The Brexit talks: opening positions on the status of UK and EU citizens


by Professor Steve Peers


One of the most high-profile issues relating to Brexit, which could potentially have the biggest direct impact on the lives of the greatest number of people, is the issue of what happens to UK citizens in the EU and EU citizens in UK after Brexit. This is one of the first issues to be addressed in Brexit talks, and both sides have now adopted their positions: the EU in the form of a Council decision on the mandate for the Commission negotiators, back on May 22, and the UK in the form of a UK government proposal, released on June 26. As we can see from these dates, it’s entirely false to suggest (as the UK Foreign Secretary has done, for instance) that this UK government proposal came first, with no EU position yet: it’s quite the opposite. (Equally it’s false to suggest, as the Brexit Secretary does, that among the EU institutions, only the EU Commission is demanding that the ECJ have a role in the agreement).

This EU position also covers the issues of the financial consequences of Brexit and its purely transitional aspects (ie court cases pending on Brexit Day), which no published UK proposal has addressed yet. However, I will focus solely on the citizens’ rights issues for now. For the sake of simplicity, the relevant parts of the EU position are repeated in the Annex to this blog post.

There is a basic choice to be made whether the position of UK and EU citizens after Brexit is based on the ‘acquired rights’ approach (ie retaining their status under EU law) or an approach based on equality with nationals. As we will see, the EU takes the former approach, while the UK takes the latter, even though during the referendum campaign the Leave side promised acquired rights to both EU citizens in the UK (‘no change’, ‘no less favourable’) and UK citizens in the EU.

The EU position

Basically, the EU position follows the ‘acquired rights’ approach, adopting a broad interpretation of that concept to include rights which will vest in future as well as those ‘in the process of being obtained’, specifically permanent residence status which can be obtained under EU free movement law after five years’ continuous legal residence. It explicitly covers both EU citizens in the UK and UK citizens in the EU, including those who previously resided on one side or the other. Protection will be based on equal treatment compared to nationals – reflecting the second option for approaching the issue – for the lifetime of each person, via ‘smooth and simple administrative procedures’.

The EU position goes on to define the personal scope of the deal: those covered by the EU citizens’ Directive (workers, self-employed and economically inactive people – implicitly subject to the limits in the Directive for ‘benefit tourists’, as discussed here), also including family members who arrive before or after Brexit Day. It will also repeat the scope of the EU social security Regulation, which addresses social security coordination in cross-border situations as distinct from immigration status, including frontier workers (ie those who work in the UK but live in France, or vice versa).

The material scope of the deal (ie the rights to be protected) should include residence rights based on the Treaties or the citizens’ Directive, as well as the procedural rules on documenting those rights; the social security coordination rules, including export of benefits and cumulating social security contributions made in different countries; the supplementary rights in the Regulation on free movement of workers, including workers’ children’s access to education; access to self-employment; and recognition of qualifications which were obtained before Brexit Day or which are in the process of being recognised on that date.

As for enforcement, the EU side wants this issue to be enforced by the ECJ, and the rules in the withdrawal agreement to be enforced in accordance with pre-Brexit case law of the Court. A separate position paper makes clear that this refers to all of the Court’s current jurisdiction, in particular references from national courts to the ECJ and Commission challenges to the UK.

The UK position

Firstly, the UK paper states that it will not alter the Common Travel Area arrangements between the UK and Ireland (and the Crown Dependencies), including ‘the rights of British and Irish citizens in each others’ countries rooted in the Ireland Act 1949’. To that end, ‘Irish citizens residing in the UK will not need to apply for settled status to protect their entitlements’. (It should be noted that some have questioned how much the Ireland Act in fact protects Irish citizens’ immigration status in the UK).

Next, the document suggests its legal form: the government ‘undertakes to treat EU citizens in the UK according to the principles below, in the expectation that the EU will offer reciprocal treatment for UK nationals resident in its member states’. It’s not clear if this is a unilateral offer conditioned on the assumption that the EU side will match it, or whether it is a proposal to be subject to negotiations with the view to being included in the withdrawal treaty. (At some other points, the document refers to ‘negotiations’ and to an ‘international law’, however).

In detail, the UK government states first that it will comply with EU free movement law until Brexit Day. Next, post-Brexit it ‘will create new rights in UK law for qualifying EU citizens resident here before our exit. Those rights will be enforceable in the UK legal system and will provide legal guarantees for these EU citizens’, alongside ‘commitments in the Withdrawal Agreement which will have the status of international law’. The paper rules out ‘jurisdiction in the UK’ for the ECJ. Furthermore, the government paper pledges to treat ‘all EU citizens equally’ compared to each other, although it is not clear how this fits with the special dispensation for Ireland referred to at the outset.

While ‘qualifying EU citizens will have to apply for their residence status’, the ‘administrative procedures’ to this end ‘will be modernised and kept as smooth and simple as possible’. But this will be a national process: ‘a separate legal scheme, in UK law, rather than the current one for certifying the exercise of rights under EU law’. This means that the UK government ‘will tailor the eligibility criteria so that, for example, we will no longer require evidence that economically inactive EU citizens have previously held “comprehensive sickness insurance” in order to be considered continuously resident’. The words ‘for example’ there suggest that there might be other (unspecified) differences between the criteria for obtaining status in the UK for EU citizens.

As part of the process, ‘all qualifying EU citizens will be given adequate time to apply for their new residence status after’ Brexit. This will take the form of a ‘guarantee that qualifying individuals will be granted “settled status” in UK law (indefinite leave to remain pursuant to the Immigration Act 1971).’ This means ‘they will be free to reside in any capacity and undertake any lawful activity, to access public funds and services and to apply for British citizenship’.

To get this status, ‘the EU citizen must have been resident in the UK before a specified date’, which is yet to be defined; but it will be in between March 29 2017 when the Article 50 letter was sent, and March 29 2019, Brexit Day (the government is expressly intending to negotiate this with the EU). They must also ‘have completed a period of five years’ continuous residence in the UK before they apply for settled status, at which point they must still be resident’. Since the criteria are national, not based on EU law, the calculation of this period might differ. As for ‘those EU citizens who arrived and became resident before the specified date’ but who have not accrued five years’ continuous residence on Brexit Day, they ‘will be able to apply for temporary status in order to remain resident in the UK until they have accumulated five years, after which they will be eligible to apply for settled status’.

On the other hand, those EU citizens who arrive after the [un]specified date ‘will be allowed to remain in the UK for at least a temporary period and may become eligible to settle permanently, depending on their circumstances – but this group should have no expectation of guaranteed settled status’. This category of people will therefore be treated quite differently than under the EU proposal.

As for family members, any ‘family dependants’ who join a qualifying EU citizen in the UK before Brexit ‘will be able to apply for settled status after five years’ (including where the five years falls after our exit), irrespective of the specified date’. Again, it is unclear what the definition of ‘family members’ will be. However, family members arriving after Brexit will be subject to the same immigration rules as the family of UK citizens, ‘or alternatively to the post-exit immigration arrangements for EU citizens who arrive after the specified date’. This suggests a willingness to negotiate special rules on this issue with the EU.

There will be an exclusion for ‘those who are serious or persistent criminals and those whom we consider a threat to the UK’; this might not match the rules permitting exclusion of criminals and security threats set out in the EU legislation and ECJ case law. As for ‘benefits, pensions, healthcare, economic and other rights, in the expectation that these rights will be reciprocated by EU member states, the Government intends that:’ settled EU citizens ‘will continue to have access to UK benefits on the same basis as a comparable UK national under domestic law’; those EU citizens who arrived before the specified date will ‘continue to be able to access the same benefits that they can access now – (broadly, equal access for workers/the self-employed and limited access for those not working)’, on their route to settled status. If they later get settled status, they will have access to benefits ‘on the same terms as comparable UK residents’. Also, export of benefits to the EU ‘will be protected for those who are exporting such UK benefits on the specified date, including child benefit, subject to on-going entitlement to the benefit’. (Note that the right to export benefits will implicitly not be offered to those who arrive after the specified date).

Furthermore, ‘the UK will continue to export and uprate the UK State Pension within the EU’; this mainly concerns UK citizens retiring abroad, but some EU citizens will have acquired such rights from their UK employment too.  Other forms of social security coordination will continue, including aggregation of national insurance contributions for UK benefits and state pensions, even if granted after Brexit, and healthcare arrangements set out in UK and EU law; in particular, the UK will ‘seek to protect the ability of individuals who are eligible for a UK European Health Insurance Card (EHIC) before the specified date to continue to benefit from free, or reduced cost, needs-arising healthcare while on a temporary stay in the EU’. Negotiations on ‘an ongoing arrangement akin to the EHIC scheme’ are planned, but there is no reference to negotiations on the other social security issues, even though it may prove technically and administratively difficult to aggregate contributions and pay benefits without a formal basis for cooperation. It is not clear if the UK plans to continue applying any of the relevant EU legislation as such; if it does not, negotiations and implementation of the rules will be more complicated.

Next, as regards education, the UK government ‘will ensure qualifying EU citizens who arrived in the UK before the specified date will continue to be eligible for Higher Education (HE) and Further Education (FE) student loans and ‘home fee’ status in line with persons with settled status in the UK’, as well as maintenance support (where it exists) ‘on the same basis they do now’. Equal treatment in tuition fees will still apply to those EU students who are enrolled during the 2017/18 and 2018/19 academic years ‘for the duration of their course’, along with ‘a parallel right to remain in the UK’ for those students ‘to complete their course’. (There’s no reference to a right to stay for other purposes after Brexit). The UK government ‘will seek to ensure that citizens with professional qualifications obtained in the EU27 prior to the UK’s withdrawal from the EU will continue to have those qualifications recognised in the UK (and vice versa)’. This matches the EU position, albeit with more equivocal language.

As for documentation, EU citizens will need to obtain evidence of ‘settled status’ eventually, but they do not need to apply now, although an application process will be set up prior to Brexit ‘to enable those who wish to do so to get their new status at their earliest convenience’. Those who have already got documentation of permanent residence will have to apply again, but ‘we will seek to make sure that the application process for settled status is as streamlined as possible’. Fees will be set ‘at a reasonable level’. There will be a grace period of perhaps two years while all EU citizens resident under the old system have an opportunity to transition to the new one. If they fail to apply to be covered by the new system, they lose their permission to stay.

Finally, the UK will ‘discuss similar arrangements with Iceland, Liechtenstein, Norway and Switzerland’ which are also subject to free movement rules ‘on a reciprocal basis’.

As for UK citizens in the EU, the government says they ‘must be able to attain a right equivalent to settled status in the country in which they reside’ and ‘continue to access benefits and services across the member states akin to the way in which they do now.’ The UK will also seek to ensure their continued right to establishment and cross-border provision of services within the EU.


Since the EU position refers to the continuation of existing law, there are few ambiguities in its meaning (besides those inherent in that existing law anyway – for instance, the exact status of same-sex marriages is pending before the ECJ, as discussed here). There are still some vague points, however. Firstly, is the reference to those who have previously resided in the EU or UK meant to be free-standing, or does it simply refer to the more detailed rules set out in the EU legislation referred to? (For instance, a UK pensioner living in Spain might be receiving a UK pension on the basis of contributions made some years ago).

Secondly, it seems that the reference to rights based on the Treaties covers non-EU parents of UK children in the UK, ie the so-called Ruiz Zambrano cases (see further discussion here). Thirdly, would UK citizens resident in the EU on Brexit Day still retain the right of free movement between Member States – ie would a UK citizen in France on that day retain full free movement rights to move on to Germany in future? Finally, how would each side distinguish between those UK and EU citizens with acquired rights on Brexit Day, and those (principally those who move afterward) who do not have such rights?

In comparison, the UK position is necessarily vaguer, since it does not refer to EU law as such. As noted above, therefore, some of its key features are unclear, notably the definition of the grounds for ‘settled’ status, the scope of persons who might be excluded from that status, and family members. Much of the UK position uses ‘weasel words’ like ‘seek to ensure’ or ‘akin’.

To the extent that its content can be discerned, the UK position is indisputably offering worse terms both for EU citizens in the UK and UK citizens in the EU. First of all, the cut-off date in the EU position is Brexit Day, whereas it might be earlier in the UK position. The UK suggests that EU citizens in the UK might not be treated equally even if they have permanent residence status by the cut-off date, since they will have to transfer to settled status; the application process to that end would not be necessary in the EU position. While the UK will exempt people from the requirement to have Comprehensive Sickness Insurance, it has been argued that the current UK law on this point breaches EU law anyway.

For those EU citizens who do not have settled status by the cut-off date, or who arrive after the cut-off date but before Brexit Day, they will be worse off than under the EU proposal, since they will not be covered by EU free movement law as regards the acquisition of EU permanent residence status. All categories of EU citizen will have a diminished right to family reunion after Brexit Day.

For UK citizens in the EU, the UK position that they should get settled status in the relevant EU country would not necessarily ensure a right equivalent to EU free movement law permanent residence status. Moreover, those who have not obtained such status as of Brexit Day will not necessarily be able to obtain it as easily as EU citizens do, since free movement law would no longer apply. The word ‘akin’ as regards equal treatment is also vague. While the UK would aim to keep their right of establishment and freedom to provide services, there is no reference to the broader free movement rights arguably inferred by the EU position.

The two sides obviously also differ on the role of the ECJ: it would keep its full current role under the EU proposal, while lose its jurisdiction in the UK under the UK proposal. The latter would leave it with jurisdiction over UK citizens in the EU, and arguably a possible limited role in dispute settlement. Note that the UK implicitly is willing to consider an alternative method of dispute settlement: this could be a new court, a form or arbitration, or use of the existing EFTA Court, which applies EU internal market and related law in Norway, Iceland and Liechtenstein, subject to a requirement to apply ECJ case law adopted before the date of the agreement and to take later case law into account. (This latter requirement matches the EU position, and nearly matches the UK plans for the Great Repeal Bill).

Taken as a whole then, the UK position is much vaguer and offers significantly less to both EU citizens in the UK and UK citizens in the EU than the EU position does, although the gap is much wider for those who do not yet have EU permanent residence status. There is also an enforcement gap as regards the role of the ECJ, although there are precedents (notably the EFTA Court, agreements with Switzerland and Turkey) for the EU not insisting that its citizens living outside the EU have their rights enforced by the ECJ. Any compromise would most likely be based on: a) the EU side accepting an alternative means of enforcement of rights other than the ECJ; b) a cut off date of Brexit Day; and c) the two sides agreeing to base protection on the acquired rights approach with certain exceptions (family members admitted after Brexit, more stringent rules for those with criminal convictions).



EU negotiation position

20 The Agreement should safeguard the status and rights derived from Union law at the withdrawal date, including those the enjoyment of which will intervene at a later date (e.g. rights related to old age pensions) as well as rights which are in the process of being obtained, including the possibility to acquire them under current conditions after the withdrawal date (e.g. the right of permanent residence after a continuous period of five years of legal residence which started before the withdrawal date). This should cover both EU27 citizens residing (or having resided) and/or working (or having worked) in the United Kingdom and United Kingdom citizens residing (or having resided) and/or working (or having worked) in one of the Member States of the EU27. Guarantees to that effect in the Agreement should be reciprocal and should be based on the principle of equal treatment amongst EU27 citizens and equal treatment of EU27 citizens as compared to United Kingdom citizens, as set out in the relevant Union acquis. Those rights should be protected as directly enforceable vested rights for the life time of those concerned. Citizens should be able to exercise their rights through smooth and simple administrative procedures.

21 The Agreement should cover at least the following elements:a) Definition of the persons to be covered: the personal scope should be the same as that of Directive 2004/38 (both economically active, i.e. workers and self-employed, as well as students and other economically inactive persons, who have resided in the UK or EU27 before the withdrawal date, and their family members who accompany or join them at any point in time before or after the withdrawal date). In addition, the personal scope should include persons covered by Regulation 883/2004, including frontier workers and family members irrespective of their place of residence.

b) Definition of the rights to be protected: this definition should include at least the following rights:

i) the residence rights and rights of free movement derived from Articles 18, 21, 45 and 49 of the Treaty on the Functioning of the European Union and set out in Directive 2004/38 (covering inter alia the right of permanent residence after a continuous period of five years of legal residence and the right as regards access to health care) and the rules relating to those rights; any document to be issued in relation to the residence rights (for example, registration certificates, residence cards or certifying documents) should have a declaratory nature and be issued under a simple and swift procedure either free of charge or for a charge not exceeding that imposed on nationals for the issuing of similar documents;

ii) the rights and obligations set out in Regulation 883/2004 on the coordination of social security systems and in Regulation 987/2009 implementing Regulation 883/2004 (including future amendments of both Regulations) covering inter alia, rights to aggregation, export of benefits, and principle of single applicable law for all the matters to which the Regulations apply;

iii) the rights set out in Regulation 492/2011 on freedom of movement for workers within the Union (e.g. access to the labour market, to pursue an activity, social and tax advantages, training, housing, collective rights as well as rights of workers’ family members to be admitted to general educational, apprenticeship and vocational training courses under the same conditions as the nationals of the host State);

iv) the right to take up and pursue self-employment derived from Article 49 of the Treaty on the Functioning of the European Union.

