takis: PhD, Special Adviser to General Secretariat of Greek government

The judgment of the European Court of Human Rights of 1 March 2016 in the case of Arlewin v. Sweden concerns the (mis)application of EU Regulation 44/2001 (which determines which Member State’s courts have jurisdiction over civil and commercial cases) on behalf of the Swedish courts and illuminates the limits of the ‘acte clair’ doctrine (the principle that final national courts do not always have to send questions about EU law to the CJEU).

The Arlewin judgment

The facts of the Arlewin case, which gave the ECHR the chance to decide on an issue concerning the application of EU law have as follows: The applicant, Raja Arlewin, is a Swedish national, a self-employed businessman, who attempted to bring private prosecution proceedings and a claim for damages for gross defamation against X. X is a Swedish national, the Chief Executive Officer of a television company and anchor-man of a popular show in which Mr. Arlewin was accused of, among other things, involvement in organised crime in the media and advertising sectors. The television programme had been produced in Sweden, in the Swedish language and was sponsored by Swedish advertisers. The programme was sent from Sweden via satellite to a London-based company (Viasat Broadcasting UK Ltd), which broadcast and transmitted it to viewers in Sweden.

In a preliminary ruling the Stockholm District Court declined jurisdiction over Mr. Arlewin’s claims because the programme had not originated from Sweden. The court applied the relevant Swedish law, mainly the Constitutional law on freedom of expression, and followed the case law of the Swedish Supreme Court. The Court of Appeal upheld this decision, finding that Mr. Arlewin had not established that the decisions concerning the content of the programme had been taken in Sweden, a fact which was a prerequisite for the jurisdiction of Swedish courts. It contested that the defamation claim should be decided by the courts of the United Kingdom where the seat of the company transmitting the programme is.

Mr. Arlewin appealed, alleging that the Swedish courts’ position ran counter to EU law and specifically to the Brussels I Regulation (44/2001), as interpreted by the Court of Justice of the European Union (CJEU). The Regulation establishes the principle that in cases relating to liability for wrongful acts, jurisdiction is to be exercised by the courts of the place where the harmful event had occurred and, according to Mr. Arlewin, in his case the harmful effects of his defamation had occurred in Sweden. He also requested that a question concerning the interpretation of the Brussels I Regulation be referred by the national court to the CJEU for a preliminary ruling. The Swedish Supreme Court rejected Mr. Arlewin’s referral request, finding no reason to request a preliminary ruling from the CJEU, and rejected the case.

Since Mr. Arlewin was not satisfied with this, he brought an application against the Kingdom of Sweden before the ECtHR claiming that he had been deprived of effective access to court and that the State had failed to provide him with sufficient protection against allegations that violated his right to privacy. In its judgment of 1 March 2016 the ECtHR notes that the core issue in this case is whether the Swedish State violated the applicant’s rights through the courts’ decisions to dismiss the defamation suit on grounds of admissibility. The Court examined the relevance to Mr Arlewin’s case of two instruments adopted within the framework of the European Union, namely the EU Audiovisual Media Services Directive (2010/13) and the Brussels I Regulation.

The Court rejected the Government’s argument that the Audiovisual Media Services Directive determined the country of jurisdiction when an individual initiated a defamation claim. In particular the Directive. More precisely Article 28, which addresses the situation where a person’s reputation and good name have been damaged in a programme, sets out only the right of reply; it does not deal with defamation proceedings or a related claim for damages. The Court therefore considered that the Audiovisual Media Services Directive did not regulate the matter of jurisdiction when it came to defamation proceedings arising out of the content of a transborder programme service.

