Wednesday 27 March 2019

Citizens of Somewhere Else? EU citizenship and loss of Member State nationality

Professor Steve Peers, University of Essex*

*This blog post is adapted in part from the forthcoming second edition of The EU Citizenship Directive (OUP, 2019), co-authored by me, Elspeth Guild and Jonathan Tomkin

Member States’ rules on the acquisition or loss of nationality are in principle a national competence. But this issue is nevertheless central to EU law, because citizenship of the EU is based on having the nationality of a Member State, according to Article 20(1) TFEU:

Citizenship of the Union is hereby established. Every person holding the nationality of a Member State shall be a citizen of the Union. Citizenship of the Union shall be additional to and not replace national citizenship.

So in light of the importance of Member State nationality to EU citizenship, can the loss of Member State nationality be reviewed for compliance with EU law? The recent judgment in Tjebbes was the latest of the CJEU’s rare opportunities to rule on this issue, and offers some important clarification of the law.


The first ruling of the Court of Justice on nationality issues was Micheletti. In this case, one Member State’s authorities were reluctant to recognize the nationality of another Member State which Mr Micheletti had acquired, given that he was also (and initially) a national of a non-EU State. The CJEU started by recognizing that ‘[u]nder international law, it is for each Member State, having due regard to Community law, to lay down the conditions for the acquisition and loss of nationality’ (emphasis added). This wording left the door slightly ajar for possible scrutiny of Member States’ nationality laws, by hinting that Community (now EU) law might place some constraints on national sovereignty on this issue.

However, it took nearly twenty years before the Court began to elaborate on such constraints. In the meantime, in the judgment in Kaur, the Court referred to the principle that Member States could define their own nationals, referring to a ‘principle of customary international law’ to this effect. According to the Court, this left the UK free to define ‘several categories of British citizens whom it has recognised as having rights which differ according to the nature of the ties connecting them to the United Kingdom’. The rights of these persons were defined in a declaration to the UK’s Treaty of Accession, which was updated later following changes in British nationality law, but that declaration did not deprive anyone of any EU law rights, for ‘[t]consequence was rather that such rights never arose in the first place for such a person’.

Subsequently, in the case of Rottmann, the Court began to set out limits to national control of the loss of nationality, as regards a person who (unlike Ms Kaur) had enjoyed the status of EU citizenship, but then was deprived of it. Mr Rottmann was an Austrian citizen and resident by birth, who later moved to Germany following the start of an investigation against him for serious fraud. He applied for and obtained German nationality, with the result that he lost Austrian nationality. But once the German authorities found out about the previous proceedings in Austria, which Mr Rottmann had not disclosed to them, they began the process of withdrawing his German nationality; and it did not appear that he met the criteria for the reacquisition of Austrian nationality either.

The CJEU began by agreeing that in line with a declaration to the TEU and the Edinburgh Decision of Member States’ heads of government of 1992, Member States had competence to determine who their nationals were. However, even if a matter falls within national competences, the exercise of those competences must have due regard to EU law. In this case, the ‘situation of’ an EU citizen who ‘is faced with a decision withdrawing his naturalisation, adopted by the authorities of one Member State, and placing him, after he has lost the nationality of another Member State that he originally possessed, in a position capable of causing him to lose the status’ of EU citizenship conferred by the Treaties ‘and the rights attaching thereto falls, by reason of its nature and its consequences, within the ambit of European Union law’.

Therefore the CJEU could rule on the ‘conditions in which a citizen of the Union may, because he loses his nationality, lose his status of citizen of the Union and thereby be deprived of the rights attaching to that status’. This ‘does not compromise the principle of international law previously recognised by the Court’ that Member States ‘have the power to lay down the conditions for the acquisition and loss of nationality’; rather it ‘enshrines the principle that, in respect of citizens of the Union, the exercise of that power, in so far as it affects the rights conferred and protected by the legal order of the Union’, such as ‘in particular’ a withdrawal of naturalization as in the Rottmann case, ‘is amenable to judicial review carried out in the light of [EU] law’.

In Mr Rottmann’s case, a withdrawal of nationality due to deception could be compatible with EU law, since it was in the ‘public interest’ and ‘it is legitimate for a Member State to wish to protect the special relationship of solidarity and good faith between it and its nationals and also the reciprocity of rights and duties, which form the bedrock of the bond of nationality’. Withdrawal of nationality on such grounds was provided for by international law, in particular the 1961 Convention on the Reduction of Statelessness and the Council of Europe’s European Convention on Nationality. It was therefore up to the national court to determine the proportionality of the decision in light of national and EU law. On this point, the national court had to ‘take into account the consequences that the decision entails for the person concerned and, if relevant, for the members of his family’ as regards the loss of EU citizenship rights, assessing ‘in particular, whether that loss is justified in relation to the gravity of the offence committed by that person, to the lapse of time between the naturalisation decision and the withdrawal decision and to whether it is possible for that person to recover his original nationality’. While EU law did not ban the withdrawal of nationality before the person concerned obtained again his original nationality – ie there was no apparent ban on making the person concerned stateless – the national court had to consider whether ‘the principle of proportionality requires the person concerned to be afforded a reasonable period of time in order to try to recover the nationality of his Member State of origin’. Finally, the Court made clear that ‘the principles stemming from this judgment’ as regards the effect of EU law on national powers in the area of nationality law ‘apply both to the Member State of naturalisation and to the Member State of the original nationality’.

What is the impact of the Rottmann judgment? First of all, as for the scope of the judgment, it did not seem to be relevant in this case that Mr Rottmann had previously exercised free movement rights; the Court referred to the loss of EU citizenship status in principle, not to the particular impact upon people who had moved within the EU pursuant to the citizens’ Directive or other provisions of EU law.

Secondly, what substantive rules apply? The Court accepts the loss of citizenship can be justified on ‘public interest’ grounds such as deception, subject to the application of the principle of proportionality. However, it did not indicate how the public interest should be weighted as compared to the individual interest in retaining EU citizenship in this particular case. While the Court confirmed its analysis in Rottmann by reference to the Council of Europe Convention on Nationality and the 1961 Convention on the Reduction of Statelessness, it should be noted that the former Convention has been ratified by fewer than half of the Member States, and only a slim majority of Member States have ratified the latter. In fact, ten Member States have not ratified either treaty.