  1. For reasons of legal certainty, the Agreement should ensure, in the United Kingdom and in the EU27, the protection, in accordance with Union law applicable before the withdrawal date, of recognised professional qualifications (diplomas, certificates and other evidence of formal qualification) obtained in any of the Union Member States before that date. The Agreement should also ensure that professional qualifications (diplomas, certificates or other evidence of formal qualification) obtained in a third country and recognised in any of the Union Member States before the withdrawal date in accordance with Union law rules applicable before that date continue to be recognised also after the withdrawal date. It should also provide for arrangements relating to procedures for recognition which are ongoing on the withdrawal date. (…)
  1. The Agreement should include provisions ensuring the settlement of disputes and the enforcement of the Agreement. In particular, these should cover disputes in relation to the following matters:

– continued application of Union law;

– citizens’ rights;

– application and interpretation of the other provisions of the Agreement, such as the financial settlement or measures adopted by the institutional structure to deal with unforeseen situations.

  1. In these matters, the jurisdiction of the Court of Justice of the European Union (and the supervisory role of the Commission) should be maintained. For the application and interpretation of provisions of the Agreement other than those relating to Union law, an alternative dispute settlement should only be envisaged if it offers equivalent guarantees of independence and impartiality to the Court of Justice of the European Union.
  1. The Agreement should foresee that any reference to concepts or provisions of Union law made in the Agreement must be understood as including the case-law of the Court of Justice of the European Union interpreting such concepts or provisions before the withdrawal date. Moreover, to the extent an alternative dispute settlement is established for certain provisions of the Agreement, a provision according to which future case-law of the Court of Justice of the European Union intervening after the withdrawal date must be taken into account in interpreting such concepts and provisions should be included.

Counter-terrorism and the inflation of EU databases

Original published on Statewatch (*) on May 2017

By Heiner Busch (@Busch_Heiner) and Matthias Monroy (@matthimon)  (Translation from DE by Viktoria Langer)

The topic of counter-terrorism in Europe remains closely linked to the development and expansion of police (and secret service) databases. This was the case in the 1970s, after 11 September 2001 and has also been the case since 2014, when the EU Member States started working on their action plans against ‘foreign terrorist fighters’.

The first effect of this debate has been a quantitative one: the amount of data in the relevant databases has increased explosively since 2015. This can be seen by looking in particular at available data on the Europol databases, like ‘Focal Points’ (formerly: Analytical Work Files) of the Europol analysis system. Since 2015 they have become one of the central instruments of the European Counter Terrorism Centre (ECTC) which was established in January 2016. ‘Hydra’, the ‘Focal Point’ concerning Islamist terrorism was installed shortly after 9/11. In December 2003 9,888 individuals had been registered, a figure that seemed quite high at the time – but not compared with today’s figures. [1] In September 2016 ‘Hydra’ contained 686,000 data sets (2015: 620,000) of which 67,760 were about individuals (2015: 64,000) and 11,600 about organisations (2015: 11,000).

In April 2014 an additional ‘Focal Point’, named ‘Travellers’, was introduced, which is exclusively dealing with “foreign terrorist fighters” (FTF). One year later ‘Travellers’ included 3,600 individuals, including contact details and accompanying persons. In April 2016 the total number increased by a factor of six. Of the 21,700 individuals registered at the time, 5,353 were “verified” FTFs. In September 2016, of 33,911 registered individuals, 5,877 had been verified as FTFs.

Since 2010 Europol and the USA have operated the Terrorist Finance Tracking Programme (TFTP), which evaluates transfers made via the Belgian financial service provider SWIFT. Until mid-April 2016 more than 22,000 intelligence leads had been arisen out of that programme, of which 15,572 since the start of 2015. 5,416 (25%) were related to FTFs.

In contrast to Europol’s analytical system, the Europol Information System (EIS, the registration system of the police agency) can be fed and queried directly from the police headquarters and other authorities of EU Member States. Here, more than 384,804 ‘objects’ (106,493 individuals) were registered at the start of October 2016, 50% more than the year before. The increase is partly due to the growing number of parties participating in the EIS. In 2015 13 Member States were connected; in 2016 19 Member States. Some of the EU States, like the UK, also let their national secret services participate in the system. 16 Member States currently use automatic data uploaders for input. The number of third parties involved has also increased (in 2015 there were four, in 2016 there were eight). Interpol, the FBI and the US Department of Homeland Security are some of them.

Europol has reported further growth in the number of “objects” linked to terrorism in the EIS. According to the Slovak Presidency of the Council of the EU’s schedule for the improvement of information exchange and information management, in the third quarter of 2016 alone these grew another 20% to 13,645. [2] The EIS includes 7,166 data sets about individuals linked to terrorism, of which 6,506 are marked as FTFs or their supporters, or are assumed to be so. For May 2016 the CTC stated a figure of 4,129. [3] The increase in terrorism linked data can also be seen in the Schengen Information System (SIS) – in the alerts for “discreet checks or specific checks” following Article 36 of the SIS Decision. According to this, suspect persons are not supposed to be arrested. However, information about accompanying persons, vehicles etc. are recorded to provide insight into movements and to keep tabs on the contacts of the observed person. At the end of September 2016 the number of such checks by the police authorities (following Article 36(2)) was 78,015 (2015: 61,575, 2014: 44,669). The number of alerts of the national secret services based on Article 36(3) was 9,516 (2015: 7,945, 2014: 1,859). “Hits” on such alerts and additional information are supposed to be sent directly to the alerting authorities and not as usual to national SIRENE offices (which deal with the exchange of supplementary information regarding alerts in the SIS). This option was only introduced in February 2015.

The Schengen states used the instrument for discreet surveillance or specific checks very differently. On 1 December 2015 44.34% of all Article 36 alerts came from authorities in France, 14.6% from the UK, 12.01% from Spain, 10.09% from Italy and 4.63% from Germany. [4] How many of these alerts actually had a link to terrorism remains unclear; a common definition has not yet been found. However, the Council Working Party on Schengen Matters agreed on the introduction of a new reference (“activity linked to terrorism”) for security agencies’ alerts. According to Federal Ministry for the Interior, German alerts are marked with this reference when concrete evidence for the preparation of a serious act of violent subversion (§§129a, 129b Penal Code) can be presented. [5]

‘Unnoticed in the Schengen area’ Continue reading “Counter-terrorism and the inflation of EU databases”

The Bratislava Declaration on migration: European irresponsibility instead of solidarity

ORIGINAL PUBLISHED ON OMNIA (Odysseus Network) SITE (27 Sep 2016)

By Phillippe De Bruycker (ULB/EUI) Evangelia (Lilian) Tsourdi (Max Weber Fellow, EUI)

The Bratislava Declaration refers on two occasions to “the principles of responsibility and solidarity”. The basic idea is to “broaden EU consensus” by devising a “long term migration policy” on the basis of the two principles.

At first look, this seems logical and even advisable. Since 2015, the EU has been unable to respond effectively to the ‘refugee crisis’. It is only the fragile ‘deal’ with Turkey that brought the illusion of a solution by externalising asylum provision to a third country. The EU remains profoundly divided about possible internal solutions. A European East-West divide has appeared, in addition to the well-known North-South division about the principles evoked in the Bratislava Declaration. Member States in the South have been complaining for years about the lack of solidarity measures, while many Member States in the Northwest have castigated them about their inability to implement their responsibilities. More recently, Member States in the Central/Eastern part of the EU (more precisely the Visegrad group consisting of Hungary, the Czech Republic, Slovakia and Poland) are refusing, ostensibly in the name of responsibility, to engage in the type of solidarity requested by no longer only the Member States in the South, but also those in the Northwest.

The objective to heal the wounds and reunify EU Member States around the same principles of solidarity and responsibility appears reasonable and even attractive in this setting. If all Member States (including those in the South) are fully responsible, the others (in particular those in the East) will demonstrate greater solidarity, so that the problem may be solved in a balanced way. This presentation based on an opposition between responsibility and solidarity is, however, simplistic and even incorrect from a legal point of view. If there is indeed in EU law a precise legal provision that can be considered to embody responsibility, applicable in the same manner throughout EU law, the same does not hold true for solidarity (1). Moreover, effective solidarity and fair sharing are a prerequisite to responsibility in EU migration and asylum policies, rather than the other way round  (2).

1. More responsibility than solidarity in EU law in general

When searching in the EU treaties for the word “responsibility”, Article 165(1) TFEU provides an excellent example of the kind of answer that appears: following this provision, “The Union shall contribute to the development of quality education by encouraging cooperation between Member States and, if necessary, by supporting and supplementing their action, while fully respecting theresponsibility of the Member States for the content of teaching and the organisation of education systems and their cultural and linguistic diversity”.Responsibility refers in this sense simply to competence.

Responsibility understood as competence can be envisaged as a power as well as a duty. It is not so surprising that this notion has been linked in the case law of the Court of Justice with the principle of loyalty, now referred to as the principle of sincere co-operation under Article 4(3) TEU. The principle embodies, respectively, a positive obligation (taking measures to ensure fulfilment of obligations), and a negative obligation (abstaining from measures that could jeopardize this fulfilment). It is this first part that is often evoked by Member State governments; with ‘responsibility’ they refer to Member States’ duty to fulfil their obligations and honour their commitments under EU law.

Loyalty has been made explicit under Article 4(3) of the TEU. The principles of loyalty and solidarity are sometimes used interchangeably in legal scholarship, with loyalty considered a facet of solidarity. Under this understanding, the responsibility of Member States to implement their obligations under EU law is a sign of solidarity to each other. This is, however, a narrow understanding of solidarity, which is a notion different from responsibility.

When searching in EU treaties for the word ‘solidarity’, one finds, in particular since the Lisbon Treaty, more results than a similar search for ‘responsibility’. In some instances, solidarity fulfils an aspirational role, providing political orientation, rather than forming the basis of legally binding duties.  For example, following article 3(5) TEU, “In its relations with the wider world, the Union shall (…) contribute to peace, security, the sustainable development of the Earth, solidarity and mutual respect among peoples…”

However, in other areas solidarity forms the basis of concrete actions and legally binding duties as in article 222(1) TFEU, following which “The Union and its Member States shall act jointly in a spirit of solidarity if a Member State is the object of a terrorist attack or the victim of a natural or man-made disaster. The Union shall mobilise all the instruments at its disposal, including the military resources made available by the Member States, to:

(a)       – prevent the terrorist threat in the territory of the Member States;

– protect democratic institutions and the civilian population from any terrorist attack;

– assist a Member State in its territory, at the request of its political authorities, in the event of a terrorist attack;

(b)       assist a Member State in its territory, at the request of its political authorities, in the event of a natural or man-made disaster”.

These latter provision shows that solidarity is not linked with the fulfilment of responsibilities but rather with providing assistance to other Member States in order to allow them to implement their obligations.

Interestingly, solidarity understood in this sense does not have the same status as responsibility understood as loyalty. There is indeed no legal provision of solidarity applicable throughout different policies that would create a general duty to support, but rather different and more or less strong expressions of solidarity. As a consequence, one has to examine each particular policy and the provisions in the EU treaties pertaining to it in order to ascertain whether there are concrete solidarity duties and what the extent of these may be. This leads us to the meaning of solidarity in policies on border checks, asylum and immigration as governed by Articles 77 to 80 TFEU.

2. More solidarity than responsibility in EU migration and asylum policies

When searching for the word ‘responsibility’ or ‘responsible’ in those provisions, there are four hits. Firstly, Article 72 states that the EU competences regarding border checks, asylum and immigration do not affect the “responsibilitiesincumbent upon Member States with regard to the maintenance of law and order and the safeguard of internal security” and, secondly, in Article 73, following which “it shall be open to Member States to organise between themselves and under their responsibility forms of cooperation and coordination as they deem appropriate between the competent departments of their administrationsresponsible for safeguarding national security”. Responsibility in those provisions refers to the notion of competence, i.e. that the Member States remain competent for the maintenance of law and order and internal security, and even exclusively competent for national security.

Another ‘hit’ is found in Article 78(2), requesting the European Union to adopt measures for a common European asylum system comprising, under point (e), “criteria and mechanisms for determining which Member State is responsible for considering an application for asylum or subsidiary protection”. This is the legal basis of the famous “Dublin System”, based on Regulation 604/2013, determining the responsible Member State for examining an application lodged in the EU. As the flaws of this system have already been analysed in numerous publications,including in this blog, it is not necessary to explain them once more.

Let us just remind ourselves that the origin of this regulation goes back to aConvention signed in Dublin on 15 June 1990 (this explains why specialists of EU asylum continue to speak about ‘Dublin’ in relation to this system). The aim of this system is to indicate which Member State is competent when an asylum application is introduced in the EU on the basis of a certain number of criteria. In practice, the responsible Member State will more often than not be the one of the legal or illegal first entry of the third-country national to the EU.

Responsibility in this regulation refers to the idea of competence regarding the examination of asylum applications, so that all Member States have to deal with the asylum applications for which they are responsible. The problem is that the Dublin system was not devised on the basis of solidarity. On the contrary, apart from exceptions based on the right to family unity, or the rights of the child, it is premised on the idea that each Member State should deal with the applications of asylum seekers whose presence is attributable to actions of that Member State. This could be either because it let them enter the EU voluntarily by issuing a visa or residence permit, or involuntarily by not controlling its external borders effectively. It is not a coincidence that the Dublin system was conceived by the North-Western Member States who drafted the Schengen Convention (France, Germany, Belgium, the Netherlands and Luxembourg) which is at the origin of the Dublin Convention. Solidarity was not an issue at that time in such a small and coherent space. Moreover, Dublin was devised in a purely intergovernmental framework, a decade before the beginning of the implementation of the supranational method with regard to asylum policy, as introduced by the Treaty of Amsterdam, without any actor such as the European Commission looking out for the general interest rather than the national interest of each State. It is an excellent example of the kind of measure that Northern governments managed to impose on other Member States of the European Union, who can try to amend it subsequently, although only with the support of those governments, which explains why this has not been possible regarding the core of the system with the regulations Dublin II in 2003 and Dublin III in 2013.

This is crucial as this policy is, like the area of external borders, characterised by asymmetric burdens between the Member States due to the fate of geography. Following this logic, Greece should have examined all the asylum applications that could have been introduced by the hundreds of thousands of third-country nationals who entered the EU through its borders during the year 2015. It should also have intercepted the persons trying to enter the EU through the Greek borders without the requested documents (a passport with very often at least a short-term visa), as well as taken their fingerprints in order to store them inEurodac, a database helping to determine in practice the responsible Member State. In this particular case, it would mean that Greek authorities should have assumed responsibility of one million third-country nationals just because they entered the EU through the Greek territory.

Does it mean that the Southern Member States are legally wrong when they ask for solidarity from the other Member States, and that they should instead, or at least firstly, fulfil their responsibilities deriving from EU law? The answer is actually much more complicated due to Article 80 TFEU, which reads as follows:The policies of the Union set out in this Chapter and their implementation shall be governed by the principle of solidarity and fair sharing of responsibility, including its financial implications, between the Member States. Whenever necessary, the Union acts adopted pursuant to this Chapter shall contain appropriate measures to give effect to this principle”.

This provision is one of those detailing the idea of solidarity in the policies for border checks, asylum and immigration. A quick reading may give the impression that this provision is precisely about two principles that have to be balanced, much like in the Bratislava Declaration. Under this reading, Member States should first fulfil their responsibilities by applying the Dublin Regulation and assuming responsibility for the asylum seekers arriving on their territory before they can expect solidarity. In the event of a failure to take up their responsibilities, they should not expect solidarity, or rather they should be found ‘undeserving’ of it.

However, this provision is about one and not two principles and, more importantly, about the principle of “solidarity and fair sharing of responsibility”.It is interesting to note that the words “fair sharing of” have simply been omitted from the Bratislava Declaration, while they completely change the meaning and content of what is at stake. Instead of an opposition between responsibility and solidarity that should be balanced against each other, the idea of fair sharing of responsibility actually reinforces that of solidarity. The policies of the Union on border checks, asylum and immigration are governed by the principle of solidarity, and responsibilities between the Member States in these areas must be shared in a fair way. If one will agree that fairness leaves some margin of discretion to the European Union, this notion refers to the ideas of equity and justice and thus provides an indication about how the EU policy on borders, immigration and asylum must be conceived and implemented.

It therefore appears that the legal obligation of the EU is not to balance the two principles of solidarity and responsibility, but rather to realise solidarity through a fair sharing of responsibilities. This means also that the concerned Member States should not be expected to implement Dublin as pre-condition for solidarity, but should instead benefit from a system aiming at a fair sharing of responsibility between all EU Member States. Some will say that Dublin is as such not contrary to EU law and that the system could be accompanied by “appropriate measures to give effect to the principle of solidarity and fair sharing of responsibility”, following the wording of Article 80 TFEU. The problem is that Dublin is the source of the asymmetric burdens between Member States, so that it seems difficult to amend or revise it without reversing the basic principle on which it is based.

Conclusion: responsibility or irresponsibility?