Rather, jurisdiction under EU law was regulated solely by the Brussels I Regulation. Under Articles 2 and 5 of that Regulation, it appears prima facie that both the United Kingdom and Sweden had jurisdiction over the subject matter of Mr. Arlewin’s case. On the one hand, X is domiciled in Sweden, and, on the other, Viasat Broadcasting UK Ltd is registered, and thus domiciled, in the United Kingdom (Article 2 gives jurisdiction to the country of domicile of the defendant). Furthermore, Article 5 gives jurisdiction to the place where the harmful event occurred, and it could be argued that the harmful event had occurred in both countries, as the television programme had been broadcast from the United Kingdom and the alleged injury to Mr. Arlewin’s reputation and privacy had manifested itself in Sweden. The programme had been produced in Sweden in the Swedish language, was backed by Swedish advertisers, was to be shown live to an exclusively Swedish audience, while the alleged harm to Mr. Arlewin’s reputation had occurred in Sweden. With the exception of the technical detail of satellite reception transmission, the facts of the case were entirely Swedish in nature.

Therefore, the Court noted that the content, production and broadcasting of the television programme as well as its implications had very strong connections to Sweden and very little to the United Kingdom. Under those circumstances the Swedish State had an obligation under Article 6 of the Convention to provide Mr. Arlewin with an effective right of access to court. The Swedish courts’ interpretation of both the Swedish and the EU law had led to a situation where the alleged victim of the defamation could not hold anyone responsible under Swedish law and could only find his way to a court in the UK. Requiring Mr. Arlewin to take proceedings before the UK courts could not be said to have been an equally viable and reasonable alternative, as far as the practical and economic obstacles to initiate proceedings abroad would render the remedy inefficient. The Swedish courts, by dismissing Mr. Arlewin’s action without examination of the merits and by referring to the alternative of initiating proceedings in another country, had impaired the very essence of his right of access to court and consequently breached Article 6 of the ECHR.


This is an ECHR judgment which directly examines the application of the EU acquis by the Member States’ courts and indirectly touches upon the issue of these courts’ obligation to apply for a preliminary ruling by the CJEU. As far as the first matter is concerned, the ECtHR clarifies the scope ratione materiae of the EU's Audio-visual Media Services Directive as not regulating every issue of substance and jurisdiction that may arise in relation to the broadcast of a television programme. The Directive only regulates a person’s right to reply to claims concerning him or her and not his or her right to initiate civil or criminal proceedings before national courts for his or her defamation.

The Court further addresses the issue of the provisions of the Brussels I Regulation and their impact on the right to a fair trial. In underlining the importance of the systematization of solutions ensured by the Regulation and the criterion of ‘strong connections’ of a dispute to a country as a jurisdictional basis (article 5 paragraph 3 of the Regulation), the judgment affirms that in general the Regulation observes the right to a fair trial and draws a fair balance between diverging interests. The special jurisdiction in matters relating to tort establishes the jurisdiction of the courts of the place where the harmful event occurred or may occur (locus damni) and constitutes a comprehensive derogation from the main rule set in Article 2. 

The ratio of this provision rests on the indissoluble ties between the component parts of liability with the evidence and of the conduct of the proceedings and the causal connection between the damage and the event from which that damage originates. That is why litigation before an alternate forum abroad would be too burdensome for the applicant and a violation of his right to enjoy practical and effective access to a court. The Court of Strasbourg relies upon the findings of the Luxembourg Court and reaffirms the existence of a direct dialogue between the two jurisdictions, with the first affirming the findings of the second in a noteworthy manifestation of its endeavour to choose -whenever possible- an interpretation of the ECHR that facilitates the proper application of EU law by national authorities.

As for the second issue at hand, namely the obligation to address a question to the CJEU, the Court doesn’t explicitly address this issue. However, it seems rather apparent that had the Swedish courts applied for a preliminary ruling, the CJEU would have affirmed that the referring courts were competent to examine the defamation claims of Mr. Arlewin by virtue of Article 5 paragraph 3 of the Brussels I Regulation. This is clearly the outcome of the CJEU judgments cited and of the paragraphs 36-39 of the ECtHR judgment concluding that ‘Brussels I Regulation requires EU Member States to make their courts available if jurisdiction is confirmed, [as] the ECJ noted in Kongress AgenturHagen GmbH v. Zeehaghe BV that the Regulation does not govern matters of procedure. This means that a court can reject a case for reasons relating to domestic procedural rules as long as the national procedural law does not impair the effectiveness of the Brussels I Regulation’.