Finally, procedurally, on the facts of Rottman there had to be a ‘reasonable period of time’ to pursue the re-acquisition of the original nationality, as an aspect of the principle of proportionality. However, while the Court mentioned judicial review there were no further details besides a reference to the national court’s obligations to apply the principle of proportionality, and no specific mention of administrative procedural rights.

The latest judgment

Tjebbes concerns four different applicants: a dual citizen (since birth) of the Netherlands and Canada; a Dutch citizen from birth who acquired Swiss nationality via marriage; her daughter, a dual Dutch and Swiss citizen from birth, who was listed on her mother’s passport when a child; and an Iranian national from birth who acquired Dutch nationality.  Each of them, when residing outside the EU, applied for a Dutch passport, but the authorities ruled that they had lost Dutch nationality automatically due to residence outside the EU. This resulted from Dutch law which provides for such loss after ten years’ residence abroad, if the person concerned has the nationality of another State. That loss could be prevented by residing in the EU for a year before the nationality was lost, or by obtaining a declaration regarding the possession of Netherlands nationality or a travel document or Netherlands identity card. There are also special rules on minors, entailing loss of Dutch nationality to correspond with the parent’s loss of that status.

The national court had doubts about the notion of a general rule on loss of nationality (as distinct from a decision to strip a person of nationality in an individual case, as in Rottman), as well as the specific issues relating to loss of nationality by children, and so asked the CJEU to rule on whether the Dutch rules are compatible with EU law.  The Court’s analysis began by specifically pointing out that none of the persons concerned had moved within the EU. So its judgment is concerned with EU citizenship as such, rather than loss of free movement rights.

Reiterating the basic points of Rottmann (loss of nationality has to be consistent with EU law, but a Member State can legitimately ‘protect the special relationship of solidarity and good faith between it and its nationals and also the reciprocity of rights and duties, which form the bedrock of the bond of nationality’, the Court accepted that in general:

when exercising its competence to lay down the conditions for acquisition and loss of nationality, it is legitimate for a Member State to take the view that nationality is the expression of a genuine link between it and its nationals, and therefore to prescribe that the absence, or the loss, of any such genuine link entails the loss of nationality. It is also legitimate for a Member State to wish to protect the unity of nationality within the same family.

Applying that principle to the facts of this case, the ten-year absence criterion in Dutch law ‘may be regarded as an indication that there is no such link’. As for children, ‘the lack of a genuine link between the parents of a child who is a minor and the Kingdom of the Netherlands can be understood, in principle, as a lack of a genuine link between the child and that Member State’.

This interpretation was supported (as in Rottmann) by the Convention on the Reduction of Statelessness which provides for the possible loss of nationality ‘in similar situations’ as long as the person concerned ‘does not become stateless’. The latter risk was ruled out by Dutch law. Also as in Rottmann, the Council of Europe Convention on Nationality is relevant: providing that nationality can be lost in the event of ‘no genuine link between that State and a national habitually residing abroad and, in the case of a minor, for children whose parents lose the nationality of that State’.

The applicable safeguard (a request within the 10-year period for the issuing of a declaration regarding the possession of Netherlands nationality, a travel document or a Netherlands identity card, interrupting the 10-year period) was also relevant.

However, again as in Rottmann, national authorities and national courts had to assess whether the loss of the nationality of the Member State (and therefore EU citizenship) respects the principle of proportionality, as regards ‘the consequences of that loss for the situation of the person concerned and, if relevant, for that of the members of his or her family, from the point of view of EU law’. In particular, loss of nationality ‘by operation of law  would be inconsistent with the principle of proportionality’ if there was no possibility for ‘at any time an individual examination of the consequences of that loss for the persons concerned from the point of view of EU law’, which could, ‘where appropriate’, mean that the person concerned could ‘recover his or her nationality ex tunc in the context of an application by that person for a travel document or any other document showing his or her nationality’. In practice, Dutch law includes this safeguard. It must entail:

an individual assessment of the situation of the person concerned and that of his or her family in order to determine whether the consequences of losing the nationality of the Member State concerned, when it entails the loss of his or her citizenship of the Union, might, with regard to the objective pursued by the national legislature, disproportionately affect the normal development of his or her family and professional life from the point of view of EU law. Those consequences cannot be hypothetical or merely a possibility.

The examination of proportionality must ensure that the national authorities and national courts assess whether the loss of nationality is consistent with the EU Charter of Rights, particularly the right to family life (Article 7 of the Charter) and the obligation to take into consideration the best interests of the child (Article 24). As regards the individual circumstances to be considered, it’s likely to be relevant that the persons concerned:

would be exposed to limitations when exercising his or her right to move and reside freely within the territory of the Member States, including, depending on the circumstances, particular difficulties in continuing to travel to the Netherlands or to another Member State in order to retain genuine and regular links with members of his or her family, to pursue his or her professional activity or to undertake the necessary steps to pursue that activity.

It’s also relevant that the person concerned ‘might not have been able to renounce the nationality of’ a non-EU country, and whether there’s a ‘serious risk, to which the person concerned would be exposed, that his or her safety or freedom to come and go would substantially deteriorate because of the impossibility for that person to enjoy consular protection under Article 20(2)(c) TFEU in the territory of the third country in which that person resides’.

There are particular safeguards for minors: the administrative and judicial authorities must ‘take into account, in the context of their individual examination’, the possibility that loss of nationality resulting from the loss of nationality by a parent ‘fails to meet the child’s best interests as enshrined in Article 24 of the Charter because of the consequences of that loss for the minor from the point of view of EU law.’


In many respects, Tjebbes confirms the Court’s ruling in Rottmann: national competence to regulate loss of nationality (and therefore EU citizenship) subject to minimum standards developed by the Court due to the importance of EU citizenship; taking account of international treaties as a source of those minimum standards; broad deference to national determination as to what the grounds of loss of nationality should be; a proportionality test which must take account of individual circumstances, including those of family members; and the requirement of judicial control.

However, a number of points are further developed or new. First of all, it’s now clear beyond any doubt that loss of nationality of a Member State raises issues because of the consequential loss of EU citizenship regardless of whether the persons concerned have exercised free movement rights. Secondly, in Tjebbes the Court seems more concerned to avoid the persons concerned becoming stateless.