Nothing about this constitutional requirement is mentioned in the Bratislava Declaration. On the contrary, the issue of the relocation of asylum seekers, as a concrete solidarity measure at the core of the debate since 2015, has simply disappeared from the agenda, despite the call of the first summit of the Mediterranean countries of the EU organized in Athens on 9 September. This is the case despite the fact that the relocation measures were based on mandatory EU rules, which most Member States do not apply, while some of them openly challenge them, for instance Hungary through the organisation of a referendumcalling the population to vote against them.

What remains is a kind of “flexible solidarity”, following the words of the joint statement of the Heads of Governments of the V4 Countries (the Visegrad group) defined as a concept that “should enable Member State to decide on specific forms of contribution talking into account their experience and potential”, knowing that “any distribution mechanism should be voluntary”. Some observers have already tried to imagine what this could entail. This will become clearer when the Council of Ministers takes a position on the Commission proposal reforming the Dublin system (Dublin IV), which contains a relocation mechanism that appears ambitious but that would in fact be dysfunctional, as underlined by Francesco Maiani in his report for the European Parliament. The European legislator should keep in mind that, despite the discretion left by this provision, Article 80 TFEU requires a strong solidarity mechanism aiming at “fair sharing of responsibility” between the Member States.

The retreat of the EU regarding the issue of solidarity had actually been announced by the President of the Commission himself in his State of the Union speech, where he stated that “Solidarity must be given voluntarily. It must come from the heart. It cannot be forced”. This clearly contradicts the mandatory character of the relocation decision, which was imposed on 22 September 2015 by a qualified majority in the Council against the opposition of Hungary, Slovakia, Romania and the Czech Republic.

The Bratislava Declaration announces a double evolution. First, a so-called principle of responsibility is prioritised over the principle of solidarity and fair sharing, the latter reduced to a “commitment by a number of Member States to offer immediate assistance to strengthen the protection of Bulgaria’s border with Turkey and continue support to other frontline States”. Secondly, “the objective to ensure full control of external borders” is prioritised over the asylum policy, which is not even mentioned in the text.

The so-called “responsibility to ensure full border controls” is nothing else than a rhetoric contrary to the Treaties, ignoring that the Schengen Borders Code is without prejudice to the rights of asylum seekers (see in particular Articles 3 and 4 of Regulation 2016/399 codifying the Schengen Borders Code). Trying to convince public opinion that asylum seekers can simply be rejected at the border without any further examination of their claim is not only illegal but also populistic. This has proven to be impossible, even in the case of a safe third-country, for example Turkey on the basis of the EU/Turkey agreement of 18 March 2016 (see in this blog Henri Labayle’sThe EU-Turkey Agreement on migration and asylum: False pretences or a fool’s bargain?).

The President of the European Council, Donald Tusk wrote in his letter of invitation to the Bratislava Summit that “Europeans all too often heard politically correct statements that Europe cannot become a fortress and that it must remain open”. This is indeed not the case of the Bratislava Declaration where the Heads of State and government want to improve the communication with citizens through the “use of clear and honest language (…) with strong courage to challenge simplistic solutions of extreme or populist political forces”. The problem is that they do exactly this by pretending to build a Fortress Europe, that is de jure impossible. They probably want to prove that this is possible de facto. This is nothing less than European irresponsibility instead of solidarity.

CS and Rendón Marín: Union Citizens and their Third-Country National Parents – A Resurgence of the Ruiz Zambrano Ruling?


by Maria Haag, PhD Researcher, European University Institute (Florence, Italy) & Michigan Grotius Research Scholar, University of Michigan Law School (Ann Arbor, Michigan)


Five years ago, the CJEU delivered its infamous Grand Chamber decision in C-34/09Ruiz Zambrano. It held that “Article 20 TFEU precludes national measures which have the effect of depriving citizens of the Union of the genuine enjoyment of the substance of the rights conferred by virtue of their status as citizens of the Union” (para 42, emphasis added). This ‘genuine enjoyment’-protection had two consequences. First, Union citizens could rely on Article 20 TFEU against their Member State of nationality without having previously made use of their rights to free movement and thus bypassing the Court’s general lack of jurisdiction in ‘purely internal’ situations. Secondly, Member States were precluded from denying a right of residence to third-country national (‘TCN’) parents or caretakers of minor citizens of that Member State, as these children would otherwise be forced to leave the territory of the EU and thus no longer able to make use of the rights granted by Union citizenship.

Shortly after the delivery of this ground-breaking judgment, the Court of Justice proceeded to interpret Ruiz Zambrano very narrowly in a series of cases (C-434/09McCarthy, C-256/11 Dereci and Others, C-40/11 Iida, C-356&357/11 O. and S., C-87/12Ymeraga and Others, C‑86/12 Alokpa and Moudoulou and C-115/15 NA) leading many to wonder about the original significance of the Ruiz Zambrano decision. In contrast to Ruiz Zambrano, these subsequent cases mostly concerned the significance of Article 20 TFEU in a host Member State. The Court held that the applicants fell outside the scope of Article 20, even if they had never moved to another Member State, i.e. had been born in a Member State other than their Member State of nationality and had never left. The most recent cases – C-304/14 CS and C-165/14 Rendón Marín – however, Ruiz Zambrano decision, fully address the right under Article 20 TFEU in the home Member State. On the 13th of September 2016, the Grand Chamber delivered these two decisions in which it considered the effect of a criminal record of a TCN parent on his or her derived residence right under Article 20 TFEU and to what extent this right can be derogated on grounds of public policy or public security.

C-304/14 CS: facts and judgment

The case in CS concerned a Moroccan national, who resided in the UK together with her British national son. In 2012, she was convicted of a criminal offence and given a prison sentence of 12 months. Following her conviction, she was notified of her deportation liability. Her subsequent application for asylum was denied. Upon her appeal, the First-tier Tribunal (Immigration and Asylum Chamber) found that her deportation would violate her child’s rights under Article 20 TFEU. The Home Secretary was granted permission to appeal this decision before the Upper Tribunal, which asked the CJEU, under which circumstances the expulsion of a TCN caretaker of a Union citizen could be permitted under EU law and whether Article 27 and 28 of the Directive 2004/38 (the ‘citizens’ Directive’, which sets out the main rules on EU citizens who move to another Member State) had any effect in this case.

In its two-part decision, the Court firstly answered the question whether a TCN parent of a Union citizen has a derived right of residence in the home Member State under Article 20 TFEU and, secondly, if such a right can be limited on grounds of public policy or public security.

The Court first firmly restated its holding in Ruiz Zambrano. It explained that Article 20 TFEU “precludes national measures which have the effect of depriving Union citizens of the genuine enjoyment of the substance of the rights conferred by virtue of their status as Union citizens” (para 26; citing Ruiz Zambrano para 42). Furthermore, this means that “a right of residence must … be granted to a third-country national who is a family member of [a minor Union citizen] since the effectiveness of citizenship of the Union would otherwise be undermined, if, as a consequence of refusal of such a right that citizen would be obliged in practice to leave the territory of the European Union as whole” (para 29). CS thus had a derived right of residence under Article 20 TFEU in her son’s home Member State.

Secondly, the Court held that, as a general rule, such a derived residence right can be derogated for reasons of public policy or public security: “where the exclusion decision is founded on the existence of a genuine, present and sufficiently serious threat to the requirements of public policy or of public security … that decision could be consistent with EU law” (para 40, emphasis added). However, a deportation decision cannot be made “automatically on the basis solely of the criminal record of the person concerned” (para 41). Thus the UK legislation at issue, which obliges the Home Secretary to make a deportation order of any non-national who is sentenced to a period of imprisonment of 12 months or more, establishes “a systematic and automatic link between the criminal conviction of a person … and the expulsion measure” (para 44) and therefore violates EU law. Instead, it is for the national courts to weigh up “the personal conduct of the individual concerned, the length and legality of his residence on the territory of the Member State concerned, the nature and gravity of the offence committed, the extent to which the person concerned is currently a danger to society, the age of the child at issue and his state of health, as well as his economic and family situation” (para 42, emphasis added).

Furthermore, derogations for reasons of ‘public policy’ or ‘public security’ must be interpreted strictly and decisions are subject to review by the EU institutions (para 37). Lastly, and most notably, the assessment of the individual situation must take account of the principle of proportionality and the rights protected in the Charter of Fundamental Rights of the European Union (‘CFREU’), especially Article 7 on the right to respect of private and family life and Article 24(2) on the obligation of consideration of the child’s best interests (paras 48 and 49).

C-165/14 Rendón Marín: facts and judgment

The facts in Rendón Marín are very similar to the ones in CS and essentially raise the same question, presumably why the Court decided these cases on the same day and why Advocate General Szpunar did not give separate opinions in these cases, but combined the two. Rendón Marín concerned a Colombian national father, who lived in Spain together with his Spanish national son and his Polish national daughter. His application for a residence permit was rejected due to his criminal record. The crucial difference between the facts of the two cases is that Mr Rendón Marín has a Union citizen daughter who lives in a host Member State and a son who lives in his home Member State. There thus exists a cross-border element in the situation of his daughter, but not in his son’s (For further discussion on the cross-border element, see C-200/02Zhu and Chen, especially para 19.).

The part of the Court’s decision concerning the son’s circumstances – a Spanish national in Spain – is almost identical to the Court’s judgment in CS. In fact, some of the paragraphs can be found in exactly the same wording in both decisions (the two cases also had the same rapporteur, Allan Rosas). Interestingly, the Court in Rendón Marínmentioned the possibility of moving to Poland, as this is the Member State of nationality of Mr Rendón Marín’s daughter. Whilst the Court noted the applicant’s objection that the family had no ties to Poland, it did not go into this discussion. (See, in contrast, footnote 109 in Advocate General Szpunar’s Opinion in CS and Rendón Marín. For more on this, see also Advocate General Wathelet’s Opinion in NA, paras 112-117.) Here the Court simply holds that “it is for the referring court to check whether … the parent who is the sole carer of his children, may in fact enjoy the derived right to go with them to Poland and reside with them there” (para 79, citing Alokpa and Moudoulou paras 34-35). The Court therefore did not deny that moving to Poland could be a possible solution in case of the father’s deportation from Spain.

As for the legal status of the daughter, the Court held that, as a Polish national and Union citizen, she could rely on Article 21 TFEU and the Directive 2004/38 to grant her a right of residence in Spain (para 44). Furthermore, the Court stated that if the daughter fulfils the conditions laid down under Article 7(1) Directive 2004/38 (i.e. having sufficient resources and comprehensive health insurance) then the derived right of residence of Mr Rendón Marín, her father and sole caretaker, cannot be refused (para 53). Whilst this derived right of residence can be limited for reasons of public policy or public security (para 57), EU law precludes such limitations on “grounds of a general, preventive nature” (para 61). Instead, it is for the national courts to do a similar weighing-up exercise as laid out in CS (see Rendón Marín, paras 59-66). Derogations from derived rights of residence on the basis of Article 20 TFEU and Article 21 TFEU thus presumably have to withstand the same test.


After a longer period of silence on this issue, the Court in these cases seems at the very least willing to explore the scope of Ruiz Zambrano. (The Court should soon decide another case, Chavez-Vilchez, which raises some further important questions about the scope of that judgment). The two recent judgments, whilst they in some sense appear to diminish the scope of Ruiz Zambrano even further, can also be seen as a restatement of the fundamental significance of the original judgment.

The cases following the Ruiz Zambrano decision made it very clear that protection under Article 20 TFEU is only applicable to a very small number of people in “very specific situations” (Rendón Marín para 74; CS para 29): essentially only to minors who reside with their TCN parents in their home Member State. CS and Rendón Marín both confirm this, but also clarify that a very high level of protection is granted to those Union citizens who fall within the scope of the ‘Ruiz Zambrano-protection’. In fact, the substantive protection against expulsion is equivalent to that of EU citizens (and their family members) who move to another Member State (the Court refers to concepts found in the EU citizens’ Directive and its predecessors, as well as relevant case law), although it is not clear if the same procedural protection applies.

The Court certainly does not exclude the possibility that “in exceptional circumstances” (CS para 50) a criminal and dangerous parent who poses a threat to a Member State’s public policy or public security could be deported. Even if this means that his or her Union citizen children are forced to leave EU territory and thus deprived of the genuine enjoyment of their EU citizenship rights. Nevertheless, the Court insists on a very stringent test before such a decision can be taken.

Most notably, the Court refers to the EU’s Charter of Fundamental Rights and stresses the fact that a deportation decision needs to take account of Article 7 and Article 24(2) of the Charter (see CS paras 36 and 48; Rendón Marín paras 66 and 85). In Dereci, the Court had previously held that “if the referring court considers … that the situation of the applicants in the main proceedings is covered by European Union law, it must examine whether the refusal of their right of residence undermines the right to respect for private and family life provided for in Article 7 of the Charter” (Dereci, para 72). In that case the Court had decided that the circumstances fell outside the scope of EU law, and that it was therefore beyond its jurisdiction to consider a violation of the Charter. In both CSand Rendón Marín, the Court found that the applicants’ circumstances fell within the scope of EU law and thus that the Charter applied.

It is also interesting to compare the protection granted in C-135/08 Rottmann against the deprivation of the legal status of Union citizenship altogether and the protection granted in CS and Rendón Marín against being deprived of the genuine enjoyment of the Union citizenship rights by means of a parent’s expulsion to a non-EU state. Whereas in Rottmann, the Court held that a decision to withdraw someone’s nationality needs to respect the principle of proportionality (Rottmann, para 59), in CS and Rendón Marín it established a list of criteria that need to be observed. Curiously, the Rottmann-test therefore appears to be narrower than the one established in CS and Rendon Marin, even if the potential outcome in circumstances like Rottmann, i.e. statelessness, might be much more serious for the individual concerned.

In its decision in CS, the Court cites the European Court of Human Rights (ECtHR) judgment in Jeunesse v the Netherlands. The EU Court states in paragraph 49:

“[A]ccount is to be taken of the child’s best interests when weighing up the interests involved. Particular attention must be paid to his age, his situation in the Member State concerned and the extent to which he is dependent on the parent (see, to this effect, ECtHR, 3 October 2014, Jeunesse v. the Netherlands, CE:ECHR:2014:1003JUD001273819, §118).”

Jeunesse v. the Netherlands, which was decided by the Strasbourg court in 2014, concerned a Surinamese national, who lived with her Dutch national husband and children in the Netherlands without a valid residence permit. The applicant argued that the refusal to allow her to reside in the Netherlands infringed her right to respect of her family life under Article 8 ECHR. The facts of this case are very similar to the ones inDereci, in which the Court of Justice held that such a denial of residence right did not conflict with EU law. The ECtHR, however, came to the conclusion that the Dutch authorities had failed “to secure the applicant’s right to respect for her family life as projected by Article 8 of the Convention” (Jeunesse v the Netherlands, §122).

So what does the reference to this judgment mean? First and foremost, the CJEU clarifies and stresses the utmost importance of taking account of the children’s best interests in these deportation decisions. Secondly, it signals the Court’s commitment to taking the fundamental rights of those who fall within the Ruiz Zambrano-protection very seriously.

Finally, the fact that the Court treats the situation of the daughter and the son separately in Rendón Marín reaffirms the Court’s findings in previous cases that a Union citizen in a host Member State first has to rely on Article 21 TFEU before Article 20 can be applied. In the NA judgment, which the Court delivered at the end of June 2016, it held that one first has to examine whether the citizen and their TCN caretaker have a right of residence under secondary EU law. Only if there is no such right, can Article 20 TFEU apply.

The NA case concerned a Pakistani national mother who lived in the UK with her German national children where she was refused a right of residence. The Court decided that because it had already held that both the children and their TCN mother had a right of residence in the host Member State under Article 12 of Regulation No. 1612/68 (paras 52-68), which guarantees children of current and former workers the right to access to education in the host Member State, with corollary residence rights for those children and their parents (for more, see CJEU decisions in C-480/08 Teixeiraand C-310/08 Ibrahim). Article 20 TFEU did not confer a right of residence in the host Member State. It is clear that the protection under Article 20 TFEU is one of last resort. Whilst the Court in NA and Rendón Marín does not directly rule out the possibility that the Ruiz Zambrano-protection might apply in a host Member State, it now almost seems impossible. It appears that that protection can only be granted by the home Member State.

Frontières de l’Union : chronique d’une « recommandation » annoncée ou la flétrissure


par Nuno Piçarra, Omnia, Université de Lisbonne

Le 12 mai 2016, le Conseil a adopté la décision d’exécution 2016/894 « arrêtant une recommandation relative à la réintroduction temporaire du contrôle aux frontières intérieures en cas de circonstances exceptionnelles mettant en péril le fonctionnement global de l’espace Schengen » (ci-après, « décision d’exécution »). Celle-ci se base notamment sur l’article 29 du Code frontières Schengen (ci-après, « CFS ») qui prévoit une « procédure spécifique » tendant à la réintroduction d’un tel contrôle. C’est la première fois que cette procédure, ajoutée au CFS par le règlement n°1051/2013, trouve à s’appliquer.