However, the Swedish courts applied the Brussels I Regulation in a manner inconsistent with EU law and in considering the jurisdictional provisions to be an acte clair, rejected the request of the applicant. According to Article 267 of the TFEU, when questions on the interpretation or validity of EU law are raised in a case pending before a court of a Member State against whose decisions there is no judicial remedy under national law, that court is required to bring the matter before the Court of Justice of the European Union and apply for a preliminary ruling. The CJEU in the Cilfit and Others case has established three limited exceptions to this rule. Under the third of these exceptions, a national court against whose decisions there is no judicial remedy doesn’t have to apply to CJEU for a preliminary ruling when it is convinced that no doubt arises as the meaning or the validity of the law to be applied is clear and unequivocal (acte clair) and therefore, immediately comprehensible.

A court or tribunal adjudicating at last instance may take the view that, although the lower courts have interpreted a provision of EU law in a particular way, the interpretation that it proposes to give of that provision, which is different from the interpretation espoused by the lower courts, is so obvious that there is no reasonable doubt as to its meaning. The Court has also made clear that the existence of such a possibility must be assessed in the light of the specific characteristics of EU law, the particular difficulties to which the interpretation of the latter gives rise and the risk of divergences in judicial decisions within the European Union (see judgments in Intermodal Transports, para 33, and in João Filipe Ferreira da Silva e Brito and Others, para 39, and discussion here and here). 

However, the Court in the latter judgment found that, in cases where a question of EU law has given rise to a great deal of uncertainty on the part of many national courts, the courts of last instance should deem themselves obliged to make a reference to the Court of Justice. A persistent uncertainty strongly indicates not only that there are difficulties of interpretation, but also that there is a real risk of divergences in judicial decisions within the European Union. Should a failure on behalf of the national court to observe this duty imposed by Article 267 TFEU cause damages to individuals (see Köbler), the Member States should either set aside the court’s decision or, where this is impossible due to the principle of res judicata, compensate the individual for infringement enacted by the court of last instance of its EU law obligation (João Filipe Ferreira, para 60).

The relationship between the refusal to request a preliminary ruling and the violation of the right to a fair trial has been the subject of two other important judgments delivered by the ECtHR. At the same time, a third application is still pending (Repcevirág Szövetkezet v. Hungary, Application no. 70750/14 vii Application no. 17120/09). In the case of Dhahbi v. Italy the Court said that refusing a request for a preliminary ruling, while providing no reasoning at all for the refusal, constitutes a breach of Article 6 ECHR. In the more recent judgment of 21 July 2015 in the case of Schipani and others v. Italy (discussed here), the Italian Cour de cassation did not make any reference whatsoever to whether the issue at hand was an acte clair and therefore, justified an exemption from the rule that renders the preliminary reference obligatory for the last instance courts. According to the ECtHR ‘it is therefore not clear from the reasoning of the impugned judgment whether that question was considered not to be relevant or to relate to a provision which was clear or had already been interpreted by the CJEU, or whether it was simply ignored’. The Court of Strasbourg again reached the conclusion that the applicants’ right to a fair trial had been breached.

Examining the cases of Dhahbi, Schipani and Arlewin of the ECtHR and João Filipe Ferreira of the CJEU together, indicates that the Courts of Luxembourg and Strasbourg have restricted the criteria of the application of the doctrine of acte clair. The national courts have to accurately reason their decisions not to apply for a preliminary ruling based on their unequivocal understanding of the EU law. The combination of the Arlewin and João Filipe Ferreira judgments shows that a breach of EU law resulting from the misapplication by the highest national courts of Article 267 TFEU may entail the responsibility of the Member State to compensate the individuals at the same time for both their material damages (a EU law consequence in line with the Francovich and Köbler rulings) and their non-pecuniary damages (a ECtHR consequence in line with the Arlewin judgment).