Thirdly, the Court confirms that the issues which Member States may consider legitimate are not just the public interest concerns arising from the deception in Rottmann, but also the simple lack of residence within the EU for a sufficient period of time. (The Court does not comment on the hypothetical prospect of loss of nationality due to the exercise of free movement rights within the EU, but this would obviously be highly problematic in that it would penalise the use of free movement rights, and therefore deter their exercise. The Court does not comment on the possible relevance of the EU/Swiss free movement treaty as regards one of the persons concerned). It was arguably already implicit from Rottmann that Member States can sanction even more serious offences such as terrorism with a loss of nationality (and therefore EU citizenship).

Fourthly, the Court accepts that nationality can be lost by means of a general rule, not just as a consequence of an individual decision. However, it considers it essential that the general rule on loss be complemented by a general safeguard – in this case, the possibility of forestalling the loss of nationality via an application for an identity document during the ten-year period. Fifthly, in both cases, the Court places strong emphasis upon a review of the proportionality of the loss of nationality, although the standards applicable are necessarily partly different. In both cases, the individual consequences must be assessed, but instead of ‘whether that loss is justified in relation to the gravity of the offence committed by that person’ (Rottmann), the test in Tjebbes is a disproportionate effect on ‘the normal development of his or her family and professional life from the point of view of EU law’, although the argument on this issue cannot be purely hypothetical.

Furthermore, the Court requires the assessment to consider limitations on the exercise of free movement rights and consular protection (which are surely inevitable, except as regards Swiss citizens – although again, the Court does not mention the EU/Swiss treaty) including difficulties in travel to the EU (possibly a bigger issue if a visa requirement would be imposed, as in the case of Iran as distinct from Canada or Switzerland) as regards family links or professional activity. For minors, it is hard to see how the best interests of the child are served by the child’s loss of EU citizenship despite any decision by that child to leave the EU as an adult. Sixthly, the Court emphasises not only judicial control, but also the obligations of the administrative authorities. Finally, the Court places significant stress on the EU Charter of Rights, which it did not mention in Rottmann.

However, there are still a number of uncertain issues. The Court’s degree of concern about the risk of statelessness seems to vary from the relative indifference in Rottmann to the implied safeguard in Tjebbes.  This leaves the important question of the degree of certainty of obtaining the other country’s nationality (an issue in UK case law and disputes) undecided. Also, it’s still unclear how relevant the two international treaties the Court refers to in both cases might be – given that a number of Member States have not ratified one or both of them.   One wonders also if various limits on judicial review or stemming from administrative procedure might be challenged for breach of effectiveness.

Of course, there’s a limit to how much the case law on this issue can develop if the Court keeps receiving about one case per decade. It’s possible that the Court will soon get more though, if the UK soon becomes a non-EU State. This raises a variety of issues, most obviously whether UK nationals – who were citizens of the EU until Brexit Day – lose EU citizenship on that date (unless they hold the nationality of a remaining Member State)? The orthodox view is that they do, on the basis of the Treaty wording: ‘Every national of a Member State shall be a citizen of the Union.’ The obvious implication is that when your State of nationality ceases to be a Member State, you are no longer a citizens of the Union. This interpretation is reflected in the wording of the draft withdrawal agreement, which distinguishes between UK nationals and EU citizens, defining the latter as citizens of Member States but not listing the UK as a Member State. It also is implicit in the case law of the Court of Justice on loss of EU citizenship, which is based on the assumption that EU citizenship is lost when a person loses the nationality of a Member State. This is arguably a quite different issue from a country ceasing to be a Member State.

The counter-argument is that the Treaty only sets out the means of acquiring EU citizenship, not losing it. It follows that Brexit cannot remove EU citizenship from those UK nationals who already had it. A Dutch court initially decided to ask the CJEU to clarify these issues (see discussion here), but an appeal court decided to block the case. The question might at some point be resolved by the CJEU one way or the other. 

Brexit will also raise other issues. In Member States that remove their nationality from citizens resident outside the EU, the UK’s non-EU status will bring more of their citizens within the scope of this rule. At the very least, should time spent in the UK when it was a Member State be ignored for the purpose of this rule? Or should the rule only apply to nationals who move to the UK after Brexit? This raises the broader question of whether time spent in the UK before Brexit day continues to have effect for the purposes of EU citizenship and free movement law (ie for those EU27 citizens who married a non-EU citizen in the UK and then returned to their State of nationality).

More generally, given the importance of nationality to the essence of national sovereignty, the absence of power for the EU to harmonise laws in this field, the TEU declaration and the Edinburgh decision, there’s a strong case that the Court’s case law interfering with the national rules on the loss of national citizenship is unjustifiable in principle. However, this is vitiated by the great deference which the Court gives to Member States on this issue in its case law. Ultimately, the importance of this case law may be as an essential safeguard against any drift towards authoritarianism in a Member State which might want to strip government critics of their nationality for defying the ‘will of the people’. The problem is that any government already inclined to flout the rule of law so flagrantly may well ignore the CJEU’s rulings too. But this is a paradox to explore another day.

Barnard & Peers: chapter 13
Photo credit:

Tuesday 26 March 2019

Guardianship, free movement and the rights of the child: the SM judgment

Professor Steve Peers, University of Essex

*This blog post builds on research which contributed to the forthcoming second edition of The EU Citizenship Directive (OUP, 2019), co-authored by me, Elspeth Guild and Jonathan Tomkin

When is a child a ‘family member’? Many people regard others they love dearly as children, parents or other relatives, but the law is rather stricter. This strictness is particularly important where children are involved, in order to ensure their welfare, and where it impacts on immigration law. Moreover, different countries have different approaches to the legal definition of family members. All these issues come to a head in today’s judgment of the CJEU in SM, a case concerning the intersection between EU free movement law and the family law of non-EU countries.

EU law background

The EU’s citizens’ Directive sets out the main rights of EU citizens and their family members to move to other EU countries. It defines family members as including, in Article 2(2)(c):  

the direct descendants who are under the age of 21 or are dependants and those of the spouse or partner as defined in point (b)

It also refers separately to ‘beneficiaries’ of the Directive, in Article 3(2):

2.      Without prejudice to any right to free movement and residence the persons concerned may have in their own right, the host Member State shall, in accordance with its national legislation, facilitate entry and residence for the following persons:

(a)      any other family members, irrespective of their nationality, not falling under the definition in point 2 of Article 2 who, in the country from which they have come, are dependants or members of the household of the Union citizen having the primary right of residence, or where serious health grounds strictly require the personal care of the family member by the Union citizen;

(b) the partner with whom the Union citizen has a durable relationship, duly attested.