1. La décision d’exécution 2016/894 du Conseil : recommandations et obligations pour les États Schengen concernés

Concrètement, le Conseil a recommandé qu’à partir du 12 mai 2016 et pendant une durée maximale de six mois, « des contrôles temporaires et proportionnés » soient maintenus (i) par l’Allemagne, à la frontière terrestre avec l’Autriche ; (ii) par l’Autriche, aux frontières terrestres avec la Hongrie et la Slovénie ; (iii) par le Danemark, dans les ports depuis lesquels sont assurées des liaisons par transbordeur vers l’Allemagne, et à la frontière terrestre avec l’Allemagne ; (iv) par la Suède, dans les ports situés dans les régions de police Sud et Ouest, et au pont de Öresund ; (v) par la Norvège, dans les ports depuis lesquels sont assurées des liaisons par transbordeur vers le Danemark, l’Allemagne et la Suède.

La Norvège n’étant pas un État membre de l’UE mais un « pays associé à l’application de l’acquis de Schengen », l’expression « États membres concernés » utilisée par la décision d’exécution, doit être entendue dans le sens de « États de l’espace Schengen » ou simplement « États Schengen ».

Au 12 mai 2016, les cinq États concernés effectuaient déjà des contrôles à leurs frontières intérieures en réponse, selon la décision d’exécution, « à une menace grave pour leur ordre public ou leur sécurité intérieure causée par les mouvements secondaires de migrants en situation irrégulière consécutifs aux défaillances graves dans les contrôles aux frontières extérieures. (…) ces mesures sont nécessaires et considérées comme proportionnées » (attendu 11).

Outre la recommandation de maintien du contrôle aux frontières intérieures identifiées, la décision d’exécution établit de vraies et propres obligations pour les cinq États concernés.

La première est, toutefois, d’exécution difficile voire impossible dans la mesure où, selon le texte qui l’établit, « avant de mettre en place de tels contrôles, les États membres concernés devraient procéder à des échanges de vues avec l’État (les États) membre(s) voisin(s) concerné(s) afin de s’assurer que ces contrôles ne sont effectués que sur les tronçons de la frontière intérieure où ils sont jugés nécessaires et proportionnés, conformément au code frontières Schengen ».

En fait, ce que la décision d’exécution recommande aux cinq États concernés est, non pas de « mettre en place », mais plutôt de « maintenir » des contrôles aux frontières intérieures qui ont été mis en place à partir de septembre 2015. Pour accomplir rigoureusement l’obligation dont il s’agit, il faudrait donc supprimer ou suspendre au préalable ces contrôles afin que l’« échange de vues » puisse avoir lieu avant leur mise en place. Ceci se heurterait, néanmoins, à un obstacle juridique supplémentaire : la décision d’exécution ayant été adoptée le 12 mai 2016 et devant s’appliquer à partir de cette même date, l’échange de vues préalable qu’elle impose ne sauraient s’effectuer en temps utile.

La deuxième obligation imposée aux États concernés est celle de notifier aux autres États Schengen, au Parlement européen et à la Commission « leur décision » de maintenir les contrôles recommandés. C’est, d’ailleurs, ce que prévoit l’article 29, paragraphe 2, quatrième alinéa, CFS. Toutefois, le même article prévoit également au paragraphe 3 que, « en cas de non application par un État membre de la recommandation visée au paragraphe 2, celui-ci en communique sans tarder les motifs par écrit à la Commission ».

Dès lors que la décision d’exécution reprend dans son texte la première disposition citée du CFS, elle aurait dû reprendre également la seconde, pour des raisons de cohérence et de complétude liées au principe de légalité. Et cela même si le Conseil était persuadé que, dans le cas d’espèce, aucun des États concernés ne se prévaudrait de la possibilité prévue par l’article 29, paragraphe 3, CFS pour refuser l’application de la recommandation.

Finalement, la décision d’exécution oblige chaque État concerné à « réexaminer régulièrement la nécessité, la fréquence, le lieu et la durée des contrôles, adapter ces derniers au niveau de la menace à laquelle ils visent à répondre, les supprimant progressivement s’il y a lieu (…) ». Il est ainsi rappelé que les critères fournis par les articles 26 et 30 CFS pour la prise d’une décision de réintroduction temporaire du contrôle aux frontières intérieures « sur tous les tronçons ou sur certains tronçons spécifiques » s’appliquent aussi en aval d’une telle décision.

2. La pratique de réintroduction, par les États Schengen concernés, du contrôle à leurs frontières intérieures avant la décision d’exécution

La date du 12 mai 2016 coïncide précisément avec celle à partir de laquelle l’Allemagne a cessé de pouvoir se prévaloir des dispositions du CFS qu’elle avait invoquées depuis le 13 septembre 2015 – et ce dans le contexte de la « crise des migrants et des réfugiés sans précédent » – pour réintroduire le contrôle à toutes ses frontières intérieures et en particulier à celle avec l’Autriche. L’exemple de l’Allemagne a été suivi trois jours plus tard par l’Autriche, le 12 novembre par la Suède, le 26 novembre par la Norvège et le 4 janvier 2016 para le Danemark.

Dans un premier temps, tous ces États se sont prévalus de l’actuel article 28 CFS, qui établit une procédure spécifique pour la réintroduction exceptionnelle et temporaire du contrôle aux frontières intérieures dans les cas nécessitant une action immédiate. Toutefois, en vertu du même article, la durée totale de cette réintroduction ne peut pas dépasser les deux mois.

Dans ces conditions, l’Allemagne le 14 novembre 2015, l’Autriche le 16 novembre suivant, la Suède le 9 janvier 2016, la Norvège le 15 janvier et le Danemark le 4 mars se sont prévalus des articles 25 à 27 CFS pour maintenir le contrôle à leurs frontières intérieures. Ces articles établissent les conditions et les limites à la réintroduction temporaire d’un tel contrôle en cas d’évènements prévisibles.

Le 23 octobre 2015, la Commission a émis un avis sur la nécessité et la proportionnalité du contrôle aux frontières intérieures réintroduit par l’Allemagne et l’Autriche ainsi que de ses prolongations. Ce contrôle a été considéré conforme au CFS. [C(2015) 7100]. Parmi les commentateurs, il a été relevé que l’avis de la Commission tend à assimiler le franchissement des frontières extérieures par un grand nombre de ressortissants de pays tiers à une menace pour l’ordre public et la sécurité intérieure, même si des preuves de risques liés à la criminalité organisée ou au terrorisme n’ont pas été apportées (voir Evelien Brouwer, « Migration flows and the reintroduction of internal border controls: assessing necessity and proportionality », 2015, et bibliographie y citée).

En vertu de l’article 25, paragraphe 4, CFS, la durée totale de la réintroduction du contrôle aux frontières intérieures en cas d’évènements prévisibles ne peut excéder six mois. Par conséquent, à partir du 12 mai 2016, l’Allemagne cesserait de pouvoir invoquer les articles 25 à 27 CFS pour maintenir le contrôle à ses frontières intérieures. Il en irait de même pour l’Autriche le 15 mai, pour la Suède le 9 juillet, pour la Norvège le 15 juillet et pour le Danemark le 4 août 2016.

Dans ces conditions, la seule possibilité légale pour les États concernés de maintenir le contrôle à leurs frontières intérieures est de le baser sur les articles 29-30 CFS. En effet, en vertu de ces dispositions, la durée totale de la réintroduction de ce contrôle peut être étendue à une durée maximale de deux ans (article 29, paragraphe 1, in fine).

3. L’autre État Schengen concerné par la décision d’exécution : la Grèce. Questions de légalité

La réintroduction ou le maintien du contrôle, par un État Schengen, à ses frontières intérieures conformément à l’article 29 CFS présuppose, outre une recommandation du Conseil sur une proposition de la Commission, l’imputation à un autre ou à d’autres États Schengen des « manquements graves persistants liés au contrôle aux frontières extérieures », à la suite d’une évaluation effectuée en application du règlement nº 1053/2013. Plus précisément, la première condition pour déclencher la procédure prévue à l’article 29 est la persistance de ces manquements, constatée par la Commission au terme d’un délai de trois mois à compter de la date à laquelle l’État Schengen concerné a fait rapport sur la mise en oeuvre de son «plan d’action destiné à remédier à tout manquement constaté dans le rapport d’évaluation ». Ce plan d’action, à son tour, doit être présenté dans un délai d’un mois à compter des recommandations concernant les mesures correctives des manquements en cause, adoptées par le Conseil sur une proposition de la Commission (article 21, paragraphe 3, CFS).

À cette première condition d’ordre simultanément substantiel et procédural, vient s’ajouter une deuxième, d’ordre substantiel : les « manquements graves persistants dans l’exécution du contrôle aux frontières extérieures» constatés par la Commission à l’égard d’un État Schengen doivent « mettre en péril le fonctionnement global de l’espace sans contrôle aux frontières intérieures ». En d’autres mots, ils doivent représenter « une menace grave pour l’ordre public ou la sécurité intérieure » dans cet espace ou sur certains de ses tronçons. Par ailleurs, une telle menace doit s’avérer insusceptible d’être effectivement atténuée, soit par le déploiement d’équipes européennes de gardes-frontières aux frontières extérieures de l’État Schengen en cause, soit par la présentation à Frontex des plans stratégiques basés sur une évaluation des risques, soit par toute autre mesure.

Aux termes de l’article 29 CFS, la réintroduction du contrôle aux frontières intérieures ne peut être recommandée par le Conseil qu’« en dernier ressort et à titre de mesure de protection des intérêts communs au sein de l’espace sans contrôle aux frontières intérieures » face aux manquements graves constatés. Une telle réintroduction doit être évaluée au regard de son adéquation pour « remédier correctement à la menace pour l’ordre public ou pour la sécurité intérieure au sein d’un tel espace », ainsi que de sa proportionnalité. À cette fin il doit être tenu compte, notamment, de la disponibilité de mesures de soutien technique ou financier auxquelles il serait possible de recourir ou auxquelles il a été recouru au niveau national ou au niveau de l’Union, ou à ces deux niveaux, ainsi que de l’incidence probable de la réintroduction du contrôle sur la libre circulation des personnes au sein de l’espace Schengen (article 30, paragraphe 1).

Les conditions extrêmement restrictives pour l’application de cette procédure spécifique de réintroduction ou de maintien du contrôle aux frontières intérieures portent à croire qu’une telle procédure serait surtout destinée à jouer un rôle préventif d’avertissement aux États Schengen pour qu’ils prennent effectivement au sérieux leurs obligations communes de vérification et de surveillance à leurs frontières extérieures, en abandonnant toute pratique incompatible.

Quoi qu’il en soit, l’application de l’article 29 permet de vouer un État Schengen à l’ostracisme et à l’exclusion, jusqu’à deux ans, de l’espace sans contrôle aux frontières intérieures. Il suffit que le Conseil recommande à tous les autres États Schengen de réintroduire le contrôle à leurs frontières communes avec le premier et que tous ces États décident de suivre une telle recommandation. Celle-ci jouera en pratique le rôle de mesure réactive de type sanctionnateur (voir en ce sens Henri Labayle, « Schengen : un espace dans l’impasse », Revue Europe, 2016).

Il convient de rappeler, par ailleurs, que dans le cadre du contentieux franco-italien de 2011 (concernant la réintroduction par la France des contrôles à sa frontière commune avec l’Italie, suite à l’arrivée d’un grand nombre de migrants en provenance de la Tunisie au territoire italien), le Conseil européen du 23-24 juin 2011, secondé par la Commission [COM(2011) 561 final], a proposé l’introduction dans le CFS d’une « clause de sauvegarde » d’application tendanciellement automatique « afin d’autoriser, à titre exceptionnel, le rétablissement des contrôles aux frontières intérieures en cas de situation véritablement critique, lorsqu’un Etat membre n’est plus en mesure de respecter ses obligations au titre des règles Schengen », quelle qu’en soit la cause (voir « Internal border controls in the Schengen area: Is Schengen crisis-proof? » Study for the LIBE committee, 2016, p. 23-25).

Ce n’est donc qu’en faisant une inexplicable table rase soit du contenu que peut revêtir une recommandation du Conseil basée sur l’article 29 CFS, soit des travaux préparatoires concernant cette disposition, que la Commission peut affirmer qu’une telle mesure « n’est pas une sanction à l’encontre d’un ou de plusieurs États membres, pas plus qu’elle ne vise à exclure quelque État membre que se soit de l’espace Schengen », mais simplement « une mesure visant à préserver le fonctionnement global» de cet espace [COM (2016)120 final].

L’application pratique de l’article 29 devient d’autant plus problématique que celui-ci ne fait pas une distinction que la « crise des migrants et des réfugiés sans précédent », survenue au cours de 2015, a rendu indispensable et urgente. Il s’agit de la distinction entre, d’une part, les « manquements graves persistants dans l’exécution du contrôle aux frontières extérieures », susceptibles d’être évités à travers d’une « diligence moyenne » de la part de l’État Schengen concerné et, d’autre part, les manquements du même type qui sont manifestement hors du contrôle de cet État, en raison notamment de l’ampleur, de l’intensité et de la persistance de la pression migratoire à ses frontières extérieures maritimes et terrestres, ainsi que de l’insuffisance au moins provisoirement insurmontable des moyens humains et matériels pour leur faire face.

Cette dernière situation, frôlant le cas de force majeure, a été vécue par la Grèce surtout à partir de septembre 2015 et au moins jusqu’à « la mise en œuvre initiale de la déclaration UE-Turquie du 18 mars 2016 ». Ce n’est donc nullement par hasard que la Grèce est le sixième État Schengen concerné para la décision d’exécution. Celle-ci lui impute des manquements graves persistants dans l’exécution des contrôles à ses frontières extérieures, certains desquels persistent et mettent en péril le fonctionnement global de l’espace Schengen.

À cet égard, la question essentielle de savoir si d’autres États Schengen ont pu manquer à leur devoir de solidarité vers la Grèce n’est même pas soulevée. La décision d’exécution se limite à relever que « du fait de sa situation géographique, la République hellénique est particulièrement touchée par [la crise des migrants et des réfugiés sans précédent] et a dû faire face à une augmentation spectaculaire du nombre de migrants arrivant sur les îles de la mer Égée », la zone la plus exposée (attendu 2). Le mot « spectaculaire » apparaît assez malheureux dans ce contexte. Le bon mot serait sans doute « incontrôlable ».

Cette critique d’ordre terminologique n’est sûrement pas la seule susceptible d’être adressée à la décision d’exécution, qui vise avant tout à légitimer, au regard du CFS, le maintien du contrôle, par les cinq États concernés, à leurs frontières intérieures au de-là du 12 mai 2016. Il convient surtout de vérifier si cette décision respecte intégralement les conditions prévues par l’article 29 CFS pour imputer en même temps à la Grèce des manquements graves persistants dans l’exécution du contrôle à ses frontières extérieures, représentant « une menace grave pour l’ordre public ou la sécurité intérieure» dans l’espace Schengen ou sur des tronçons de cet espace, qu’aucune autre mesure « ne peut effectivement atténuer ». Cette imputation constitue en effet une conditio sine qua non pour le maintien du contrôle aux frontières intérieures recommandé par le Conseil.

Or, c’est la décision d’exécution elle-même qui, dans son attendu 13, lance des doutes considérables à cet égard : « La République hellénique a accompli des progrès importants dans la correction de nombreux manquements que présente la gestion de ses frontières extérieures, constatés au cours de l’évaluation de novembre 2015. En outre, la mise en œuvre initiale de la déclaration UE-Turquie du 18 mars 2016 ainsi que les opérations en cours menées par Frontex et l’OTAN ont entraîné une diminution sensible du nombre de migrants en situation irrégulière et de demandeurs d’asile qui partent de la Turquie pour gagner la République hellénique. Cette réduction substantielle du flux de migrants en situation irrégulière et de demandeurs d’asile vers la République hellénique, ainsi que le soutien apporté par les agences de l’Union européenne et d’autres États membres dans les centres d’enregistrement, ont permis à la République hellénique d’améliorer sensiblement l’enregistrement des migrants en situation irrégulière et des demandeurs d’asile nouvellement arrivés » (italiques ajoutés).

Au vu de ces constatations – et si les mots employés par le Conseil ont un sens précis –, il n’est pas du tout évident que la conclusion puisse être celle qui est tirée à l’attendu 16 : « comme aucune autre mesure n’a pu effectivement atténuer la menace grave constatée, il s’ensuit que les conditions pour appliquer l’article 29 (…) en dernier recours sont remplies ». En effet, plusieurs mesures ont pu effectivement atténuer la dite « menace grave », même si, le cas échéant, elles ne l’auraient pas totalement éliminée. En tout état de cause, ce n’est pas cette dernière exigence que fait l’article 29, même si dans sa proposition de décision d’exécution [COM (2016) 275 final, p. 5] la Commission semble utiliser ces deux verbes comme synonymes.

Dans ce contexte, la décision d’exécution reproche encore à la Grèce le fait que la surveillance à sa frontière avec l’ancienne République yougoslave de Macédoine « n’est actuellement pas totalement conforme au code frontières Schengen » (attendu 14). Comment pourrait-elle l’être dans l’immédiat si ce pays tiers a carrément fermé cette frontière commune avec la Grèce pour des périodes considérables, en bloquant sur le territoire grec les migrants et requérants de protection internationale désireux de la franchir ?