Barnard & Peers: chapter 9, chapter 10

Photo credit:

Friday 18 March 2016

The final EU/Turkey refugee deal: a legal assessment

Steve Peers

The EU and Turkey have now reached an agreement on refugee issues, which has aroused considerable legal and political controversy. To examine the arguments about the deal, I present here the main text with my legal assessment of each point annotated. This builds upon my comments (together with Emanuela Roman) first of all in general on the relevant points last month, and then secondly on the leaked draft text of the final deal earlier this week (I have reused here some of the latter analysis where relevant). The agreement should be read alongside the EU summit conclusions, as well as the Commission communication on the deal. It incorporates the March 7 EU/Turkey statement which addressed the same issues in less detail.

The text of the deal is underlined below. The sections in bold have been added during negotiations, and the sections in strike-out have been removed. I have already discussed the legal status of the deal in the prior post earlier this week: it’s a statement that is not subject to approval or legal challenge as such; but its implementation in the form of specific laws or their application to individual asylum-seekers can be challenged.

1.       All new irregular migrants crossing from Turkey into Greek islands as from 20 March 2016 will be returned to Turkey. This will take place in full compliance with EU and international law, thus excluding any kind of collective expulsion. All migrants will be protected in accordance with the relevant international standards and in respect of the principle of non-refoulement. It will be a temporary and extraordinary measure which is necessary to end the human suffering and restore public order, Migrants arriving in the Greek islands will be duly registered and any application for asylum will be processed indvidually by the Greek authorities in accordance with the Asylum Procedures Directive 2013/32/EU, in cooperation with UNHCR. Migrants not applying for asylum or whose application has been found unfounded or inadmissible in accordance with the said directive will be returned to Turkey, Turkey and Greece, assisted by EU institutions and agencies, will take the necessary steps and agree any necessary bilateral arrangements, including the presence of Turkish officials on Greek islands and Greek officials in Turkey as from 20 March 2016, to ensure liaison and thereby facilitate the smooth functioning of these arrangements. The costs of the return operations of irregular migrants will be covered by the EU. Migrants having been returned to Turkey will be protected in accordance with the international standards concerning the treatment of refugees and respecting the principle of non-refoulement.

The newly added first sentence is a flagrant breach of EU and international law – but the rest of the paragraph then completely contradicts it. To be frank, anyone with a legal qualification who signed off on this first sentence should hang their head in shame. Returning ‘all’ persons who cross from Turkey to the Greek islands would contradict the ban on collective expulsion in the EU Charter and the ECHR, as well as EU asylum legislation. However, it does appear from the rest of the paragraph – including the newly added reference to non-refoulement (not sending people back to unsafe countries) – that this is not really the intention.

As for the rest of point 1, the first question is how ‘temporary’ this arrangement will be. Secondly, point 1 makes clear that the EU’s asylum procedure directive will apply to those who reach the Greek islands, as legally required. Note that the text does not refer to Greek waters: but the Directive explicitly applies to them too. It does not apply to international or Turkish waters. It is not clear what is planned as regards those intercepted before they reach the Greek islands.

As for ‘migrants not applying for asylum’ the crucial question is whether they will be given an effective opportunity to apply for asylum, as the Directive (and ECHR case law) requires. If an irregular migrant does not apply for asylum then in principle there is no legal obstacle to returning them to Turkey, subject to the conditions set out in the EU’s Returns Directive. Note that the Greek authorities will have to consider the applications, which is a significant administrative burden; this implicitly reiterates the closure of the route via the Western Balkans. The EU’s decisions on relocation of asylum-seekers from Greece and Italy (discussed here) will implicitly continue to apply, but they only commit to relocating a minority of those who arrive in Greece, and they are barely being applied in practice.  

If an application is ‘unfounded’ that means it has been rejected on the merits. If it is ‘inadmissible’ that means it has not been rejected on the merits, but on the grounds that Turkey is either a ‘first country of asylum’ or ‘safe third country’ (there are other grounds for inadmissibility, but they wouldn’t be relevant). The Commission paper briefly suggests that Turkey could be a ‘first country of asylum’ (for more analysis on that, see the prior blog post). Most of the debate is on whether Turkey is a ‘safe third country’.