The host Member State shall undertake an extensive examination of the personal circumstances and shall justify any denial of entry or residence to these people.

The people listed in Article 3(2) are referred to in UK law implementing the Directive as ‘extended family members’, and I will use that term for brevity’s sake. In contrast, I’ll refer to the Article 2 family members, including ‘direct descendants’, as ‘core’ family members.

CJEU case law has touched on the definition of core family members several times, for instance clarifying the concept of ‘dependants’ in Article 2(2)(c): see most recently the CJEU judgment in Reyes, discussed here). But the Court had never been asked to clarify the meaning of ‘direct descendants’, although it had ruled when interpreting previous free movement legislation that children include an EU citizen’s step-children (Baumbast).

As for extended family members, previous CJEU case law (Rahman and Banger) clarified that they have in principle a less far-reaching right than core family members. While core family members have a right to enter and reside, for extended family members ‘entry and residence has only to be facilitated by that Member State’. So the Directive ‘does not oblige the Member States to grant every application for entry and residence submitted by persons who show’ that they fall within the scope of Article 3(2). However, the Court said it was ‘clear from the use of the words “shall facilitate” that that provision imposes an obligation on the Member States to confer a certain advantage, compared with applications for entry and residence’ by other third-country nationals, on applications by extended family members. This obligation meant that Member States had to ‘make it possible’ for the extended family members to ‘obtain a decision on their application that is founded on an extensive examination of their personal circumstances and, in the event of refusal, is justified by reasons’. When examining such an application, the authority concerned had to ‘take into account of the various factors that may be relevant in the particular case, such as’ the factors specifically listed in the preamble to the Directive – namely ‘their relationship with the Union citizen or any other circumstances, such as their financial or physical dependence on the Union citizen.’

Moreover, in light of the reference to national legislation in Article 3(2), and in the absence of more specific rules in the Directive, the Court ruled that each Member State ‘has a wide discretion as regards the selection of factors to be taken into account’. But there is a core substantive obligation for Member States: their legislation must ‘contain criteria which are consistent with the normal meaning of the term “facilitate” ’—presumably as defined by the Court—‘and of the words relating to dependence used in Article 3(2), and which do not deprive that provision of its effectiveness’. Finally, the Court conceded that Article 3(2) was not directly effective, although ‘an applicant is entitled to a judicial review of whether the national legislation and its application have remained within the limits set by that Directive’. The subsequent judgment in Banger elaborated upon these procedural rights.

Background to the case

According to the UK Supreme Court judgment which referred questions to the CJEU, this case concerns a French couple married in the UK, who were married in 2001 but were unable to have children of their own. They went to Algeria and applied for the guardianship of a child by means of application of the kafala family law system (the term is also used to refer to a controversial system of migrant worker sponsorship, but this case only concerns family law). Under the Islamic approach to family law, adoption as such is not permitted, but kafala instead provides for a form of guardianship, and is particularly applied for the benefit of orphaned or abandoned children.

The Algerian authorities deemed the French couple suitable guardians, and they were soon accorded guardianship of a girl abandoned at birth (called ‘Susana’ by the Supreme Court: I’ll use that name rather than the impersonal ‘SM’). The husband returned to the UK to work, while the wife stayed in Algeria to look after Susana, seeking to obtain entry clearance for her. The UK authorities refused on the grounds that this was not an adoption as such, and so a long dispute through the legal system got underway.

At first instance, the immigration tribunal ruled in favour of the UK authorities, holding that Susana was neither a core family member or extended family member. On appeal, the Upper Tier tribunal ruled that she was an extended family member. The authorities appealed in turn to the Court of Appeal, which ruled that she was neither. The Supreme Court gave leave to appeal on the ‘extended family member’ point, but subsequently decided to consider both points. Ultimately, Lady Hale’s judgment (with which the other judges agreed) concluded that it was obvious that Susana was an extended family member, but decided to ask the CJEU whether she was a core family member.

Judgment of the CJEU

First of all, the Court noted that the Directive ‘makes no express reference to the law of the Member States for the purpose of determining the meaning and scope of’ the term ‘direct descendant’. Next, it followed (in line with the Court’s usual approach) that ‘the need for a uniform application of EU law and the principle of equality require that the terms of that provision must normally be given an independent and uniform interpretation throughout the European Union’. In the absence of any definition of ‘direct descendant’, the term should be interpreted considering ‘not only its wording, but also the context in which it occurs and the objectives pursued by the rules of which it is part’.

In the Court’s view:

the concept of a ‘direct descendant’ commonly refers to the existence of a direct parent-child relationship connecting the person concerned with another person. Where there is no parent-child relationship between the citizen of the Union and the child concerned, that child cannot be described as a ‘direct descendant’ of that citizen for the purposes of Directive 2004/38.

Elaborating further, ‘[a]lthough that concept primarily focuses on the existence of a biological parent-child relationship,’ the aim of the citizens’ Directive is to facilitate free movement rights, and so therefore it must be ‘construed broadly’, including the core family member definition. That meant ‘it covers any parent-child relationship, whether biological or legal’, including adopted children. However, it did not extend to a guardianship system which did not create a legal parent-child relationship.

However, the Court went on to confirm that Susana fell within the concept of an extended family member, as the relevant definition in the Directive is ‘capable of covering the situation of a child who has been placed with citizens of the Union under a legal guardianship system such as Algerian kafala and in respect of whom those citizens assume responsibility for its care, education and protection, in accordance with an undertaking entered into on the basis of the law of the child’s country of origin’. The Court reiterated Member States’ discretion in such cases as referred to in previous cases, but considerably narrowed the discretion on the facts of this case.

This was because of the right to family life in Article 7 of the EU Charter of Rights, which has the ‘same meaning and scope’ of the corresponding right in the European Convention on Human Rights. Case law of the European Court of Human Rights had confirmed that the child-guardian relationship under the kafala system ‘may fall under the definition of family life, having regard to the time spent living together, the quality of the relationship, and the role which the adult assumes in respect of the child’ (referring to Chbihi Loudoudi and Others v. Belgium). Article 8 ECHR ‘protects the individual against arbitrary action by the public authorities and requires those authorities, where the existence of a family tie has been established, to enable that tie to be developed and to establish legal safeguards that render possible the child’s integration in his family’ (referring to Harroudj v. France and Chbihi Loudoudi).

Furthermore, Article 24 of the Charter obliges Member States to ‘take into consideration the best interests of the child’. It followed that when exercising their discretion as regards admission of extended family members, national authorities have to make ‘a balanced and reasonable assessment of all the current and relevant circumstances of the case, taking account of all the interests in play and, in particular, of the best interests of the child concerned’. This assessment has to ‘take into consideration, inter alia, the age at which the child was placed under the Algerian kafala system, whether the child has lived with its guardians since its placement under that system, the closeness of the personal relationship which has developed between the child and its guardians and the extent to which the child is dependent on its guardians, inasmuch as they assume parental responsibility and legal and financial responsibility for the child.’

Conversely, the assessment has to ‘take account of possible tangible and personal risks that the child concerned will be the victim of abuse, exploitation or trafficking’. But these risks can’t simply ‘be assumed’ just because the process of the Algerian kafala system ‘is less extensive’ than the adoption procedure in the State where the EU citizens are living, or because Algeria has not signed up to the 1996 Hague Convention on parental responsibility. Rather, it was necessary to weigh such factors up against the other factors which the Court set out.

If this assessment establishes that

the child placed under the Algerian kafala system and its guardians, who are citizens of the Union, are called to lead a genuine family life and that that child is dependent on its guardians, the requirements relating to the fundamental right to respect for family life, combined with the obligation to take account of the best interests of the child, demand, in principle, that that child be granted a right of entry and residence as one of the other family members of the citizens of the Union for the purposes of Article 3(2)(a) of [the citizens’ Directive], read in the light of Article 7 and Article 24(2) of the Charter, in order to enable the child to live with its guardians in their host Member State.

This analysis ‘applies a fortiori’ where the refusal to admit the child meant that one of the child’s guardians is in practice prevented from living together in a Member State because one of them needs to stay in a non-EU State to care for the child. As noted above, these were the facts in this case.

Finally, the Court declined to answer the Supreme Court’s separate question about refusing to admit a child if there was a risk of abuse, exploitation or trafficking, because the question was only raised in the event that Susana was considered a core family member. In any event, it should be noted that the Court referred to such issues being considered as part of the assessment of the application to admit Susana as an extended family member.


The Court’s confirmation that adopted children are covered by the concept of ‘direct descendant’ is not surprising, although it raises the question as to what happens if there are divergences between the adoption rules in the State where the child was adopted and the Member State in which the child lives now. Then again, the previous case law applying free movement law to cover step-children already raised questions (not yet answered) about the interaction between family law and free movement law (ie, who has custody of the step-child), on top of the long-standing case law about the impact of divorce and separation upon free movement rights (see most recently the CJEU’s NA judgment of 2016, which I discussed critically here). One quite striking feature of the Court’s ruling is its apparent assertion that only a parent-child relationship could fall within the concept of ‘direct descendant’: this is, with respect, not consistent with the obvious meaning of the words, as evidenced by the Supreme Court’s assumption that it could also cover grand-children.

The ruling that the term ‘direct descendants’ could not apply to guardianship followed from the Court's definition of ‘direct descendants’ to mean a parent-child relationship, given that the kafala system deliberately does not classify the child’s guardians as parents. Despite that, the French couple in this case understandably sought to bring Susana within the scope of the ‘direct descendants’ concept, because they believed that otherwise the UK would retain residual discretion not to admit her as an extended family member.

This brings us to the most remarkable aspect of the Court’s ruling: the removal of the Member State’s discretion to admit Susana, on condition that the assessment of the factors was positive. Susana’s case appears to fall within the scope of the factors the Court refers to as part of that assessment (the age at which the child was placed under the Algerian kafala system, whether the child has lived with its guardians since its placement under that system, the closeness of the personal relationship with the guardians, the degree of dependency, and the assumption of responsibility for the child). It’s impossible from the outside of this case to know if the risks of abuse or trafficking exist, although of course in the best interests of the child the Court was right to confirm that the assessment should examine whether such factors are present. In other cases with different circumstances, the factors may point to a different conclusion than they do in this case. And today’s judgment raises the interesting question of whether there might be reduced discretion to refuse entry in other extended family cases too – notably cases involving children or partners, and where the impact of refusal might similarly be to prevent free movement rights from being exercised.

Finally, it’s worth putting this case in the broader context. For eight-year old Susana, abandoned at birth by her parents, her guardians have been litigating nearly her entire life to secure her right to stay with them in the UK. It took six years of litigation before four levels of the judicial system before they could confirm that the UK was even obliged to consider her application.  Now the CJEU rules that (depending on the assessment) the UK may be obliged to let her stay. But that’s mainly because of the EU Charter of Rights; and this may be the last judgment concerning the UK and the Charter. If you seek its monument, look at this little girl’s case.   

Barnard & Peers: chapter 13
Photo credit:

Monday 25 March 2019

Accountability and Independence of the Governors of National Banks: Any role for the Court of Justice of the European Union?

Dr Marios Costa, Senior Lecturer, City, University of London

The European Union (EU) has long been criticised for administrative inadequacies and for structural deficiencies. There have been a number of reports and commentaries highlighting that the EU suffers from political irregularities. Similarly, alleged corruption, maladministration and money laundering at the national level of governance, as shown below, is of equal concern for the citizenry and the Union alike.

On 26 February 2019, the Court of Justice (Grand Chamber) gave a significant judgment on two joint cases brought by the suspended Governor of the Central Bank of Latvia, Ilmārs Rimšēvičs, and the European Central Bank (ECB) against the Republic of Latvia. Mr Rimšēvičs and the ECB argued that the Latvian Anti-corruption Office has unlawfully prohibited him from carrying out his duties as the Governor of the Central Bank which included participation in the Governing Council of the ECB. Rather unexpectedly, the Court of Justice annulled, for the very first time in the history of EU law, the national act from the Anti-corruption Office which restricted Mr Rimšēvičs from exercising his duties.

This judgment raises broader constitutional ramifications. It is therefore necessary to examine whether the Court has now gone beyond the jurisdiction set out in the Treaty framework. With all due respect, the ruling comes as a big surprise. This commentary examines the appropriateness of the recent judgment and concludes that the Court of Justice has stretched its powers of judicial review unprecedentedly. Yet, it can be set at the outset that the Court’s ruling is exceptional and closely related to the EU’s monetary regime. It remains to be seen whether this rather extraordinary case will mark the beginning of a new judicial trend with extended jurisdiction well beyond the EU legal order. 

Facts of the Case

Mr Rimšēvičs, the Governor of the National Bank of Latvia, was accused of soliciting bribery in the form of a free leisure trip as well as accepting the amount of EUR 750 000 in exchange of exercising influence in favour of a private Latvian Bank. The Latvian Anti-Fraud Office initiated investigations into the serious bribery allegations which resulted to the imprisonment of Mr Rimšēvičs. On 19 October 2018, he was released following a prohibition on performing decision-making, control and monitoring duties within the Central Bank of Latvia.

Mr Rimšēvičs and the ECB challenged the legality of the decision to relieve him from the office before the Luxembourg Court. The Court annulled the decision of the Latvian Anti-corruption Office to the extent that it has prevented the Governor of the Central Bank to exercise his EU (and national) duties.

EU Legal Framework on the Governors’ Accountability

EU law is not silent on the issue of the Governors’ accountability. Article 14(2) of the Statute of the European System of Central Banks (ESCB) and of the ECB, entitled ‘National central banks’, provides:

A Governor may be relieved from office only if he no longer fulfils the conditions required for the performance of his duties or if he has been guilty of serious misconduct. A decision to this effect may be referred to the Court of Justice by the Governor concerned or the Governing Council on grounds of infringement of these Treaties or of any rule of law relating to their application. Such proceedings shall be instituted within two months of the publication of the decision or of its notification to the plaintiff or, in the absence thereof, of the day on which it came to the knowledge of the latter, as the case may be.

Additionally, Article 130 TFEU provides:

When exercising the powers and carrying out the tasks and duties conferred upon them by the Treaties and the Statute of the ESCB and of the ECB, neither the European Central Bank, nor a national central bank, nor any member of their decision-making bodies shall seek or take instructions from Union institutions, bodies, offices or agencies, from any government of a Member State or from any other body. The Union institutions, bodies, offices or agencies and the governments of the Member States undertake to respect this principle and not to seek to influence the members of the decision-making bodies of the European Central Bank or of the national central banks in the performance of their tasks.

Judicial Review of EU Law acts

The Treaty on the Functioning of the European Union (TFEU) provides for two different methods of judicial control designed to ensure the legal exercise of power by EU institutions, offices, bodies and agencies. The relevant provisions are Articles 263, concerning direct actions for annulment, and 267, concerning indirect review via the preliminary reference procedure from the national courts. Overall, the EU system of judicial review reflects the fundamental principles of subsidiarity provided in Articles 4 and 5 TEU. Consequently, the annulment of a national act falls within the exclusive competence of the Member States and the CJEU has jurisdiction to annul a national measure only where there is an explicit power to do so in the Treaties. Rimšēvičs is therefore a unique case as it is relates to the annulment of a national measure by which the Governor of the national bank of Latvia was “relieved from office”. It becomes pressing to examine the appropriateness of the Court’s ruling and to assess whether EU law explicitly empowers the CJEU to annul the national measure adopted against the central banker.

The findings of the Court

The CJEU ruled that it has jurisdiction to annul a national measure so long as it suspends the Governor of the national bank. In doing so, the Court interpreted that “both the literal and the systemic and teleological interpretations of Article 14(2) of the statute entail the action provided for in that Article being classified as an action for annulment” (para 66). The Court went even further to explain that the statute of the ESCB derogated from the usual distribution of judicial review powers between the national court and the EU courts. The justification, according to the Court, was that the “ESCB represents a novel legal construct in EU law which brings together national institutions, namely the national central banks, and an EU institution, namely the ECB, and causes them to cooperate closely with each other, and within which a different structure and a less marked distinction between the EU legal order and national legal orders prevails.” (Para 69).  

Comment and Analysis

In Rimšēvičs the Court clarified the abovementioned provisions as regards independence and accountability of the Governors of the National Banks. Surely, any failure by an individual Governor to meet the standards described in Article 14 (2) ESCB, as set out above, can lead to a significant damage of the public image of the ECB and consequently cause a significant damage to financial stability in the EU. One can understand that the concerns and commitment to high standards exercised by the Latvian Anti-corruption Office are perfectly legitimate.

Yet the Latvian authorities were asked in a number of instances by the Court to support their serious allegations with evidence. The failure to produce any evidence supporting the suspension of Mr Rimšēvičs from office is remarkable. The lack of evidence against the serious background of alleged bribery and money laundering is related, closely, to the fact that individual Governors need to operate impartially and independently, without influence and pressure from external sources, whether national governments or private individuals. The Court protected the independence of the ECB and its Governing Council and emphasised, rightly, that under EU law any form of pressure cannot be accepted. Overall, the Court highlighted that independence should be protected under any circumstances in order for the Governor and the ECB to adopt and implement their decisions based upon technical and up-to-date scientific expertise. They adopt critical monetary related decisions that will have little or no use if they are subjected and influenced to any pressure.

The Governors must meet the highest possible standards and should perform their duties without any external influence due to their high ranking. In particular, as already explained above, Article 14 (2) ESCB requires Governors to be free from any external influence. This point is vital if the ECB is to stay independent of Member States or individuals. But let’s assume for a while that there was enough evidence that Mr Rimšēvičs obtained pecuniary advantages from the Latvian private bank. Assume further that the Latvian Anti-corruption Office concluded that the Governor needs to be held accountable for infringing his duty to behave with integrity and avoid maladministration. Is the national procedure that relieved him from the office in accordance with EU requirements to respect the rule of law? Or, does it represent an abusive behaviour exercised by the Latvian executive authorities? Relieving the national Governor whose independence is protected under EU law without evidence and without given the opportunity to see and respond to concrete evidence that supports the serious allegations constitutes a manifest violation of the notion of independence which is clearly safeguarded by the ESCB statute. Surely this is not something that can be accepted or justified. Anything that compromises the independence of the national central bankers is illegal under EU law.

Pursuant to the EU Treaties, clearly the Court of Justice has jurisdiction to decide on cases related to the accountability and independence of the ECB. Yet, the interpretation of the Court to extend its power to annul a national decision is surprising, at least. Taking into account the absence of any evidence and also the factual background of the case, one can realise that a number of irregularities by the Latvian authorities took place. The Court has made the right decision in clarifying the set EU law requirements that shield the ECB and the ESCB from any pressure. Additionally, and perhaps most importantly, the judgment provides us with clear boundaries on how to safeguard the independence of the national Governors. Independence has been a key factor in deciding Rimšēvičs, a factor which came with the cost to strike down a national decision by the CJEU in order to safeguard it.

Barnard & Peers: chapter 10, chapter 19
Photo credit: New Europe

Thursday 21 March 2019

Brexit and Extending EU Membership: The Legal Issues

Professor Steve Peers, University of Essex*

*This blog post was updated following the approval of the official decision extending EU membership

As the Brexit day deadline of March 29th loomed without approval of the proposed withdrawal agreement, the European Council meeting agreed conclusions offering an extension of UK membership of the EU on March 21st. The formal decision on extension was then adopted on March 22nd, after the UK agreed to it. What are the legal issues concerning the extension of EU membership?

EU law issues

The extension decision states that:

In the event that the Withdrawal Agreement is approved by the House of Commons by 29 March 2019 at the latest, the period provided for in Article 50(3) TEU is extended until 22 May 2019.

In the event that the Withdrawal Agreement is not approved by the House of Commons by 29 March 2019 at the latest, the period provided for in Article 50(3) TEU is extended until 12 April 2019. In that event, the United Kingdom will indicate a way forward before 12 April 2019, for consideration by the European Council

If the House of Commons approves the withdrawal agreement next week, there's an extension to 22 May to sort out the details (notably the Act of Parliament necessary to implement the withdrawal agreement). If the agreement is not approved next week, there's a shorter extension to 12 April, and the UK will indicate what it sees as the way forward before this date.

The starting point for legal discussion of this issue is Article 50(3) TEU, which provides, as regards a Member State withdrawing from the EU:

The Treaties shall cease to apply to the State in question from the date of entry into force of the withdrawal agreement or, failing that, two years after the notification referred to in paragraph 2, unless the European Council, in agreement with the Member State concerned, unanimously decides to extend this period. (emphasis added)

Unanimity can still apply despite abstentions (Article 235 TFEU), and the European Council can, if necessary, act by written procedure (see its Rules of Procedure). (In the event, it seems that neither was necessary). It logically follows that the Member State concerned has to agree to the text of the relevant European Council decision, in particular as regards the length of extension - which is what happened in practice. There’s no formal role for the European Parliament or national parliaments, but it's possible that their expressions of opinion had some influence. Article 50(3) is silent on whether or not there can be multiple extensions.

During the extension, the UK has the same rights and responsibilities as it would ordinarily have as a Member State otherwise, including the continuing right to revoke its notification of intention to leave the EU unilaterally (see the CJEU’s Wightman judgment; the general point is confirmed by preamble clause 10). This follows on from two earlier CJEU rulings (discussed here and here), in which the Court confirmed that the UK remained a fully-fledged Member State throughout the main two-year period after notifying its intention to leave the EU. In other words: Membership means Membership. (This rule doesn’t prevent political conditions being attached to the EU’s decision to agree to an extension though, such as those in the decision). Those who claimed that the EU would demand vast sums or abolition of opt outs as a condition of extension of membership were incorrect. 

This general rule raises a specific issue relevant to extension of the UK’s EU membership. Would the UK have to hold European Parliament elections, set for May 23-26? (The date is based on the basic law on EP elections, since amended in 2018 on other points).  The answer would obviously be no, if the extension did not go past May 22. Arguably the answer would still be no if the extension went no further than June 30 – as requested by the Prime Minister – because the new European Parliament would only take office after that point. However, this is disputed (see the recent EU room document on extension, Wednesday’s Commission paper, and the contrasting views of Professors Barnard and Weatherill, Professor Spaventa and the UK’s CJEU Advocate-General Eleanor Sharpston). Ultimately the European Council took the view that membership after May 22 does require holding EP elections, and noted that April 12 would be the last date on which the UK could decide to hold them, explaining why these dates appear in the decision (see clause 10 in the preamble).

In the event of extension after June 30, the UK would certainly be in breach of its obligation to hold EP elections, unless a special exception was granted to it. This is because the Treaties refer to election for five-year terms, and a Council decision (see link above) sets out the election dates. Both the Treaties and the basic law on EP elections need unanimous national ratification to amend them. Article 50 does not refer to granting exceptions from EU law, except as regards the departing Member State’s absence from discussions concerning its departure. While new Member States do have temporary exceptions from the obligation to elect MEPs, they are covered by a different legal framework: Article 49 TEU, which deals with accession, expressly allows for “adjustments” to the EU Treaties. As noted above, tonight’s conclusions make no specific mention of the elections issue (it appeared in an earlier draft, but was dropped).

Would the EP (and the new Commission which the EP has a big role in appointing) be acting illegally, if the UK had not elected MEPs in time? The EU courts have answered a similar question before, as regards the Santer Commission which resigned irregularly and was not immediately replaced. In the British Airways judgment, it was held that the Commission’s actions adopted during this period were not illegal.  There was also a brief period when the Commission’s status was uncertain, as the Lisbon Treaty came into force a month after a new Commission (with fewer Commissioners) was due to be appointed in 2009. In practice, the new Commission was appointed after a short delay, with the old Commissioners spending a short extra period in their jobs.

If the UK held EP elections in (say) September, there could be a delay in appointing a new Commission or adopting new EU legislation until then. (A new EP takes a while to have a big role in adopting legislation anyway, since a lot of proposed legislation is wrapped up before the outgoing EP dissolves for elections).  There might, however, be some complication as regards the number of MEPs per Member State (see the relevant European Council decision).

UK legal issues

In order to change ‘exit day’ as far as Westminster legislation is concerned, it’s necessary to adopt secondary legislation – a statutory instrument approved by Parliament – under s 20 of the EU Withdrawal Act. It’s been suggested that a Statutory Instrument could be tabled next week. (The UK doesn’t need to adopt a new law to hold EP elections, since the repeal of the legislation concerned has not been commenced yet). In the event that an extension decision is adopted, but the definition of ‘exit day’ has not changed in the UK, it’s been argued that this does not mean that the UK leaves the EU with no deal, but rather that it stays in the EU with no domestic legal framework to give effect to EU law (although it might be possible to fix this quickly). (Updated March 24th: see also Professor Mark Elliott's post on this issue, and the reply of Brexit minister Lord Callanan).

Apart from the legal issues directly related to the extension, next week is likely to see a number of crucial votes relating to the UK’s imminent and future relationship with the EU. Time will tell what indirect effect they may have on the length and circumstances of the UK’s extended EU membership – and in particular whether the UK government and parliament is able to agree upon a way forward in the event that the withdrawal agreement is not approved.

Barnard & Peers: chapter 27
Photo credit:

Sunday 17 March 2019

Of extension of UK membership and basic democratic principles: why the UK will need to hold EP elections if its membership of the EU is prolonged beyond the 22nd of May

Professor Eleanor Spaventa, Bocconi University

To say that the political situation is the UK regarding Brexit negotiations is volatile is an underestimation of reality; Mrs May’s tactic of running down the clock might or might not be successful, but it also appears to have backfired, since now there is an unavoidable need to seek an extension to Article 50 TEU. Should the Withdrawal Agreement finally be accepted by the British Parliament this week, the UK will need an extension to the exit day in order to pass all the legislative instruments needed before withdrawal from the EU.

Should the Withdrawal Agreement be rejected, then the UK will need an extension to either prepare for a non-agreed exit, or to decide where its future lies. But once again the political turmoil, and the options available, are constrained by the constitutional framework of the EU. In particular, the issue is whether the UK would have to hold European Parliament elections if its membership of the EU is to continue beyond the 23rd of May. The problem of course is that holding EP elections is symbolically and politically problematic for the Government, and yet, even opting for a shorter extension, the UK might need more time to pass all the required legislation.

The European Commission’s view seems to be that the UK will not be able to seek an extension beyond 23rd May unless it holds EP elections. The Commission’s position is based on a very simple and yet, in the writer’s opinion unchallengeable premise: EU citizens have a Treaty right to vote for the European Parliament and, as long as the UK is a member of the EU, that right is bestowed on its citizens and on EU citizens living in the UK.

Yet, illustrious EU lawyers have argued in this blog and elsewhere that an extension could stretch to the end of June, without compromising the legitimacy of the newly elected European Parliament. The same point was made in a legal opinion to the European Parliament. Yet, to focus simply on the legitimacy of the European Parliament misses the point: democracy is not a mere exercise in institutional balance – if it were so there would be nothing preventing any Member State or the EU to prolong the term of their Parliaments beyond the fixed mandatory term. After all, if one Member State were to be allowed not to hold European Parliament elections because of its own political problems, why not another Member State for its own political problems. So at first sight, the idea that the UK could be a member of the EU and withhold the right of its citizens, and of EU citizens resident in its territory, to vote for the European Parliament elections seems to be an anathema for a polity which is based on the rule of law. And the legal impossibility of such a scenario is backed not only by the Treaties, and the Charter, but also by the obligations bestowed on the UK and all of the other Member States by the European Convention on Human Rights.

The law on EP elections as it stands

Article 14(3) of the TEU provides that members of the European Parliament shall be elected for a term of five years by direct universal suffrage, whereas Article 20 (2)(b) TFEU provides that every citizen of the EU shall have a right to vote and stand for elections in the State where they reside under the same conditions as nationals. The Court of Justice, in its ruling in Delvigne (discussed here), clarified that the franchise for European Parliament elections is a matter falling within the scope of EU law, even in relation to own citizens. For this reason, the Charter of Fundamental Rights applies; the Court found that the deprivation of the right to vote of a own citizen is a limitation of the right conferred in Article 39(2) Charter; as such it must be justified under the conditions provided for in Article 52 Charter. The limitation to the right to vote must therefore be provided by law (which in the case of the UK would require a new Act since at present the UK simply has not made any provision for the EP elections), must be justified in relation to a public interest recognized by the Union, be necessary for the attainment of that interest and be proportionate. The disenfranchisement of an entire nation would fail to meet these requirements: it would be difficult to identify a public interest pursued by such disenfranchisement (unless one believes that protecting the Conservative Party from self-implosion can be qualified as a public interest), nor would disenfranchisement be proportionate or necessary to ensure withdrawal.

Furthermore, rights conferred by the Charter cannot fall below those provided for in the ECHR: the European Court of Human Rights has clarified in Matthews that Article 3 of Protocol 1 of the ECHR, the right to free elections, applies in relation to the European Parliament. Member States are then obliged, under the Convention, to guarantee the franchise for it, as well as respect basic democratic principles. For this reason, the UK was in violation of its ECHR obligations for disenfranchising citizens in Gibraltar who did not (at the time) have a right to vote for the European Parliament.

If the UK seeks an extension it will remain a Member State of the EU, and as such bound by the Treaties and the Charter. The UK is also bound by the ECHR in this matter, both as a Member of the EU and as a Member of the Council of Europe. It therefore seems that it would not be able, under both EU and ECHR law, to disenfranchise its own citizens.

Barnard and Weatherill, as well as AG Sharpston, suggest that the democratic issue could be overcome by prolonging the mandate of the current British MEPs – yet, such a suggestion is not only contrary to the Treaty but also inimical to the very basics of democratic principles (it is dictators that usually resort to these systems). Their suggestion that, since those MEPs have been democratically elected in 2014, they can continue to represent British citizens past the life of this European Parliament, without seeking a new mandate is far from being persuasive, not only because there is no such possibility in the Treaties, but also given that the passive electorate has changed in the past five years and that prolonging MEPs would entail the disefranchisement of those who have come to voting age in the past 5 years. And even if it were legally possible under EU law to provide such a transitory arrangement, through a derogation from the Treaty provisions as suggested by AG Sharpston, such a transitional arrangement would still breach Article 3 of Protocol 1 of the ECHR.

In as much as we might want to avoid Brexit chaos, the end cannot justify the means in this case; disenfranchisement is illegal and infringes the founding principles upon which the EU is founded (democracy, rule of law and protection of fundamental rights). Furthermore, following the ruling in Wightman (discussed here), the UK can at any point withdraw its notice under Article 50. As clearly stated by the Court, up until that point in which the Withdrawal agreement enters into force or the UK has exited by virtue of the passing of time, the UK is and remains a full member of the EU (also during the extension) – and its citizens therefore are subject to all its rules and derive all the rights stemming from the Treaties.

To allow the UK to remain a member of the EU without participating to the European Parliament elections would not only be illegal, it would be very dangerous (not least in these political times): Brexit is bad enough as it is – it cannot be allowed to destroy the very values upon which the EU is founded and which it already struggles to protect.

Barnard & Peers: chapter 3, chapter 27
Photo credit: Evening Standard