Toujours selon la décision d’exécution, « Cette situation accroît le risque de mouvements secondaires de migrants vers d’autres États membres ». Toutefois, pour que « le risque persistant de mouvements secondaires de migrants en situation irrégulière » (que la décision d’exécution ne cherche même pas à démontrer,) puisse légitimer le maintien du contrôle aux frontières intérieures en cause, il faudrait également examiner au préalable, au regard du principe de proportionnalité. son « incidence probable » sur la libre circulation des personnes au sein de l’espace Schengen [article 30, paragraphe 1, lettre c), CFS]. C’est un exercice que la décision en cause a manifestement omis de faire, dans la mesure où elle a accepté sans plus les justifications fournies par les cinq États destinataires de la recommandation.

Par ailleurs, pour ce qui est de la légalité procédurale de la décision d’exécution, il semble également critiquable qu’elle ait été adoptée avant l’expiration du délai de trois mois prévu par l’article 16, paragraphe 4, du règlement nº 1053/2013 pour que, à la suite des recommandations de mesures correctives adoptées par le Conseil le 12 février 2016, la Grèce rende compte de la mise en oeuvre de son plan d’action. Ce délai expirerai précisément le 12 mai 2016. Toutefois, dès lors que la Grèce avait fait ce compte-rendu le 26 avril, la Commission a estimé qu’elle « ne devrait pas attendre l’expiration du délai de trois mois pour évaluer si, au 12 mai 2016, la situation persistera (…) ». Le 4 mai elle a présenté sa proposition de « recommandation annoncée » au Conseil.

Ces anticipations ne semblent pourtant pas compatibles au moins avec l’esprit de l’article 29 CFS – qui exige que l’État Schengen dont le contrôle à ses frontières extérieures a été considéré gravement défaillant puisse avoir une opportunité effective d’essayer de résoudre tous les problèmes jusqu’à l’expiration des délais prévus à cette fin (voir dans ce sens Steve Peers, EU Justice and Home Affairs Law, Volume I, Oxford : OUP, 2016, p. 113).

Il est certes vrai que, nonobstant la menace grave et nullement atténuée imputée illégalement à la Grèce, la décision d’exécution n’a pas recommandé aux cinq États concernés de réintroduire le contrôle à leurs frontières communes avec la Grèce, lesquelles ne comprennent pas des frontières terrestres. Cette décision ignore ainsi la relation directe établie, bien ou mal, par l’article 29 entre défaillances qualifiées dans les contrôles aux frontières extérieures d’un État Schengen et réintroduction/maintien, par d’autres, du contrôle aux frontières intérieures avec premier.

Il ne s’ensuit cependant pas que, dans cette mesure, la décision d’exécution doive être considérée aussi illégale. En effet, l’article 29 ne semble pas exiger que, en cas de constatation de manquements graves liés au contrôle des frontières extérieures, mettant en péril le fonctionnement global de l’espace Schengen, la réintroduction du contrôle aux frontières intérieures qui peut en découler, le cas échéant, doive comprendre forcément l’État qui a fait l’objet d’une telle constatation.

Quoi qu’il en soit, même si dans le cas d’espèce la Grèce n’a pas été exactement suspendue de l’espace Schengen voire vouée à l’ostracisme, il n’en reste pas moins qu’en passant, la décision d’exécution produit sur elle l’effet d’une flétrissure, avec tout ce que cela implique du point de vue de la confiance mutuelle, sans laquelle l’espace Schengen ne peut pas subsister.

Est cela le signe d’un temps où, s’agissant de la refondation de l’Union européenne souhaitée par certains, la Grèce ainsi que d’autres États membres n’auront pas de place en pied d’égalité, et la Commission, au lieu de maintenir son rôle essentiel de « gardienne des traités », assumera le rôle d’adjoint d’un Conseil « directorial » ? Ou (et) s’agit-t-il d’un signe d’un temps où l’Union européenne se permet déjà de traiter plus sévèrement les États membres dont les gouvernements ne reflètent pas intégralement des partis ouverts aux « grandes coalitions », au sens germanique de l’expression ?

Navigation de l’article


The text below is the final version of the EU Regulation on the European Border and Coast Guard as revised by the Jurist Linguists of the EU institutions. Formally adopted this week as a “corrigendum” by the European Parliament and by written procedure by the Council it will be published on the Official Journal in the coming weeks. Presented, negotiated and adopted in extremely short time ([1]) under the pressure of the European Council the new EU Regulation on the European Border and Coast Guard could be seen at the same time a main evolutionary step and a revolutionary one in the relation between the EU and its Member States in the freedom security and justice area. 

Even if the main subject of the text is the border management it covers also directly and indirectly other EU policies such as refugee law, international protection, migration and even internal and external security. Not surprisingly  such an ambitious objective was difficult if not impossible to achieve in such a short time and several commentators and representatives of the civil society have already considered (see Peers , Carrera [1], Rijpma [2], and, more recently, De Bruycker [3])  that the text on one side does not deliver what it announces and on the other side is still rooted in an old intergovernamental model. Maybe from a legistic point of view instead of bringing all these objectives in a single legislative text it would had been more elegant to focus its content only on the organisational and operational aspect of the “new” Frontex  and deal with the general framework of the integrated EU border management in the Schengen Border Code where general rules on the definition, negotiation adoption and implementation would had been better placed together with the rules on its evaluation and on the adoption of extraordinary measures in case of emergency. However these have probably been considered by the Commission legal niceties to be dealt with in times with less political pressure.. 

With so many objectives it is not surprising that the final result is far from the expectations and the text is somewhere still elusive and somewhere too detailed. It can then be interesting to  compare the negotiation position of the three institutions as it result from a very interesting Multicolumn document leaked by Statewatch during the “confidential” legislative trilogies. It shows that the European Parliament has tried to improve the original Commission proposal and has obtained some concessions from the Council but regrettably, it had lost the main targets such as the definition in codecision of the European Border Strategy (instead of a simple decision of the Agency’s Management Board) and even on the procedure to appoint of the Agency Director where its position will be to express an opinion …which can be disregarded.

Further comments will follow. EDC


[1] See the CEPS study of Sergio Carrera and Leonhard den Hertog “A European Border and Coast Guard: What’s in a name?”

[2] See Jorrit RIJPMA study for the Civil Liberties Committee of the EP “The proposal for a European Border and Coast Guard: evolution or revolution in external border management?”

[3] See Philippe DE BRUYCKER “The European Border and Coast Guard: A New Model Built on an Old Logic


It is the latest (and quite likely not the last) of a chain of legal texts by which the EU has tried in the recent years to legally frame the issue of human mobility and human security in the EU by taking in account the new legal framework after the entry into force of the Lisbon Treaty and of the EU Charter of fundamental rights.

[1] A rather detailed and updated collection of the legislative preparatory works can be found here :

[2] As as verified by the Jurist Linguist and endorsed by the EP according to art 231 of its Rules of procedure)


REGULATION (EU) 2016/…OF THE EUROPEAN PARLIAMENT AND OF THE COUNCIL of … on the European Border and Coast Guard and amending Regulation (EU) 2016/399 of the European Parliament and of the Council and repealing Regulation (EC) No 863/2007 of the European Parliament and of the Council, Council Regulation (EC) No 2007/2004 and Council Decision 2005/267/EC


Having regard to the Treaty on the Functioning of the European Union, and in particular Articles 77(2)(b) and (d) and Article 79(2)(c) thereof,

Having regard to the proposal from the European Commission,

After transmission of the draft legislative act to the national parliaments,

Having regard to the opinion of the European Economic and Social Committee1,

Acting in accordance with the ordinary legislative procedure2,

Whereas: Continue reading “WORTH READING : the final text of the EUROPEAN BORDER AND COAST GUARD REGULATION”

Will UK citizens have to pay to visit the EU after Brexit?


by Steve Peers

Following a Guardian article on Saturday, and the Home Secretary’s confirmation on Sunday, it’s clear that the EU is planning to institute some kind of Electronic System of Travel Authorisation (ESTA) in future, which could well apply to UK citizens visiting the EU after Brexit. I’ll examine the background, context and consequences in this post.


What is an ESTA?

First of all, let’s establish what an ESTA is not. It’s not a means of regulating longer-term migration as such, although there is an indirect link between long-term migration rules and ESTA systems, as discussed below. Rather it’s a means of regulating short-term visits for tourism or other reasons.

Nor is an ESTA a tourist visa. A lot of people have confused it with one, perhaps because a Guardian sub-editor initially put an inaccurate headline on the original story (I see the online headline has since been corrected). A tourist visa is a bigger hassle for visitors than an ESTA, since travellers must visit a consulate or pay an agency to handle their application. It entails higher fees and a longer waiting period, and probably a bigger risk of rejection.

During the Brexit referendum campaign, the prospect of a visa regime between the UK and EU was not raised by the Leave side generally. However, it was raised by a junior minister, Dominic Raab, and at the time I trashed the idea here. Since then, Theresa May has shown sufficient judgment to return Raab to the backbenches, so hopefully we have heard the last of this idea for a while.

So what is an ESTA? It’s a way of gathering travellers’ information in advance of travel, usually for citizens of countries subject to a visa waiver, for instance the USA and Japan. In fact, the best-known example of an ESTA is the American version, although there are several other countries with one.  If a traveller fails to complete an ESTA in advance of travel, they will likely be denied boarding or admission at the border.  The US version includes a fee for administration and tourism promotion. Usually the form is completed, and the fee paid, online. It’s recommended to complete the ESTA form several days in advance, although on my last trip to the USA, I did it just before dashing out of the house to catch my plane. (I am not suggesting this as best practice).

The EU context

The EU has been considering an ESTA for a while. It would form part of the Schengen system of standardised external border controls, which are paralleled by the abolition (in principle) of internal border controls between Schengen States. The Schengen states comprise all the EU countries except Ireland – although Romania, Bulgaria, Cyprus and Croatia do not fully participate yet – plus four non-EU Schengen associates (Norway, Iceland, Switzerland and Liechtenstein).

A key feature of EU law in this area is that the Schengen system interacts with EU free movement law. So because the UK and Ireland have signed up to the free movement of EU citizens as EU Member States, their citizens are fast-tracked across the Schengen external borders. The same is true of the Schengen associates, because they have all signed up to free movement of their citizens with the EU as well.

Other non-EU citizens are subject to more intensive checks at the Schengen external border, as set out in the Schengen Borders Code. There’s a simple reason for this: they don’t have an underlying right to stay in the country, whereas citizens of EU Member States and the Schengen associates do – subject to exceptions. There are also distinctions between non-EU countries: some (like the US or Canada) have a visa waiver from the Schengen countries, while others (like India and China) don’t.

An ESTA was first discussed in a Commission discussion paper back in 2008. This was followed up by a very detailed study in 2011 which recommended against the idea, after which the Commission dismissed it.  In 2013, the Commission decided instead to propose an entry-exit system, which would record the movements of non-EU citizens (besides the Schengen associates) into and out of the Schengen external borders. Discussions on that proposal moved slowly, and the Commission proposed a new version of it in spring 2016. The intention is to agree on this system by the end of the year, although it will take several years afterward to get the system up and running in practice.

At the same time, the Commission revived talk of a possible EU ESTA, in a discussion paper on EU information systems. This excited many Member States, as can be seen by a Dutch EU Council Presidency paper published in the spring, which argued that the system could be a quid pro quo for visa waivers with countries like Ukraine and Turkey. Now the idea is on the agenda for the summit of the ‘EU27’ (ie the EU without the UK) to be held this week. It is being pushed by France and Germany in particular. Surely only a cynic would link this to the upcoming elections in those countries…


Like the entry-exit system, an EU ESTA would take some time to set up. The details of how it would work would remain to be determined: the Commission is due to make a proposal this autumn, which would then be agreed by the Council (only Schengen States get a vote, so not the UK) and the European Parliament. So it might not follow the US model exactly, in terms of fees or the link to the broader border control system, or the two-year period of validity.

For one thing, some of the EU documents suggest an EU ESTA will apply at external land borders, whereas the US system does not. Also, some EU papers suggest an ESTA will be used as a method of screening people and denying them entry in advance, while others refer to it simply as generating information for border guards to use to speed up their work. It’s not clear whether an ESTA would apply to those UK citizens who live in the EU already, if they (for instance) visited the UK and then returned to France.

But it does seem very likely that it will apply to all non-EU countries which don’t have a treaty on free movement of citizens with the EU. This would follow the existing model of the Schengen Borders Code, the Schengen Information System (which includes data on non-EU citizens to be refused entry) and the proposed entry-exit system. It’s simply common sense: fast-track entry at the border for those who are not subject in principle to immigration controls, but scrutiny at the border (or in advance of it) for those who are.

It’s been suggested that the application of an EU ESTA to the UK would be an act of ‘spite’. This is simply ridiculous. If a country leaves the EU, it leaves behind both the pros and cons of membership. In short: divorce doesn’t come with bed privileges.

Many on the Leave side argued that the UK should leave the EU and then stop applying free movement law, so that it could exercise more control over EU citizens at the border. Applying an EU ESTA to UK citizens would just be exactly the same principle in reverse. Equally UK citizens would no longer be fast-tracked at Schengen external borders, would be subject to the EU entry-exit system and (for a few) would be listed in the Schengen Information System as people to be denied entry into any Schengen State. This isn’t ‘scaremongering’: it’s simply a description of existing and proposed EU law.

So will the UK be subject to an EU ESTA after Brexit? The obvious way to avoid it (and the other forms of stepped-up border control) would be to conclude a deal on free movement of persons with the EU (this need not mean joining Schengen). Arguably even a free movement deal with derogations – for instance, limiting the numbers of EU citizens who can work in the UK in some way – could justify an exemption from stepped-up border controls, as long as those UK controls are not applied at the border. I can foresee the counter-argument that ‘the EU will never negotiate an exception to free movement of people’; but has it occurred to anyone that this might simply be a negotiating position?

If an EU ESTA does end up being applied to UK citizens, the UK could reciprocate with a system of its own, applied to EU visitors. But this doesn’t rule out some form of deal on immigration flows between the UK and the EU, which could be agreed in return for continued UK participation in the single market.  The mere existence of a UK ESTA – perhaps accompanied by some other form of immigration safeguard on EU citizens – might arguably go some way to satisfying those who want additional border controls. It could be accompanied by further mutual sharing of data on serious convicted criminals, for use in the ESTA process. Latvia’s daft decision to release a convicted murderer after only a few years in prison should not have had tragic consequences in the UK, or any other Member State.

Extradition to non-EU countries: the limits imposed by EU citizenship


Steve Peers

One of the best-known EU laws created the European Arrest Warrant (EAW), which replaced the prior rules governing extradition between Member States. But on the other hand, in principle EU law has nothing to do with extradition to non-EU countries, except where the EU has agreed a treaty on this issue (as it has with the USA and Norway and Iceland), or as regards asylum-seekers (the EU’s asylum procedures law limits their extradition to their country of origin, because it’s necessary to determine first if the country which seeks to prosecute them is in fact persecuting them).

Yesterday’s CJEU judgment in Petruhhin altered this legal position. EU law does apply to such issues, and the Court clarified some relevant issues while leaving others open. Furthermore, the judgment raises the question of future UK/EU relations on extradition following Brexit.


Extradition between the EU and non-EU countries is governed by a combination of national law and bilateral and multilateral treaties – most notably the Council of Europe Convention on Extradition, which has been ratified by all 47 European countries plus three non-EU countries. There are four further Protocols to this Convention, which have been ratified by fewer States.

A key feature of extradition law is that in principle most States will not extradite their own citizens, although within the EU, the EAW law has overruled any absolute ban on surrendering nationals as between Member States. While the refusal to extradite citizens could run a risk of impunity if those citizens commit criminal offences in another country, most States avoid that risk by extending their criminal jurisdiction to cover acts of their citizens committed outside their territory. In fact many EU laws and international treaties require States to assert such extraterritorial jurisdiction as regards specific transnational crimes.

The EAW law says a little about possible conflict between EAWs and extradition requests from third countries. It states simply that in the case of such a conflict, the national authority should decide which takes priority ‘with due consideration of all the circumstances’, including the relevant treaty and ‘the relative seriousness and place of the offences’, the relevant dates of the requests and whether the extradition request or EAW aims to obtain custody of a fugitive for trial or to serve a sentence already imposed.

This compares with the original proposal for the EAW law, which always gave priority to an EAW if the extradition request came from a country which was not party to the Council of Europe Convention. That clause was dropped following intensive lobbying from the US government, while the law was being negotiated in autumn 2001 (the EAW law was largely motivated by the 9/11 terrorist attacks, although it is not limited to terrorist offences).


The Petruhhin case concerned an Estonian citizen sought by the Russian government for prosecution for organised crime offences. He was safe from extradition to Russia as long as he stayed in Estonia, since that country will not extradite its nationals outside the EU. But at one point he was arrested in Latvia, which decided to implement the Russian request. While Latvia also refuses to extradite its own citizens outside the EU, an Estonian citizen prima facie obviously cannot rely on that protection.

So Mr Petruhhin tried to rely on his transnational form of citizenship instead, arguing that since he was an EU citizen in another Member State, he was entitled to equal treatment with Latvians – therefore protecting him from extradition from Latvia to Russia, just like them. (Logically if his argument had worked, he would also be protected from extradition from any other Member State which refuses to extradite its own citizens to Russia).

The CJEU ruled first of all that the despite the absence of EU law on this issue, the dispute fell within the scope of EU free movement law, since Mr Petruhhin was exercising free movement rights. Therefore he had a right to equal treatment with nationals of Latvia in principle. However, a breach of that equality right could be justified on the grounds of avoiding impunity from prosecution for alleged criminal offences: Latvia, like most States, extends its criminal jurisdiction to cover acts of its own citizens abroad, but not the acts of citizens of other countries abroad. This distinction between the position of Latvians and citizens of other Member States can justify different treatment as regards protection from an extradition request.

Having said that, the Court added a crucial rider. To limit the effect of its ruling upon free movement rights (the proportionality principle), it ruled that Latvia has to contact the Estonian authorities first, to see if they wish to prosecute him there on the basis of theirextraterritorial jurisdiction, before handing him over to Russia. That’s an important proviso, as many people believe they are more likely to be treated fairly in the courts of their own State. At any rate, this likely means they will have access to defence and court proceedings in their own language, with any pre-trial detention closer to friends and family.

Finally, the Court stated that any extradition to Russia was subject to the ban on torture or other inhuman or degrading treatment set out in Article 4 of the EU Charter of Fundamental Rights, which transposes the standards set out in Article 3 ECHR and the relevant case-law of the European Court of Human Rights on this issue. That means that if there’s a real risk of such treatment in Russia, the fugitive cannot be extradited there in any case. It should be noted that the Charter equally bans extradition to face the death penalty.


First of all, the Court was correct to assert the link between EU free movement law and extradition to third States, although its rather abstract reasoning could be improved upon. The best argument supporting this part of the ruling is rather that Mr Petruhhin would have been deterred from leaving Estonia for another Member State if he ran a risk of being extradited to Russia every time he left the country.

Next, would EU law also apply to cases where a Member State considers extraditing itsown citizen to a non-EU country? The question may not arise often, since as noted already, many Member States don’t do this at all. But where they do, logically the case law on citizenship of the European Union (as distinct from free movement applies). As developed since the Zambrano judgment, this prevents citizens of their own Member State from being forced outside the EU in principle, as they are thereby deprived of EU citizenship. But logically the same limits apply by analogy: extradition of citizens can be justified on grounds of preventing immunity, but that is qualified if the extraditing State subjects its own citizens to prosecution for acts committed abroad (most do, as noted already).

Third, could there be other grounds justifying extradition to a non-EU state, besides preventing impunity? This isn’t clear from the judgment. But logically the judgment would apply by analogy to cases where a fugitive has already been sentenced to a term of imprisonment. In such cases, Latvia (say) would contact Estonia to see if the latter country could take over the punishment imposed by Russia, pursuant to the Council of Europe treaty on the transfer of prisoners or another relevant international treaty.

Could there be further grounds preventing extradition to a non-EU state, besides the Charter ban on the death penalty and torture et al, and the priority for EAWs? For instance, what if the person concerned has already been tried in a Member State, or in a third State? The EU has a cross-border ban on double jeopardy, but it only applies to Member States and Schengen associates, not to countries like Russia and the USA. Extradition treaties sometimes address this issue, but don’t always cover every double jeopardy scenario. A pending case before the CJEU should clarify this issue.

Next, logically the ruling would also apply by analogy if a third Member State could apply its jurisdiction: say an Estonian in Latvia was wanted by Russia but could potentially be prosecuted or serve a sentence in France, due to an [alleged] assault against a Russian citizen in France, or a French citizen in Russia. (Some countries assert criminal jurisdiction where one of their citizens was a victim of crime).

This brings us to the issue of conflict between an EAW and an extradition request from a non-EU state. The CJEU didn’t have to comment on this issue in its judgment, because no EAW had been issued yet. But the Court’s judgment necessarily means that there is more likely to be such a conflict in future, if Estonia indeed issues an EAW. And if that happens, the new judgment implies that the open-ended conflict rule in the EAW has to give way to the primary law of the Treaties: so the Estonian EAW has to take precedence over the Russian extradition request. The Court has in effect enshrined priority for EAWs over (almost) all non-EU extradition requests, whereas the original Commission proposal, as noted above, would have given such priority only over extradition requests from non-Council of Europe states. Perhaps the Americans should also have lobbied the Court of Justice.

But then, the USA has its own extradition deal already with the EU, as noted at the outset. (The deal with Norway and Iceland has not been ratified yet). The Court says several times in its judgment that the general rules it elaborates here are without prejudice to extradition treaties concluded between the EU and third countries. Presumably it can interpret the EU/USA treaty, since it can interpret any treaties which the EU signs with non-EU states. In fact, there’s a pending case before the CJEU which asks the Court to interpret this very treaty.


This judgment is probably relevant for Brexit. The UK government has recently hintedthat it will seek some continuation of criminal law cooperation with the EU. There will be transitional issues with EAWs pending on Brexit Day, which the EU/UK withdrawal treaty concluded under Article 50 TEU will hopefully address. In fact there are already possible complications arising from Brexit in this area, as there are several challenges in Irelandto the execution of UK EAWs on the grounds that Brexit is coming. The CJEU may well be called on to address these issues even before Brexit Day.

For the position after Brexit, it’s undoubtedly possible for the EU to conclude an extradition arrangement with the UK, as the Court’s judgment actively encourages the EU to sign such treaties. In fact, the judgment might arguably be the basis of an argument for EU exclusive competence over extradition treaties with non-EU countries, on the basis that any Member State agreements would affect the operation of the EAW law, at least as regards EU citizens. That would mean that the UK could no longer sign extradition deals with individual EU countries, but only with the EU as a whole.

If no deal were reached, the UK and EU could fall back on the Council of Europe extradition Convention. But as I have noted before, this would mean far less extradition (and much slower extradition) as compared to the EAW.

If there were a UK/EU deal, Member States may still want to refuse to extradite their own citizens to the UK, as they have under the treaty with Norway and Iceland. But even if they are willing to extradite them to the UK, on whatever treaty basis, it may be arguable on the basis of the new judgment that they can’t, as long as the fugitive can be tried or serve her sentence in the remaining EU. And although the UK can still assist an EU Member State in prosecuting its own citizens, that will be far more expensive for the UK authorities than trying the person in the UK.

Accord politique ou juridique : Quelle est la nature du “machin” conclu entre l’UE et la Turquie en matière d’asile?

ORIGINAL PUBLISHED HERE  on 10 Friday Jun 2016

Par Olivier Cortenet Marianne Dony, Professeurs ordinaires à l’Université libre de Bruxelles,

Alors que trois demandeurs demandeurs d’asile – apparemment deux Pakistanais et un Afghan dans les affaires T-192/16, 193/16 et 257/26 – ont demandé au Tribunal de l’Union Européenne l’annulation de l’accord conclu le 18 mars 2016 entre l’UE et la Turquie, il est permis de s’interroger sur la nature exacte de ce “machin” considéré par le service juridique du Parlement Européen comme un simple accord politique, sachant cependant que la recevabilité du recours sera tout d’abord au coeur des débats…

Les autorités européennes affirment à l’unisson que la « déclaration UE-Turquie », dont le contenu a été détaillé dans un communiqué de presse du Conseil européen du 18 mars dernier, n’est pas un accord international mais une simple déclaration d’intention. Qu’en est-il vraiment, au regard des règles du droit de l’Union européenne et du droit international public ?

Si l’on s’en réfère au droit de l’Union européenne, il est permis d’en douter au vu de la jurisprudence de la Cour de justice (de l’Union européenne).

Ainsi, dans un arrêt du 23 mars 2004, C-233/02, France c. Commission (points 42 à 45), la Cour s’est interrogée sur le point de savoir si des « lignes directrices » finalisées en février 2002 par communication entre les négociateurs des services de la Commission et leurs homologues américains et sur lesquelles aucune signature n’avait été apposée pouvaient être considérées comme un accord international. La Cour a indiqué que le critère décisif était de savoir si ces lignes directrices avaient, ou non, force obligatoire et qu’à cette fin, il fallait s’en référer à l’intention des parties. En l’espèce, la Cour a alors constaté qu’il résulte explicitement du texte de ces lignes directrices que les parties avaient l’intention de les appliquer sur une base volontaire et que de surcroît l’intention des parties de ne pas contracter d’engagements obligatoires avait été à maintes reprises expressément réitérée durant la phase de négociations des lignes directrices. C’est sur cette base que la Cour conclut que ces lignes directrices ne constituent pas un accord international et ne sont donc pas visées par l’article 300 CE, devenu article 218 TFUE.

Par ailleurs, dans l’arrêt du 26 novembre 2014, C-103/12 et C-165/12, Parlement et Commission c. Conseil (points 60 à 74) la Cour a analysé le contenu et le but d’une déclaration relative à l’attribution de possibilités de pêche dans les eaux de l’Union européenne à des navires de pêche battant pavillon » de la République du Venezuela pour considérer qu’elle devait s’analyser comme une offre adressée à cette dernière, subordonnée au respect de certaines conditions précises et que la République du Venezuela  en transmettant des demandes d’autorisation de pêche dans les eaux concernées avait consenti à cette offre. Elle a dès lors conclu qu’un accord avait bien été conclu entre ces dernières, en ajoutant que « le fait qu’un tel accord est formalisé dans un seul document commun ou dans deux ou plusieurs instruments écrits connexes est dépourvu de pertinence ».

Il résulte donc clairement de ces arrêts que la forme n’importe pas. Ce n’est pas parce que le choix d’une déclaration, ou d’un communiqué de presse, a été fait qu’il ne peut s’agir d’un accord international. Au contraire, il faut analyser le but et le contenu de cette déclaration pour déterminer si elle contient des engagements ayant force obligatoire.

Et, en passant à l’analyse de la déclaration UE-Turquie, une lecture même rapide permet de relever que l’UE et la Turquie « sont convenues » d’un certain nombre d’actions pour atteindre un objectif commun. Au nombre de ces actions:

  • il est d’abord prévu que « tous les nouveaux migrants en situation irrégulière qui partent de la Turquie pour gagner les îles grecques à partir du 20 mars 2016 seront renvoyés en Turquie », étant entendu que tous les « coûts des opérations de retour des migrants en situation irrégulière seront pris en charge par l’UE » ;
  • ensuite, « pour chaque Syrien renvoyé en Turquie au départ des îles grecques, un autre Syrien sera réinstallé de la Turquie vers l’UE », un mécanisme devant, à cette fin, être « mis en place, avec le soutien de la Commission, des agences de l’UE et d’autres États membres » ;
  • de plus, la Turquie « prendra toutes les mesures nécessaires pour éviter que de nouvelles routes de migration irrégulière, maritimes ou terrestres, ne s’ouvrent au départ de son territoire en direction de l’UE »  et l’UE « accélèrera (…) le versement du montant de trois milliards d’euros initialement alloué au titre de la facilité en faveur des réfugiés en Turquie ».
  • Finalement la déclaration affirme que tous « ces éléments progresseront en parallèle et feront l’objet d’un suivi mensuel mené conjointement ».
  • A titre accessoire, on mentionnera que le considérant 4 de la proposition de décision modifiant la décision du Conseil du 22 septembre 2015 instituant des mesures provisoires en matière de protection internationale au profit de l’Italie et de la Grèce, déposée par la Commission le 21 mars 2016, indique que « les chefs d’État ou de gouvernement sont convenus, le 7 mars, d’une série de principes devant constituer la base d’un accord avec la Turquie … », que la proposition entend partiellement mettre en œuvre. Nous noterons que c’est la Commission elle-même qui a souligné ces termes dans la présentation de sa proposition.

L’ensemble de ces éléments est de nature à indiquer que la déclaration ne contient pas simplement un certain nombre d’actions que les parties entendent appliquer sur une base volontaire mais bien des engagements à caractère obligatoire. On peut donc conclure que cette déclaration constitue bien en réalité un accord international.

Conséquences en droit européen

La conséquence en est que cette déclaration relève de l’article 218 TFUE. En effet, ainsi que l’a indiqué expressément la Cour dans son arrêt du 26 novembre 2014 (point 83), cet article « régit la négociation et la conclusion des accords entre l’Union et des pays tiers », étant entendu que le terme « accord » utilisé à cet article « doit être compris dans un sens général, pour désigner tout engagement pris par des sujets de droit international et ayant une force obligatoire, quelle qu’en soit la qualification formelle ».

En outre, comme l’a souligné la Cour de justice dans un arrêt du 16 juillet 2015C-425/13, Commission c. Conseil (point 62), cette disposition « constitue une norme autonome et générale de portée constitutionnelle, en ce qu’elle attribue aux institutions de l’Union des compétences déterminées », en « visant à établir un équilibre entre ces dernières ».

Que prévoit cette disposition? (voy. à ce sujet, M. Dony,   Droit de l’Union européenne, points 358 à 368). Il résulte de l’article 218, paragraphe 6, TFUE, que c’est le Conseil, sur proposition du négociateur (qui est normalement la Commission), qui adopte, en principe à la majorité qualifiée, une décision portant conclusion de l’accord, et ce après approbation du Parlement européen, lorsque cet accord couvre un domaine auxquels s’applique la procédure législative ordinaire, ce qui est le cas pour la politique d’asile (article 78, paragraphe 2, TFUE) et d’immigration (article 79, paragraphe 2, TFUE).

Il est patent qu’en l’espèce, cette procédure n’a absolument pas été respectée : en effet, il n’y a pas trace d’une quelconque proposition de la Commission ; la « déclaration UE-Turquie » a fait l’objet d’un communiqué du Conseil européen et non d’une décision du Conseil et enfin, last but not least, le Parlement européen, non seulement n’a pas approuvé l’accord, mais n’a tout simplement pas été impliqué du tout. Il est même permis de penser que le terme « déclaration » a été consciemment choisi pour notamment éluder l’application de la procédure de l’article 218 TFUE.

En vertu d’une jurisprudence constante, le non-respect de la procédure de l’article 218 TFUE est de nature à affecter la légalité de la décision de conclusion d’un accord. C’est ainsi que la Cour, dans son arrêt précité du 26 novembre 2014, a annulé la décision adoptant la déclaration en cause au motif que le Parlement avait été simplement consulté et n’avait pas donné son approbation à l’accord. De même, dans un arrêt du 9 août 1994, C-327/91, Conseil c. Commission, la Cour a annulé l’acte par lequel la Commission avait entendu conclure un accord avec les Etats-Unis concernant l’application de leur droit de la concurrence, après avoir jugé que l’accord aurait dû être conclu par le Conseil, qui seul détient la compétence de conclusion des accords internationaux.

Cela étant, deux difficultés se présentent dans le cadre du droit de l’Union européenne:

D’abord, la question se pose de savoir comment il est possible de contester la validité du communiqué de presse contenant la « déclaration UE-Turquie », qui doit être considérée comme « la décision de conclusion » de cet accord.

La voie à laquelle chacun pense spontanément est l’action en annulation prévue par l’article 263 TFUE. La forme qu’a prise cette « décision » n’est pas un obstacle, car il est de jurisprudence constante que « recours en annulation doit être ouvert à l’égard de toutes les dispositions prises par les institutions de l’Union, indépendamment de leur nature ou de leur forme, à condition qu’elles visent à produire des effets de droit », comme l’a rappelé la Cour de justice dans son arrêt précité du 16 juillet 2015 (point 26).

Ce recours est ouvert sans conditions aux Etats membres, au Parlement européen, au Conseil ou à la Commission. En revanche, les personnes physiques ou morales doivent quant à elle établir que l’acte dont elles demandent l’annulation les concerne « directement et individuellement » et il paraît très difficile voire téméraire de soutenir que tel pourrait être le cas d’un tel acte. Elles devraient donc s’en remettre à une action introduite par un des requérants dits privilégiés, ou plus précisément au Parlement européen ou un Etat membre.

Une voie plus indirecte, et plus incertaine, pourrait être envisagée, qui consisterait, en application de l’article 267 TFUE, à saisir une juridiction nationale et de lui demander de poser à la Cour de justice une question préjudicielle en appréciation de validité de cette « décision ».

Ensuite, il résulte d’une jurisprudence constante que la Cour de justice peut seulement annuler ou invalider la décision de conclusion d’un accord international mais non l’accord lui-même. Ainsi, dans l’arrêt précité du 9 août 1994, la Cour a interprété le recours en annulation de l’accord introduit par la France comme étant dirigé contre l’acte par lequel la Commission a entendu conclure cet accord. La question de la conséquence d’une telle annulation sur la validité de l’accord international doit être abordée au seul regard du droit international public, vers lequel il importe donc de se tourner.