Is it? The commitments on treatment in Turkey have been moved from this statement to the separate summit conclusions. Treatment in Turkey will need to match EU rules in the procedures Directive, which define a ‘safe third country’ as a country where: the people concerned do not have their life or liberty threatened on ground of ‘race, religion, nationality, membership of a particular social group or political opinion’ (this test is taken from the Geneva Convention on refugee status); there is ‘no risk of serious harm’ in the sense of the EU definition of subsidiary protection (death penalty, torture et al, civilian risk in wartime); the people concerned won’t be sent to another country which is unsafe (the non-refoulement rule, referring specifically to the Geneva Convention, plus the ban on removal to face torture et al as laid down by ECHR case law); and ‘the possibility exists to request refugee status and, if found to be a refugee, to receive protection in accordance with the Geneva Convention’.

As set out in the previous blog post, the last point is questionable because Turkey does not apply the Geneva Convention to non-Europeans, and the best interpretation of this requirement is that it must do so in order for the clause to apply. However, this interpretation is not universally shared: the Commission, the Council, Greece and some academics take the view that it is sufficient that Turkey applies equivalent standards in practice. (Note that the Commission only selectively quotes the Directive to make this argument). Even if this latter interpretation is correct, whether Turkey does apply equivalent standards in practice might itself be open to question.

Furthermore, again as discussed in the previous post, many NGOs argue that refugees are not always safe from mistreatment in Turkey itself, although no one argues that all of them are mistreated there.  Equally Turkey allegedly returns some people (but clearly not all of them) to unsafe countries, and the deal explicitly plans for a ‘safe zone’ in Syria.  Such a zone is conceivable in theory, but whether it would indeed be safe would have to be judged when and if it happens; and it may become less (or more) safe in light of events. To address these issues the procedures Directive says that the asylum-seeker must be able to argue that ‘the third country is not safe in his or her particular circumstances’. Everything will then turn on the assessment of an argument along these lines.

A critical here is whether the case can be fast-tracked. The procedures Directive contains lists of cases where the administrative procedure can be fast-tracked, and where the appeal against a negative decision to a court doesn’t automatically entitle an asylum-seeker to stay. Note that those lists don’t refer to fast-tracking ‘safe third country’ cases, although in practice it may be quicker to decide a case without examining the merits. It is possibly arguable that the lists aren’t exhaustive. If Greece wants to take this view, the interpretation of these clauses will be crucial. If the cases can’t be fast-tracked, it will obviously take longer to return people to Turkey in practice. Member States can set up special ‘border procedures’, but there is no reference to fast-tracking applications in this context. Furthermore, Member States can’t apply fast-track or border procedures to ‘vulnerable’ applicants, as broadly defined, and can’t apply border procedures to unaccompanied minors.

Odd as it might seem, the general state of human rights in Turkey (for example, as regards freedom of expression) is not directly legally relevant to returning refugees or other migrants there. The question is whether Turkey is unsafe, as defined in EU asylum law, for refugees and migrants. However, the general state of human rights in Turkey is relevant for a different reason: the Commission has separately proposed that Turkey be designated a ‘safe country of origin’, so that any refugee claims by Turkish citizens can be more easily rejected. I argued last September that this proposal was untenable in light of the human rights record of Turkey. In light of developments since, I’ll update my assessment: the suggestion is now utterly preposterous. But this proposal is not part of the deal.

2.    For every Syrian being returned to Turkey from Greek islands, another Syrian will be resettled from Turkey to the EU taking into account the UN Vulnerability Criteria. A mechanism will be established, with the assistance of the Commission, EU agencies and other Member States, as well as the UNHCR, to ensure that this principle will be implemented as from the same day the returns start. On resettlement based on 1-for-l principle: a) Priority will be given to migrants Syrians who have not previously entered or tried to enter the EU irregularly, On the EU side, resettlement under this mechanism will take place, in the first instance, by honouring the commitments taken by Member States in the conclusions of Representatives of the Governments of Member States meeting within the Council on 22/7/2015, of which 18.000 places for resettlement remain. Any further need for resettlement will be carried out through a similar voluntary arrangement up to a limit of an additional 54.000 persons. The Members of the European Council welcome the Commission's intention to propose an amendment to the within the limits and in accordance with the distribution set out in [relocation decision of 22/9/2015 to allow for any resettlement commitment undertaken in the framework of this arrangement to be offset from non-allocated places under the decision. - non-allocated places]. Should these arrangements not meet the objective of ending the irregular migration and the number of returns come close to the numbers provided for above, this mechanism will be reviewed. Should the number of returns exceed the numbers provided for above, this mechanism will be discontinued. the number of returns exceed the numbers provided for by these commitments, this agreement will be subject to review.

The idea of a ‘1-for-1’ swap of irregular migrants for resettled Syrians has been controversial, but does not raise legal issues as such. Resettlement of people who need protection from the countries they have fled to is common in practice, but is not a binding legal obligation under international or EU law. The legality of return of people to Turkey has to be judged separately (as discussed above) from the question of whatever trade-offs might be made in return for this. However, I certainly share the view of those who find a de facto ‘trade in human misery’ morally dubious. The ethos of resettlement is humanitarian; to demand a pay-off for one’s humanitarian actions contradicts their ethical foundations.

The final text makes clear that resettlement will focus on the most vulnerable people. Note that if all resettlement from now on takes place from Turkey, then no-one will be resettled by the EU from Lebanon and Jordan, which also host large numbers of Syrian refugees. On the ‘low priority’ cases, it is open to Member States to prioritise resettlement on whatever criteria they like. Obviously the intention here is to deter people from attempting unsafe journeys via smugglers; whether that would work depends on the numbers who might be resettled.

Overall, the EU has not increased the numbers of people that Member States are willing to accept: the first 18,000 are the remainder of the 23,000 people that the EU committed to resettle from non-EU countries last year, and the next 54,000 are the remainder of those who were going to be relocated from Hungary, before that state rejected the idea last September. However, unlike the mandatory quotas under the EU’s relocation decision, these numbers will be voluntary. The final deal makes clear that the maximum member of people who will be returned on this basis is 72,000: this part of the deal ends once the number of returned irregular migrants hits that number, or if the levels of irregular migration stop. In the latter case, the EU will move to a voluntary humanitarian admission scheme, discussed below. In the former case, it is not clear what will happen.

3)   Turkey will take any necessary measures to prevent new sea or land routes for illegal migration opening from up out of Turkey and into to the EU, and will cooperate with neighbouring states as well as the EU to this effect.

This refers to Bulgarian concerns that people might try to cross the Black Sea as a new entry route. Of course, if people do make to Bulgarian territory or waters, the EU asylum laws would apply, as they do for Greece.

4)   Once the irregular crossings between Turkey and the EU have come to an end are ending, or at least have been substantially and sustainably reduced, the Voluntary Humanitarian Admission Scheme will be activated. EU Member States will contribute on a voluntary basis to this scheme.

This scheme is set out in a Commission Recommendation from December, as discussed in detail here. Note that the text was amended to make clear that irregular crossings would not have to stop entirely; that was an obvious fantasy.

5) The fulfilment of the visa liberalisation roadmap will be accelerated vis-à-vis all participating Member States with a view to lifting the visa requirements for Turkish citizens at the latest by the end of June 2016, provided that all benchmarks have been met. To this end Turkey will take the necessary steps to fulfil the remaining requirements to allow the Commission to make, following the required assessment of compliance with the benchmarks, an appropriate proposal by the end of April on the basis of which the European Parliament and the Council can make a final decision.

This commitment is transposed from the March 7 statement. The waiver of short-term visas only applies to the Schengen States, and applies for stays of three months. Under the EU/Turkey readmission agreement, Turkey will have to take back anyone who overstays. It will still be necessary for Turkey to meet the relevant criteria, and for the EU Council (by qualified majority vote) and the European Parliament to approve this change in EU law.