Conséquences en droit international

A cet égard, en droit international public, la notion de traité est entendue assez largement, comme en témoigne cette définition codifiant le droit coutumier reprise dans la Convention de Vienne de 1986 sur le droit des traités entre Etats et organisations internationales ou entre organisations internationales : « […] l’expression ‘traité’ s’entend d’un accord international régi par le droit international et conclu par écrit […] que cet accord soit consigné dans un instrument unique ou dans deux ou plusieurs instruments connexes, et quelle que soit sa dénomination particulière » (voy. les commentaires de cet article et des autres dispositions pertinentes des conventions de Vienne sur le droit des traités dans O. Corten et P. Klein (dir.), Les Conventions de Vienne sur le droit des traités. Commentaire article par article).

La dénomination, ou plus généralement la forme de l’accord, n’importe donc pas. Ont ainsi été considérés comme des traités entre Etats (mais on peut sans aucun doute transposer ces enseignements aux traités entre Etats et organisations internationales) : un échange de lettres (C.I.J., Affaire du Différend territorial (Libye/Tchad), Recueil 1994, not., § 31) , un simple procès-verbal de réunion (C.I.J., Affaire de la Délimitation maritime et des questions territoriales entre le Qatar et Bahreïn, Recueil 1994, § 23.), un communiqué conjoint (C.I.J., Affaire du Plateau continental de la Mer Egée, Recueil 1978, §§ 95-98) ou encore … une déclaration commune (C.I.J., Affaire de la Frontière terrestre et maritime entre le Cameroun et le Nigéria, Recueil 2002, § 263). Le caractère informel de la « déclaration UE-Turquie » ne constitue donc aucunement un obstacle à sa qualification de traité. Le droit international et le droit européen se rejoignent sur ce point.

La seule condition potentiellement problématique en l’espèce est celle selon laquelle l’accord doit être « régi par le droit international », ce qui suppose une volonté des parties de produire des engagements juridiques relevant du droit international public. Ici aussi, les critères du droit international sont très similaires à ceux dégagés par la Cour de justice qui d’ailleurs s’y réfère explicitement.

En l’espèce, et comme indiqué ci-dessus (point 4), la « déclaration UE-Turquie » a pour objet la circulation des personnes et le statut de réfugié, un domaine qui relève indéniablement de l’ordre juridique international. La terminologie utilisée témoigne d’ailleurs d’une volonté de s’engager, comme en témoignent les extraits suivants : « l’UE et la Turquie ont décidé ce jour de … »,  « sontconvenues des points d’action complémentaires suivants… »… « La Turquieprendra toutes les mesures nécessaires pour éviter que… » ; … « un programme d’admission humanitaire volontaire sera activé. Les États membres de l’UE ycontribueront… ». Les termes ainsi soulignés ont une portée qui, loin d’être simplement exhortative ou indicative, est normative et prescriptive. De manière tout à fait explicite, les parties précisent encore que : « Cela se fera en totale conformité avec le droit de l’UE et le droit international » ou encore « conformément aux normes internationales applicables ».

Bref, au vu des termes mêmes de cette déclaration, il paraît difficile de lui dénier la qualité de « traité », au sens du droit international public. Cela étant, le non-respect des procédures internes de l’UE lors de la conclusion de ce traité pourrait avoir des conséquences non sur l’existence, mais sur la validité de ce dernier. Le droit coutumier en la matière est exprimé dans deux dispositions de la Convention précitée de Vienne de 1986 :

« Article 27 […]

2. Une organisation internationale partie à un traité ne peut invoquer les règles de l’organisation comme justifiant la non-exécution du traité.

3. Les règles énoncées dans les paragraphes précédents sont sans préjudice de l’article 46 ».

« Article 46 […]

2. Le fait que le consentement d’une organisation internationale à être liée par un traité a été exprimé en violation des règles de l’organisation concernant la compétence pour conclure des traités ne peut être invoqué par cette organisation comme viciant son consentement, à moins que cette violation n’ait été manifeste et ne concerne une règle d’importance fondamentale.

3. Une violation est manifeste si elle est objectivement évidente pour tout Etat ou toute organisation internationale se comportant en la matière conformément à la pratique habituelle des Etats et, le cas échéant, des organisations internationales et de bonne foi ».

Il en découle que, si, en principe, l’UE ne pourrait se prévaloir de la violation de ses propres règles pour justifier la non-exécution du traité, une exception est cependant réservée aux cas de violation « manifeste » concernant une règle d’une « importance fondamentale ».

Pour ce qui est de la deuxième condition,  l’accord a été conclu  en violation totale du prescrit de l’article 218 TFUE (point 7). Or cette disposition doit indéniablement être considérée comme une règle d’une importance fondamentale. La Cour internationale de justice a ainsi indiqué que  « les règles relatives au pouvoir de signer des traités au nom d’un Etat sont des règles constitutionnelles d’une importance fondamentale » (Affaire de la Frontière terrestre et maritime entre le Cameroun et le Nigéria, précitée, § 265) et cette affirmation peut assurément être transposée au droit d’une organisation internationale comme l’UE, d’autant que, comme déjà indiqué, la Cour de justice de l’Union européenne lui a expressément accordé ce statut. On peut d’ailleurs penser que c’est ce qui explique que le Conseil européen ait préféré dénier la qualité d’accord à la déclaration commune avec la Turquie : par le biais de cette qualification, sans doute espérait-il échapper à l’obligation de respecter l’article 218 TFUE…

Peut-on également considérer que cette violation était « manifeste » et donc « objectivement évidente » pour la Turquie ? Dans l’affaire que l’on vient de citer, la Cour internationale de justice est restée très prudente sur cette question, puisqu’elle estime que la connaissance par un Etat d’une règle constitutionnelle de droit interne d’un autre Etat suppose que ce dernier l’ait « rendu[e] publique de manière appropriée ».

Dans notre cas, on n’est cependant pas devant une règle de droit interne, mais devant des règles conventionnelles qui sont par définition publiques, les traités européens faisant l’objet de diverses publications, écrites et électroniques. Plus précisément, il est difficile d’imaginer que la Turquie, candidate à l’adhésion et engagée dans des négociations à cet effet depuis des années, n’ait pas connaissance des principes régissant la conclusion des accords internationaux de l’Union et en particulier des principes selon lesquels le Parlement européen doit être impliqué dans la procédure de conclusion de ces accords et que le Conseil européen n’est  en revanche pas habilité à y intervenir.

En conclusion

Dans ces conditions, et toujours si l’on s’en tient au droit international public, il apparaît que la « déclaration UE-Turquie » peut a priori être qualifiée de traité, mais un traité dont il est sérieusement permis de douter de la validité. Rien ne devrait dès lors sérieusement empêcher de remettre en cause les effets juridiques de cette déclaration, que ce soit en droit européen ou en droit international public.

The Orbanisation of EU asylum law: the latest EU asylum proposals


by Steve Peers

There have been a number of EU proposals to deal with the perceived ‘refugee crisis’ in Europe over the last year. The latest batch, issued this week, are perhaps the most significant to date. They concern three related issues: visas (notably a short-term Schengen visa waiver for Turkish nationals); Schengen (partly suspending the open borders rules for six months); and asylum (changing the Dublin system on responsibility for asylum seekers, and creating a new EU asylum agency). Further proposals on legal migration and other EU asylum laws are coming in the months ahead.

Essentially, these proposals amount to the ‘Orbanisation’ of EU asylum law. They copy and entrench across the EU the key elements of the Hungarian government’s policy, which was initially criticized: refusing essentially all asylum-seekers at the external border and treating them as harshly as possible so as to maintain the Schengen open borders system.


The surge in the number of refugees and migrants coming into the EU since 2014 led initially to a discordant response from Member States, with Germany and Sweden initially welcoming the arrivals and Hungary trying to stop them. Last September, in a bid to modestly assist the ‘frontline’ border states of Greece and Italy with the large numbers of asylum-seekers, the EU adopted two Decisions on ‘relocation’ (discussed here), in principle taking up to 160,000 asylum-seekers off those countries’ hands and distributing them among other Member States. However, this ‘Plan A’ was ineffective, as some Member States refused to cooperate (even launching legal action) and the remainder relocated very few people.

So ‘Plan B’ was developed: an EU/Turkey deal whereby Turkey either prevented the large number of refugees on its territory from leaving, or readmitted them back from the EU if they did reach EU territory (which in practice usually means the Greek islands). To implement this, Greece agreed to treat Turkey as a ‘safe third country’ or a ‘first country of asylum’ under EU asylum law, with the result that claims were treated as inadmissible. As discussed earlier on this blog, this is a highly dubious interpretation of the law. To induce Turkey to cooperate, the EU agreed to spend money on the welfare of Syrian refugees in Turkey, and to drop the short-term visa requirement for Turkish citizens to visit the EU countries in the Schengen system. (It also agreed to open one more ‘negotiating chapter’ relating to Turkish accession to the EU, but this is a trivial concession: only one of these 35 chapters has been agreed to date, in 11 years of accession negotiations).

In the meantime, many Member States became concerned about the numbers of migrants and refugees reaching their territories, and so resumed checks on the previously open borders between Schengen states. However, under the relevant Schengen rules dating from 2013 (on which, see my thinktank report on the Schengen system here), the authority to do this will soon expire, unless the EU as a whole agrees to suspend the Schengen system for one or more periods of six months. This prospect has been mooted since December 2015 (as discussed in detail here).

So this week’s proposals aim to implement and entrench these policy developments: waiving the visa requirement for Turkey; allowing a limited suspension of Schengen; and amending the Dublin system to reflect the EU/Turkey agreement, to deter asylum-seekers from moving between Member States (allowing Schengen to be fully reinstated) and to incorporate a new version of the failing relocation rules.  All of these measures are related, but I will examine each of them in turn.


There are three separate proposals to amend the EU visa list. All of them need to be agreed by the European Parliament, as well as a qualified majority of participating Member States in the EU Council.  The proposals, if adopted, would not apply to the UK and Ireland, which have their own laws on visa requirements (or waivers) for non-EU countries, due to an opt-out from the EU’s visa laws. That opt-out forms part of those countries’ overall opt-out from the Schengen system, which allows the UK to check people at its borders and refuse entry to non-EU citizens based (mostly) on UK law. It is therefore dishonest to suggest that the proposals would lead to an increased migrant influx into the UK. Indeed the UK’s withdrawal from the EU would not change the rules at all as regards non-EU citizens seeking to enter the UK from (the rest of) the EU – other than the small minority who apply for asylum or who are family members of EU citizens.

These proposals would, in turn: a) waive visa requirements for Turkish citizens; b) waive visa requirements for Kosovo; and c) make it easier to reimpose visa requirements in the event of immigration control issues. It should be noted that the Commission also recently proposed to waive visa requirements for Ukraine and Georgia; those proposals are still under discussion. All these proposals would, if adopted, amend the EU’s main law on visa lists, which dates initially from 2001. That law has been amended many times since, without any official codification of those amendments, but I have codified it unofficially here. Note that the visa waiver would apply to Turkish citizens, not to Turkish residents like the refugees from other countries living there.

The visa waiver for Kosovo is not linked to the overall refugee crisis, but rather to the policy of strengthening relations with EU neighbours, in part as an incentive for them to settle their own disputes. The Commission report on Kosovo fulfilling the requirements for visa waivers refers in particular to a recent border agreement between Kosovo and Montenegro. It also refers to meeting the requirements as regards readmission, reintegration, document security and organised crime.

As for Turkey, there is obviously a direct link with the EU/Turkey refugee deal. A fast-track visa waiver was promised to Turkey as part of that deal. But it is still subject to Turkey meeting the EU’s conditions. According to the Commission’s report, Turkey meets all but 7 of 72 requirements: the exceptions relate to issues like readmission, corruption, terrorism and document security, and the Commission believes that they will be fulfilled by the time the visa waiver is granted. In any event, the document security point is addressed by limiting the visa waiver to those with biometric passports.

A longer staff working document elaborates on this assessment, but it is not convincing on several points. As regards asylum issues, it states that the obligation to lift the geographical limitation on the Geneva Refugee Convention (which means that Turkey only fully recognises Europeans as refugees) is met by Turkey because that country treats non-Europeans just as well as if they are refugees. But it skips over the lack of work permits for refugees who are not Syrians. It also concludes that Turkey does not refoule refugees to dangerous countries (as alleged by NGOs) simply by accepting Turkey’s word to the contrary. The Commission also waives the obligation for Turkey to ratify Protocol 7 to the European Convention on Human Rights, on the grounds that its national law offers equivalent protection. But if so, why be afraid of the supervision of the European Court of Human Rights on these issues? And it is only clear reading the staff working document that the (unresolved) concerns about ‘terrorism’ laws are actually concerns about misuse of terrorism law to crack down on freedom of expression. The main report does not even flag this as one of the most significant concerns. And the existence of these concerns gives the lie to the Commission’s argument (in an earlier proposal, still under discussion) that human rights in Turkey are so well protected as to classify Turkey as a ‘safe country of origin’ for asylum purposes.

The proposal to reimpose visa requirements more easily is implicitly linked to the Turkish visa waiver proposal, although in fact it could apply to any State on the visa waiver list (the ‘white list’). The current rules, dating from 2013, allow ‘emergency’ reimposition of a visa requirement by the EU Commission for a six-month period, renewable for another six months if the Commission proposes to amend the law to make this permanent. This temporary Commission decision can be blocked by Member States, but does not need the approval of the European Parliament. The grounds for it are ‘sudden and substantial’ increases in irregular migration, rejected asylum applications or rejected readmission applications from the country concerned.

There are some further details of these rules in the preamble to the 2013 law.  A ‘substantial’ increase is an increase above 50%, and a low rate of recognition of asylum applications constitutes 3% or 4%, although in either the Commission could choose to use a different number.  Reimposition of visas is not automatic: there is a diplomatic phase during which the Commission talks to the officials of the other country and warns them to take action in light of the impending threat.  The Commission will only propose reimposition if it is not satisfied with the outcome of these talks. So far it has not done so.

Basically the new proposal would make it easier to reimpose visas in several ways.

First of all, it would no longer be an ‘emergency’ or ‘last resort’ decision, and the increases in irregular migration, rejected asylum applications or rejected readmission applications would no longer have to be ‘sudden’.

Secondly, the reference period for examining the increased irregular migration, etc would no longer be over six months, but over two months.

Third, the increase in asylum applications would no longer have to lead to ‘specific pressure’ on asylum systems; so there would need not be a large absolute number of asylum applicants from the country concerned, just a large relative increase in the number of applications.

Fourth, the rejected readmission applications would relate not only to citizens of the country concerned, but also to citizens of other countries who transited through that State’s territory. This is obviously aimed at enforcing the key feature of the EU/Turkey plan: the readmission of refugees to Turkey.

Fifth, the possibility of triggering reimposition of visas as compared to the period before the visa requirement was dropped would now apply indefinitely, and would no longer expire after seven years. The immediate impact of this change would be on Western Balkans countries, where (apart from Kosovo) the EU waived visa requirements in 2009 and 2010.

Sixth, the Commission can trigger the clause, not just Member States. It could act on the same grounds plus an additional ground of failure to apply a readmission deal with the EU as a whole.

Again, the final point aims at enforcing the EU/Turkey refugee deal. If Turkey does stop readmitting refugees, the EU can swiftly react by reimposing visa requirements. This works both ways, of course: if the EU threatens to reimpose visas on Turkish citizens on some other ground, such as an increase in Turkish citizens overstaying without authorization, then Turkey will likely refuse to take back refugees. Indeed, as discussed above, Turkey is threatening to do this if the EU does not waive the visa requirements in the first place – which accounts for the EU’s haste on this point.

Finally, a side issue (relating only to Turkey) is worth discussing. The EU/Turkey association agreement has a Protocol, signed in 1970, that sets a standstill on the free movement of services and freedom of establishment. That means the EU and its Member States can’t make the rules on these issues stricter than they were when the Protocol was signed. The CJEU has also ruled that if the rules are made more liberal than when the Protocol was signed, they can’t be made less liberal after that point without violating the standstill (Toprak and Oguz). While the standstill rule doesn’t apply to tourist visas (Demirkan), it does apply to visas for short-term economic activity (Soysal).

So would the standstill rule in the association agreement prevent the EU from reimposing visas for economic activity by Turkish citizens? In its case law (see most recently Genc, discussed here), the CJEU has said that the standstill rule can be overridden on public interest grounds. So far the case law on this point has concerned integration of family members, although it could also be argued that the objective of preventing irregular migration is also a valid ground to override the standstill. In fact, the CJEU has been asked whether migration control objectives can override it, in the pending case of Tekdemir. However, this case won’t be decided until well after June (when Turkey wants the visa waiver in place); and like the earlier cases, it concerns legal migration.


The idea of suspending Schengen for up to two years was originally mooted back in December – as I discussed in detail at the time. The mechanics of the process, as I detailed there, have been grinding away for some time. Now we have nearly reached the final stage: a Commission Recommendation for a Council Recommendation to suspend Schengen. Once the Council adopts this (by a qualified majority of Schengen states), the suspension can go ahead.

However, the Commission has tried to limit this suspension in time and in space. It would only apply to Germany, Austria, Sweden, Denmark and Norway (where the unilateral authority to suspend border controls is about to expire), and only for an initial period of six months. The Commission argues that the tightening of EU immigration and asylum law should have had sufficient effect by then, so a further suspension would not be justified. Time will tell if this is true: the Schengen rules allow for three six-month extensions of the initial suspension.

For legal reasons, as I discussed in the earlier blog post, the suspension has to be based on blaming a Member State for insufficient control of its external borders. Obviously, the Commission has named Greece. But it has warm words for Greece’s efforts in the last few months, and flights to and from Greece to the Schengen zone will not be affected. This rather measured and proportionate approach contrasts with the Commission’s asylum proposals – to which we now turn.


Again, there are three separate proposals, all of which need to be agreed by the European Parliament, as well as a qualified majority of participating Member States in the EU Council.  First of all, the current Dublin III Regulation, which sets out rules determining which Member State is responsible for an asylum application, would be replaced by a new Regulation – which I will call ‘Dublin IV’. Secondly, the current Eurodac Regulation, which supplements the Dublin Regulation by providing for the storage and comparison fingerprints of asylum-seekers and those who crossed the border irregularly, will also be replaced by a new Eurodac Regulation. Thirdly, thecurrent law establishing an EU agency known as EASO (the European Asylum Support Office), would be replaced by a new law creating an EU Agency for Asylum (the ‘EU Asylum Agency’).

This is just one batch of proposals: as the previous Commission communication from April (discussed here) set out, it will also soon propose new laws to amend the existing laws on qualification (definition) of refugees and people needing subsidiary protection status, asylum procedures, and reception conditions for asylum-seekers. In effect, this will amount to a third phase of the Common European Asylum System.

Currently, the UK and Ireland have opted in to the EU laws regarding Dublin, Eurodac and EASO. They opted out of the second-phase asylum Directives, but are covered by the first-phase Directives (except Ireland never opted in to the first-phase reception conditions Directive). Denmark and the Schengen associates (Norway, Iceland, Switzerland and Liechtenstein) participate in these laws on the basis of treaties with the EU. It would be up to the UK and Ireland to decide whether to participate in the new proposals; if not, the current Regulations continue to apply. If they opt out of the discussions on the proposals, they could still opt in later after adoption of the legislation, if they find that the final result is more to their liking than they had feared at the outset. Denmark and the Schengen associates could refuse to participate, but in that case their treaties with the EU will automatically terminate.

In the event of Brexit, the UK would no longer be subject to any of the EU asylum laws it is now participating in, unless the EU and the UK negotiate an agreement to that effect. It should be noted that the EU has in practice only ever been willing to extend the Dublin rules to non-EU States if those States are also Schengen associates. (Indeed in some cases, the Dublin and Schengen association treaties have been negotiated as a package).

The EU Asylum Agency

I will start with the least contentious of the new proposals.

Currently, EASO has a number of practical cooperation tasks. In particular, it must: ‘organise, promote and coordinate’ the exchange of information and identify and pool good practice, as well as activities relating to country-of-origin information (ie, information about conditions in asylum seekers’ countries of origin), including gathering and analysis of that information and drafting reports on that information; assist with the voluntary transfer of persons granted international protection status within the EU; support training for national administrations and courts, including the development of an EU asylum curriculum; and coordinate and exchange information on the operation of EU external asylum measures. For Member States under ‘particular pressure’, the Office must gather information concerning possible emergency measures, set up an early warning system to alert Member States to mass influxes of asylum seekers, help such Member States to analyse asylum applications and establish reception conditions, and set up ‘asylum teams’.

For its contribution to the implementation of the Common European Asylum System, the Office gathers information on national authorities’ application of EU asylum law, as well as national legislation and case law on asylum issues. It also draws up an annual report on the situation regarding asylum in the EU. At the request of the Commission, the Office may draw up ‘technical documents on the implementation of the asylum instruments of the Union, including guidelines and operating manuals.’ The Office can also deploy ‘asylum support teams’ on the territory of a requesting Member State, in order to provide ‘in particular expertise in relation to interpreting services, information on countries of origin and knowledge of the handling and management of asylum cases’.

How would the EU Asylum Agency be different? As with the parallel proposal for a European Border Guard (discussed here), the Agency would not replace national administrations, but play a bigger role coordinating them.  The main changes are: an obligation to exchange information with the Agency; a stronger role in analysis of the situation of countries of origin, including advice on alleged ‘safe countries of origin’; the development of guidance on applying EU asylum law; monitoring of the Common European Asylum System, including the capacity of Member States to apply it; and increased operational and technical assistance for Member States. An indication of the bigger role for the Agency as compared to EASO will be the planned increase in staff – from about 150 to around 500.


The current Regulation requires Member States to take the fingerprints of all asylum-seekers and irregular border crossers over 14 years old. This information is then stored in the Eurodac computer system. Every asylum-seeker’s fingerprints are compared with those already in the system, to see if he or she has either applied for asylum already or crossed the border irregularly. This is taken as evidence as regards which Member State is responsible for the asylum application under the Dublin rules.

Eurodac can also be used for other purposes. In 2013, the Eurodac law was revised to give police forces and the EU police agency, Europol, limited access to the fingerprint data for the purposes of criminal investigations. Member States may choose to check the fingerprints of an irregular migrant against the system, for the purposes of identification, without storing that data.

The proposed new Regulation would make some key changes to these rules. First of all, it would significantly enlarge the amount of personal data that will be taken and stored. Member States will have to take information on children from the age of six (rather than fourteen), and facial images as well as fingerprints. Eurodac will also now store data on the names, nationalities, place and date of birth, travel document information. For asylum-seekers, it will store the EU asylum application number (see the Dublin IV proposal), as well as information on the allocated Member State under the Dublin rules, for the first time. For irregular border crossers and irregular migrants, it will store the date of the removal from the territory.

There will no longer be an option merely to check data on irregular migrants; rather Member States will be obliged to take and store this information. While the rules on police and Europol access to Eurodac data will not be changed as such (although the Commission will review those rules soon), there will be more personal data for them to access: they will be able to get facial image information, and more individuals will have their personal data recorded in Eurodac in the first place.

Secondly, it will be possible for fingerprint data to be taken not only by national officials, but also (as regards asylum-seekers and irregular border crossers) by the new EU Border Guard and EU Asylum agencies. Thirdly, while asylum-seekers’ data will still be retained for ten years, data on irregular border crossers will now be retained for five years – up from 18 months at present. Data on irregular migrants will also be retained for five years. The data will be marked if a Member State gives a residence permit to an irregular migrant. Finally, Eurodac data will now be made available to third countries for the purposes of return, on certain conditions, including a refusal to disclose if the person who has applied for asylum. But the non-EU country might guess that the person has applied for asylum; in fact the EU’s procedures Directive requires that country to be informed of this in some cases.

The Commission justifies these changes by the need to strengthen the EU’s return policy as regards irregular migrants, and to keep track of them if they make movements across the EU. It believes that taking fingerprints and photos of young children is justified for child protection reasons. Collecting personal data on facial images is justified because some persons refuse to have their fingerprints taken.

This proposal obviously raises huge data protection issues, and it will be important to see what concerns are raised by national data protection authorities, as well as the EU’s Data Protection Supervisor. The arguments about child safety should be independently assessed by child protection experts. It is conceivable that taking facial images would avoid the need to insist upon taking fingerprints coercively, but it’s not clear why the Commission believes that storing data on names, birthdates et al is justified. The use of Eurodac to underpin EU return policy obviates much need to use or expand the Schengen Information System (which currently contains data on non-EU citizens who are meant to be refused entry) for similar purposes, and raises the question of whether there need to be two different databases addressing the same issue. The choice between the two databases is particularly significant for the UK, since it will have access to the Eurodac returns data (if it opts in to the new proposal), but doesn’t have access to the immigration alerts in the Schengen Information System, and indeed can’t have access to those alerts unless (rather improbably) it fully joins Schengen. (However, the UK does have access to the criminal law alerts in the Schengen Information System, such as alerts on suspected terrorists: see my further discussion here. It could lose that access after Brexit, as I discuss here).

Dublin IV

As noted at the outset, the amendments to the Dublin Regulation essentially aim to entrench the EU/Turkey deal and to save Schengen by deterring secondary movements of asylum-seekers, while also making a fresh attempt to establish relocation rules. To accomplish each of these objectives, the Commission proposes an extreme solution which is probably legally and/or politically unfeasible.

Let’s examine each element in turn. In order to entrench the EU/Turkey deal (and possibly future heinous deals with countries like Libya), the proposal transforms a current rule which gives Member States an option to apply to state that a non-EU state is a ‘safe third country’ for an asylum applicant in accordance with the asylum procedures Directive, rather than send the applicant to another Member State or consider the application after a transfer from another Member State under the Dublin rules. The CJEU recently took a permissive view of this provision (Mirza). In place of this option, there would be an obligation to assess the inadmissibility of an application on ‘safe third country’ or ‘first country of asylum’ grounds before applying any of the rules on responsibility for applications. This confirms the current practice as regards asylum-seekers coming from Turkey to Greece, which aims to return as many of them as possible to Greece despite the dubious designation of Turkey as a ‘safe’ country for asylum-seekers.

This doesn’t matter much in cases where Greece would anyway be responsible for considering the application under the Dublin rules, because it was the first country where the applicants entered. (Moreover, due to recent closure of the Greece/Macedonia border and other controls and fences on internal and external Schengen borders, it’s now very difficult to leave Greece even for those asylum-seekers not in detention). But contrary to popular belief, that is not the only ground for assigning responsibility under the Dublin rules. There’s also an obligation to bring family members together, where one of the family members has status as a refugee or asylum-seeker or otherwise has legal residence in another Member State.

The Mirza judgment did not address whether these family rules take priority over the ‘safe third country’ option, but the Dublin IV proposal is clear.  If a case is inadmissible on the dubious ‘safe third country’ or ‘first country of asylum’ rules, then the Member State in question is responsible, regardless of the family or humanitarian clauses in the Regulation. It’s arguable that this is a breach of the right to family life set out in the EU Charter of Fundamental Rights. But it’s certain that this change completely undercuts the broadening of the definition of ‘family member’ contained in the Regulation – extending it to cover siblings and families formed after leaving the country of origin (while Syrians were living in Turkey, for instance). Those changes therefore amount to a legal ‘Potemkin village’ – a cynical façade intending to mislead a naive onlooker.

It might be argued that family members should not be encouraged to pay smugglers and take unsafe routes to reach their loved ones who are already in the EU. Fair enough – but in that case, the EU should take steps to ensure their safe passage (note that the EU’s family reunion Directive requires Member States to admit family members of refugees). There’s nothing in this week’s batch of proposals to do that. The EU’s informal arrangements with Turkey do provide for ‘nuclear family’ members as one category of Syrian refugees to resettle. But these arrangements are not binding and (at time of writing) not even officially published (see this entry in the Council register of documents). They only apply to the ‘nuclear’ family, and only for Syrians.

Next: the attempt to deter secondary movements of asylum-seekers, in order to reinstate the Schengen system. Most notably, there will be punishments for asylum-seekers who do not stay in the responsible Member State. In that case the asylum procedure will be accelerated, and they will lose all benefits (health, education, welfare and accommodation) except for emergency health care. (However, the grounds for detention of asylum-seekers in the Dublin Regulation will not change – though the future proposal to amend the reception conditions Directive might seek to amend the detention rules there instead.) This will overturn the CJEU ruling in CIMADE and GISTI, which was based on the right to dignity in the EU Charter. Let’s put it plainly: asylum-seekers who flout the Dublin rules will be left to starve in the streets – even children, torture victims and other vulnerable people. And fast-tracking their asylum application implicitly aims at refouling them to their country of origin, with only limited suspensive effect of any appeal to the courts.

The violations of the Charter don’t stop there. According to the CJEU case law on the current Regulation, unaccompanied minors can move to another Member State and apply there. This ruling (MA) is also based on the Charter (rights of the child), but the Commission wants to overturn that too – in the process trashing its own proposal dating from 2014. Again, any attempt to argue that this aims to protect children by deterring them from moving is undercut by the prioritisation of inadmissibility rules over family reunion rules (even for unaccompanied children), as well as the failure to insert rules to ensure that the Dublin family rules are actually applied (such as the recent UK ruling on a requirement for DNA tests). If the EU and its Member States care so much about asylum-seeking children, why have they detained so many in Greece in poor conditions, and shrugged as so many suffered in northern Greece – shirking the legal obligations which they accepted to relocate them?

Furthermore, the proposal limits both the substantive and procedural remedies for applicants. They will only be able to challenge a decision on the responsible Member State on the grounds that the asylum system has broken down, or that they should be with their family member. This overturns the opinion in the pending cases of Karim and Ghezelbash (although it is possible that the Court will not follow this opinion). Also, they will only have seven days to appeal: this risks a breach of the Charter right to an effective remedy, given that in the Diouf case the CJEU found that a 14-day time to appeal was acceptable.

The proposal doesn’t only aim to restrict asylum-seekers in order to ensure that Dublin works effectively; it will also restrict Member States to the same end. The essentially unlimited discretionary ‘sovereignty’ and ‘humanitarian’ clauses will be amended to severely limit the circumstances in which a Member State can examine an application that is not its responsibility. If Angela Merkel (improbably) wanted to repeat her open-door policy of summer 2015 in future, the proposal would make that illegal. Various deadlines for Member States to act would be speeded up (although Member States have said before that this is impractical). Conversely, other rules which limit Member States’ obligations will be dropped: there will be longer periods of responsibility after issuing a visa or residence permit, and responsibility for those who cross a border without authorisation, or who abscond or who leave the EU and then come back, will be endless.

This brings us to the relocation rules. These will be triggered once a Member State is responsible for more than 50% of the asylum applications which objective criteria (based on income and population) indicate that it ‘should’ be responsible for. In other words, if Greece ‘should’ be responsible for 50,00 asylum applications under those criteria, other Member States would be obliged to relocate asylum-seekers from Greece once it was responsible for 75,000 applications. But Member States can’t relocate asylum seekers whose applications are inadmissible under the new rules discussed above, so this may have little impact on Greece anyway. Indeed, if the EU/Turkey deal breaks down, the combination of these rules would in principle put Greece in a worse position than it is currently. A new emergency relocation Decision would have to derogate from the Dublin rules again.

Then the proposal becomes truly surreal. The Commission suggests that Member States may opt out of relocating asylum-seekers, but they will have to pay €250,000 per asylum-seeker if they wish to do this. This is a fantasy on top of a fantasy. Member States have already shown that they are unwilling to apply the relocation Decisions of last September, or to adopt the proposal to amend the Dublin rules to this end that was tabled at that time. The idea of financial contributions in place of accepting individuals, whatever its merits, is perceived to be a ‘fine’ and was already rejected by Member States last year. That idea will not suddenly appear more attractive to Member States by doubling down on it, and suggesting a contribution set at an obviously absurd and disproportionate level, which the Commission does not even try to justify.

So why did the Commission jump the shark here? Perhaps someone in the Commission lost a bet. Or perhaps this is a legislative homage to the Belgian surrealist tradition of Magritte, et al. More seriously, it might be intended as a negotiating position. But such a ridiculous position will just backfire: it’s as if management started the latest pay talks with the unions by arguing that the workers should start paying the company for the privilege of working there. Or perhaps it’s a subtle way of addressing Greece’s debt problems: rejecting the relocation of a mere 10,000 asylum-seekers from Greece would transfer €2.5 billion to the Greek treasury – where it would rest briefly on its route to Germany.

I have another theory, well known to followers of British politics. Maybe the €250,000/person proposal is the Commission’s equivalent of ‘throwing a dead cat on the table’. The phrase is borrowed – like the EU’s current asylum policy – from Australia. It means that if the political conversation is particularly damaging to a certain politician, an ally of that politician suddenly does or says something outrageous. Everyone will start talking about that outrageous thing, just as they would be talking about the unfortunate feline; which means that no-one is talking about the original issue any more.  In this case, it means that everyone is talking about the €250,000 – and no-one is talking about the suspension of Schengen, or of the families who would be split up, or the people who would be made hungry and homeless, by the Commission’s Dublin IV proposal.


The Commission’s proposals are not a done deal, of course. Some Member States and Members of the European Parliament have misgiving about a visa waiver for Turkey, on migration control or human rights grounds. MEPs fought for years for many of the provisions in the Dublin III Regulation (on family members and unaccompanied minors in particular) which the Commission now seeks to overturn. As I pointed out above, some of the proposed changes to the Dublin rules are highly vulnerable to challenge in the CJEU, if adopted. The red herring of a €250,000 sanction is already floating on the surface of the pond. And the whole EU/Turkey deal might anyway be overturned at the whim of Turkish President Erdogan – the only politician whose ego makes Donald Trump’s look small by comparison. Nevertheless, EU asylum policy is already becoming more Orbanised in practice, and I would expect at least some elements of the further Orbanisation proposed by the Commission to be adopted.

For over twenty-five years now, the EU and its Member States have been attempting to get the Dublin system to work. The continued abject failures of those attempts to get this pig to fly never seem to deter the next attempt to launch its aviation career.  With this week’s proposals, the Commission is in effect trying to get the poor beast airborne by sticking a rocket up its backside. It might be best to stand back.