6)   The EU, in close cooperation with and Turkey, will further speed up the disbursement of the initially allocated 3 billion euros under the Facility for Refugees in Turkey and ensure funding of additional further projects for persons under temporary protection identified with swift input from Turkey before the end of March. A first list of concrete projects for refugees, notably in the field of health, education, infrastructure, food and other living costs, that can be swiftly financed from the Facility, will be jointly identified within a week. Once these resources are about to be used to the full, and provided the above commitments are met, Furthermore, the EU will mobilise decide on additional funding for the Facility of an additional 3 billion euro up to the end of 2018. [X] billion for the period [Y] for the Turkey Refugee Facility.

The amount and timing of additional money from the EU and its Member States was agreed during negotiations. Details of the timing of disbursements and the nature of the spending projects have also been added. Note that this money is not, as is widely assumed, simply handed over to Turkey: legally speaking it can only be spent on projects that assist the Syrian refugee population. The Commission paper sets out further details of how the money will be spent, starting with a contract to provide food aid to over 700,000 Syrians.

7) The EU and Turkey welcomed the ongoing  work on the upgrading of the Customs Union.
This refers to an intention to extend the existing customs union to cover services and investment issues.
8) The EU and Turkey reconfirmed their commitment to re-energise the accession process as set out in their joint statement of 29 November 2015. They welcomed the opening of Chapter 17 on 14 December 2015 and decided, as a next step, to open Chapter 33 during the Netherlands presidency. They welcomed that the Commission will put forward a proposal to this effect in April. Preparatory work for the opening of other Chapters will continue at an accelerated pace without prejudice to Member States' positions in accordance with the existing rules.
Ultimately the EU and Turkey agreed to open only one new chapter out of 35 which need to be agreed in order for Turkey to join the EU. Only one chapter has been closed so far in a decade of negotiation. There is no commitment to open or close any further chapters. Even if an accession deal is ever negotiated, there are many legal and political obstacles in the way of it being approved, as all Member States’ parliaments would have to agree.
9) The EU and its Member States will work with Turkey in any joint endeavour to improve humanitarian conditions inside Syria, in particular in certain areas near the Turkish border which would allow for the local population and refugees to live in areas which will be more safe.
This refers to an intention (as noted above) to create a ‘safe zone’ within Syria. Whether this is viable or not remains to be seen. If there is any dispute about its safety, then returning Syrians to Turkey would be problematic if Turkey intends to send them further on to the alleged safe zone.


Overall the final deal tries to address the two main legal concerns about the March 7 ‘deal’. It makes clear that the EU asylum laws will apply to those who reach Greece (subject to the caveat about what happens to those intercepted in Greek waters), and that Turkey will have to meet the relevant standards when taking people back. The intention to ‘make the deal legal’ is clearly undermined by the extraordinary statement that ‘all’ irregular migrants will be returned. The key legal question will be how these commitments are implemented in practice.

The main legal route to challenging what happens should be by asylum-seekers through the Greek courts. Those courts could refer questions to the CJEU about EU asylum law (the CJEU could fast-track its replies). Alternatively if the asylum-seekers have gone through the entire Greek court system, or cannot effectively access the Greek system they could complain to the European Court of Human Rights (which is separate from the EU), and claim that there is a breach of the European Convention of Human Rights. In practice, however, it may be that access to lawyers and courts is more theoretical than real.

It is unfortunate, to say the least, that the EU did not try to ensure beyond doubt that the deal was legal, by putting in place some sort of effective monitoring of Turkish commitments as regards the treatment of refugees and migrants, in particular asking Turkey to fully apply the Geneva Convention to all refugees as a condition of the deal. After all, the EU will now be meeting a significant proportion of the costs of housing refugees in that country. It is even more disturbing that some Member States want to arrange for expedited returns to Libya. Surely before too long, the CJEU will asked to interpret the definition of ‘safe third country’ in EU asylum law. That finding will be crucial in determining whether it really is legal to return people to Serbia, Turkey, Libya and possibly other countries besides.

Barnard & Peers: chapter 26
JHA4: chapter I:5

Photo credit: