Thursday, 26 May 2016

The proposed new Audiovisual Media Services Directive: Key Features

Lorna Woods, Professor of Media Law, University of Essex

After a draft was leaked last week, the Commission proposal to revise the Audiovisual Media Services Directive (AVMSD) is now out.  Once again we see the Commission proposing the roll-out rather than the roll-back of regulation in the face of sector change.  The following provides an overview of some of the issues.

The first change is an extension of material scope.  The Commission explains in its Memo/16/1895 that a ‘limited extension’ will occur as the new proposal applies to ‘video-sharing platforms’, such as YouTube.  “Video-sharing platform services” are defined in new Article 1(aa) AVMSD (Art. 1(1)(b) of the proposal):

‘… a service, as defined by Articles 56 and 57 of the Treaty on the Functioning of the European Union, which meets the following requirements:
(i)                  the service consists of the storage of a large amount of programmes or user-generated videos, for which the video-sharing platform does not have editorial responsibility;
(ii)                the organisation of the stored content is determined by the provider of the service including by automatic means or algorithms, in particular by hosting, displaying, tagging and sequencing;
(iii)               the principal purpose of the service or a dissociable section thereof is devoted to providing programmes and user-generated videos to the general public in order to inform, entertain or educate;
(iv)              the service is made available by electronic communications networks within the meaning of point (a) of Article 2 of Directive 2002/21/EC.’

The phraseology and organisation here is different from the leaked draft. It makes clear the cumulative nature of the conditions but also clarifies that the organisational features of the video-sharing platforms identified are illustrative not an exclusive list.  It is also starting to engage with the issues surrounding editorial choice in an environment where ‘suggestions’ are made by programming – following big data profiling or just paid prominence. Moreover, the proposal integrates the point that such platforms can be caught if a ‘dissociable segment’ satisfies the definition, whereas the leaked version had a separate subclause (a ter) that applied a principal purpose test not just to video-sharing platforms but services defined in (1)(a)(i).

No doubt there will be much comment on the workability of this definition – not least where it draws the boundaries. Will there be difference in treatment between Instagram, Flickr and other photo-sharing sites, Twitter and Facebook (both of which have video capability, or link to videos) and Youtube, Vine and Vimeo; and are these sites similar to Dailymotion and For now, note the centrality of the concepts of ‘programme’, which by contrast to the leaked draft, gets a new definition (in Article 1(1)(b) AVMSD, replaced by Art. 1(1)(c) proposal), and ‘user-generated video’ (added to the AVMSD as (1)(ba)). This latter definition covers material created by end-users, but also material that such users may be re-using by uploading. This means the (unlawful) uploading of professional falls within the definition, but also material the creators of which are unknown.

The definition of programme does not apply just to video-sharing platform services, but is a central element in determining the scope of the AVMSD.  So, Rec. (3) (which was Rec 11 in the leaked draft) specifies that the AVMSD ‘should remain applicable only to those services the principal purpose of which is the provision of programmes to inform, entertain or educate’.  The purposes here are so broad that they can exclude nothing; the determinative element is therefore the programme.  This issue was the subject of litigation in the context of a press site which contained short video clips in New Media Online GmbH v. Bundeskommunikationssenat (Case C-347/14), in which the ECJ ruled that videos under a subdomain of a newspaper website could fall under the definition of a ‘programme’ within the AVMSD (an approach from which OFCOM has arguably differed in respect of its interpretation of the UK implementing regulations).   This position is reflected in Rec. 3, which notes that stand alone parts of newspaper sites can fall within AVMSD as can channels within video-sharing platforms.  Radio remains outside the AVMSD.

Under the current AVMSD,
(b)   ‘programme’    means    a    set    of    moving    images    with    or   without   sound   constituting   an   individual   item   within   a   schedule   or   a   catalogue   established   by   a   media   service   provider    and    the    form    and    content    of    which    are    comparable  to  the  form  and  content  of  television  broadcasting.   Examples   of   programmes   include   feature-length   films,   sports   events,   situation   comedies,   documentaries,   children’s  programmes  and  original  drama;

The proposal removes the phrase ‘and the form and content of which are comparable to the form and content of television broadcasting’ suggesting a move away from traditional television as the benchmark and towards a more open and arguably broader conception of just ‘an individual item’.

Rec 12 of the leaked draft is now found at Rec 26 and 27, slightly amended so as not to be limited to ‘video sharing’ platforms, though these are seen to raise particular issues.  Rec. 13 of leaked draft is now at Rec. 28 in a slightly amended form.  This change reflects the fact that the recitals refer to content restrictions rather than to scope, though Rec 28 contains the implicit acknowledgment that the proposal takes the possibility of regulation beyond those with editorial responsibility (even that at a very blunt level of choice – as in OTT services).  The Explanatory Memorandum skirts this issue, recognising that there will be a point of interplay with Articles 14 and 15 of the e-Commerce Directive (ECD). Those provisions provide immunity from damages for hosts with no knowledge of problematic content and prohibit the imposition of monitoring requirements (see also Rec. 30).  The proposal also recognises the need to include those services providers that are not established within the EU but are part of a group so as to ensure effectiveness of protection (Art. 28b).  In this context, we might be reminded of the reasoning of the Court in determining jurisdiction under the Data Protection Directive in Google Spain: legal form was not determinative of this question, but instead the business reality.  OFCOM in its response to the Commission’s consultation last year expressed concern about rules that would be ‘disproportionate and impractical’.  These provisions need also to be understood about the on-going trade negotiations with third countries, such as TTIP, which may affect their feasibility.

One of the main concerns with regard to video-sharing platforms is hate speech the understanding of which – in relation to all regulated platforms - ‘should, to an appropriate extent’ be aligned to Framework Decision 2008/913/JHA (concerning criminal expressions of racial hatred), specifically as regards the grounds on which hatred may be incited (Rec. 8).  What this means in practice, given the qualifiers used, as well as the relationship with the ECD in respect of video-sharing platforms, is uncertain (see Art. 28a(5)).  New Art. 6 AVMSD simply imposes on Member States the obligation to use ‘appropriate means’, the meaning of which is elaborated in Art. 28a. That provision points to a balancing of competing interests, which may allow for a certain degree of subjectivity and variation across Member States.  Art. 6a deals specifically with the protection of minors from a wider range of content – that likely to impair physical mental or moral development.   This envisages the need to give information to viewers so that they may make appropriate choices of viewing, rather than the imposition of technical measures. 

Another contentious issue has related to the country of origin (COO) principle, specifically where AVMS providers engage in forum shopping and ‘broadcast back’ to the ‘original’ target country.  This has always been problematic, with a body of jurisprudence on abuse of rights leading to specific exception provisions in the AVMSD.  The idea of COO, however, has always been popular with industry players as it avoids re-versioning costs and other costs associated with separate markets.  Whatever the view on COO, the anti-abuse provisions in AVMSD were complex and the issue of establishment open to interpretation.  The principle of freedom of re-transmission is restated but the possibility of derogating is extended to all audiovisual media services, not just broadcasting as is currently the case (see proposed Art. 3(2) – note differences in procedure between linear and non-linear services apply).  The grounds are those set out in Art 6, which contains an extended category of grounds of hate speech prohibited, and new Art. 12 which contains the ‘pornography provision’: transmission of relevant material is permitted, but in a way so that minors cannot access the material.  This applies to all AVMS providers.  The current broadcasting-only, protection of minors provisions (Art. 27 AVMSD), which currently act as triggers for the Art. 3 procedure, will be deleted.

While the AVMSD was a minimum harmonisation directive, recognised by Art. 4(1) AVMSD which allowed Member States to impose higher standards in respect of all fields coordinated by the directive, the proposal is now to limit the issues in respect of which Member States may impose stricter rules to Art. 5 (information obligations), 6 (hate speech), 6a (development of minors), 9 (standards for commercial communications), 10 (sponsorship), 11 (product placement), 12 (protection of minor- technical measures), 13 (on-demand quotas), 16 (tv European quotas), 17 (tv independent quotas), 19-26 (advertising and teleshopping rules), 30 (NRA) and 30a (ERGA).  In respect of the other provisions, it seems the AVMSD provides maximum standards.  It is notable that this latter category includes the provisions that are specific to video-sharing platforms as well as long-standing provisions such as the news reporting provisions.

There has been some ‘alignment’ of rules for linear and video on demand services (e.g. Art. 12).  This at an abstract level makes sense.  Commentators suggest that the industry trend is for entertainment, television and similar companies to focus on making and assembling content for distribution across the multiplicity of digital platforms available, in ways appropriate to those platforms but between which there may be overlap of form and content. Certainly, there is inter-platform competition.  So these changes are aimed at ensuring the mythical ‘level playing field’.  

The provision that has caught some attention when the proposal was leaked was that which imposes a European quota requirement on on demand AMS providers: at least 20% of the catalogue has to be European, and these works should be given prominence (Art. 13(1)).   It replaces the current provision which, in the words of the Commission ‘leaves room for testing different approaches’ but which potentially ‘unlevels’ the playing field.  Note that there is no ‘so far as practicable’ phraseology in this obligation (by contrast with the long-standing obligation on broadcasters), although member States may waive obligations in relation to small and micro enterprises (Art. 13(5)).  The obligation of ‘prominence’ is also not qualified (contrast the UK rules regarding ‘due prominence’ of PSB). Presumably it is intended at address the point made by the Society of Audiovisual Authors that currently on Netflix ‘where you have to look for European works (or even national works) under the rubric “Foreign Films”’ – not necessarily the most enticing branding.  

The current TV quotas rules have not addressed the problem of scheduling undermining the effectiveness of the quotas, a point noted in the response to the Commission’s consultation.   The definition of “European” has been left unchanged – as have the tv quotas.  This proposal will no doubt please the EU film industry, though it is likely to be less popular with the distribution sectors, which are already warning about strangling a still not mature industry.  Against this background it is noteworthy that Netflix has produced a series in Europe (Marseille – perhaps to get a stronger foothold in the French language market) and is about to launch a second, as well as engaging with local broadcasters (e.g. “Kiss Me First” with Channel 4; “Suburra” with RAI). 

The proposal also introduces a requirement for Member States to set up legally distinct and functionally independent regulators, in many aspects following the Recommendation of the Council of Europe (Rec (2000) 23). While the desirability of independent regulators is recognised in most Member States as a way of safeguarding freedom of expression while achieving other societal and political goals, there is no such obligation in the current AVMSD framework. The need to introduce such a requirement may be a response to developments in some of the Member States where there have been changes to the regulatory architecture in respect of the media with consequent concerns about media independence. It further specifies with a non-exhaustive list the remit on such regulators:  media pluralism, cultural diversity, consumer protection, internal market and the promotion of fair competition.  These roles must be established in law and carry with them enforcement powers.  A right of appeal for viewers/end-users must be provided.  Significantly, this requirement applies across all AVMS providers, including video-sharing platforms. 

The proposal also formalises the European Regulators Group for Audiovisual Media Services (ERGA) (which was established on the basis of a Commission Decision in 2014), in response to a perceived need for greater senior level cooperation in European audiovisual policy developments. The response to the group has been mixed, some questioning whether it adds anything to the existing groups, such as the Contact Committee and the European Platform of Regulatory Authorities (EPRA) which exists outside the EU framework. Alternatively, given the proposed expansion of the AVMSD and the uncharted territory awaiting the regulators, a mechanism for coordination may be important for the functioning of the COO principle.  ERGA has already produced reports for the Commission on independence; on the protection of minors in a converged environment; and on material jurisdiction as part of the preparation for the review of the AVMSD.

Historically, the broadcasting and now the audiovisual sector has revealed deep divides between member states and also between various sectors of industry.  The Commission has no doubt attempted to produce a balance of interests after an extensive review process.  What will remain once the Council and the European Parliament start to look at this, especially after what is likely to be intensive lobbying, is anybody's guess.  It may even be affected by Brexit; while the directive should be agreed before any UK exit, surely the UK’s negotiating position would be weakened between any ‘no’ vote and actual exit, shifting the balance between the free market and dirigiste Member States.

Barnard and Peers: chapter 14
Photo credit: Theon Greyjoy

Tuesday, 24 May 2016

EU law and the ECHR: the Bosphorus presumption is still alive and kicking - the case of Avotiņš v. Latvia

Stian Øby Johansen, PhD fellow at the University of Oslo Faculty of Law*

Yesterday, 23 May 2016, the Grand Chamber of the European Court of Human Rights (ECtHR) delivered its judgment in the case of Avotiņš v. Latvia. This seems to be the ECtHR’s first detailed appraisal of the so-called Bosphorus presumption (the rule on the relationship between EU law and the ECHR) after the Court of Justice of the European Union (CJEU) in Opinion 2/13 rejected a draft agreement providing for the accession of the EU to the European Convention of Human Rights (ECHR). It also provides a first glimpse of how the ECtHR views the EU law principle of mutual trust, which has become particularly dear to the CJEU over the last couple of years.


For the uninitiated: the Bosphorus presumption refers to a doctrine in the case-law of the ECtHR that goes back to the 2005 judgment in Bosphorus Hava Yolları Turizm ve Ticaret Anonim Şirketi v. Ireland. In that judgment the ECtHR first stated, in line with previous case-law, that member states of an international organization (such as the EU) are still liable under the ECHR for “all acts and omissions of its organs regardless of whether the act or omission in question was a consequence […] of the necessity to comply with international legal obligations” (Bosphorus para 153). It also recognized “the growing importance of international cooperation and of the consequent need to secure the proper functioning of international organisations” (Bosphorus para. 150). In an attempt to reconcile these two positions, the ECtHR established what is now known as the Bosphorus presumption or the presumption of equivalent protection of ECHR rights by the EU, even though the EU is not a party to the ECHR:

155. In the Court’s view, State action taken in compliance with such legal obligations is justified as long as the relevant organisation is considered to protect fundamental rights, as regards both the substantive guarantees offered and the mechanisms controlling their observance, in a manner which can be considered at least equivalent to that for which the Convention provides […]. By “equivalent” the Court means “comparable”; any requirement that the organisation’s protection be “identical” could run counter to the interest of international cooperation pursued […]. However, any such finding of equivalence could not be final and would be susceptible to review in the light of any relevant change in fundamental rights protection.
156. If such equivalent protection is considered to be provided by the organisation, the presumption will be that a State has not departed from the requirements of the Convention when it does no more than implement legal obligations flowing from its membership of the organisation.
However, any such presumption can be rebutted if, in the circumstances of a particular case, it is considered that the protection of Convention rights was manifestly deficient.

Many have been curious about whether the ECtHR would modify the Bosphorus presumption following the rather belligerent rejection of EU accession to the ECHR by the CJEU in Opinion 2/13. In the foreword of the ECtHR’s 2015 Annual Report its President, Guido Raimondi, indeed seemed to signal an interest in shaking things up (emphasis added):

The end of the year was also marked by the delivery on 18 December 2014 of the Court of Justice of the European Union’s (CJEU) eagerly awaited opinion on the draft agreement on the accession of the European Union to the European Convention on Human Rights. [T]he CJEU’s unfavourable opinion is a great disappointment. Let us not forget, however, that the principal victims will be those citizens whom this opinion (no. 2/13) deprives of the right to have acts of the European Union subjected to the same external scrutiny as regards respect for human rights as that which applies to each member State. More than ever, therefore, the onus will be on the Strasbourg Court to do what it can in cases before it to protect citizens from the negative effects of this situation.

Yet, in the ECtHR Grand Chamber judgment in the case of Avotiņš v. Latvia, it can clearly be seen that – spoiler alert – the Bosphorus presumption is still alive and kicking. Indeed, as I will show below, the ECtHR for the first time applies it to a case concerning obligations of mutual recognition under EU law. This is notable, since one of the main arguments the CJEU put forward in Opinion 2/13was that EU accession to the ECHR posed such a big threat to the principle of mutual trust that it would “upset the underlying balance of the EU and undermine the autonomy of EU law” (Opinion 2/13 para 194).


Before we look at how the Grand Chamber applied the Bosphorus, it is necessary to summarize the key facts of the case. Mr Pēteris Avotiņš is a Latvian national, who in May 1999 borrowed 100 000 US dollars from a company named F.H. Ltd. and undertook to repay that sum with interest by 30 June 1999. The loan contract was governed by Cypriot law, and Cypriot courts had non-exclusive jurisdiction to hear any disputes arising out of it.

In 2003, F.H. Ltd. brought proceedings against Avotiņš in a Cypriot district court, alleging that he had not repaid the above-mentioned debt. Since Avotiņš did not reside in Cyprus, notice of the proceedings and summons to appear had to be served on the applicants by Latvian authorities. There is some factual disagreement regarding the serving of this application (see para 19 of the judgment). It seems as if the summons slip had been signed, but the signature on the slip did not appear to correspond to the applicant’s name. Nevertheless, the Cypriot court ruled in Avotiņš’ absence on 24 May 2004, and ordered him to pay F.H. Ltd. 100 000 US dollar plus interest. According to the Cypriot judgment, the applicant had been duly informed of the hearing, but had not attended.

In February 2015, F.H. Ltd. applied to the Riga City District Court seeking recognition and enforcement of the Cypriot judgment. This request was first rejected, due to discrepancies regarding the postal address of Mr. Avotiņš. This rejection was appealed by F.H. Ltd. to the Riga Regional Court, which quashed the District Court’s rejection. Upon reexamination of F.H. Ltd.’s application by the District Court the application was granted in full – without the parties being present.

According to Avotiņš, it was not until 15 June 2016 that he became aware of the Cypriot judgment and the District Court order for its enforcement. He contacted the District Court immediately and acquainted himself with the Cypriot judgment and the Latvian order. Interestingly, before the ECtHR the Latvian authorities did not dispute these facts.

This is where things get complex (see paras 27-35 of the judgment), and I can for the sake of brevity only give a brief summary of the facts from this point out. First, Avotiņš did not attempt to appeal the Cypriot judgment. However, he decided to appeal the Latvian enforcement order on the grounds that it violated the Brussels I regulation (concerning jurisdiction over and recognition of civil judgments), which is part of EU law, as well as rules of Latvian civil procedure. Second, the Regional Court in October 2006 accepted Avotiņš’ submissions, and quashed the enforcement order. The District Court seemed to find that the Cypriot judgment was not enforceable due to the lack of the certificate referred to in Article 54 of the Brussels I regulation. Third, F.H. Ltd. appealed the October 2006 order of the Regional Court to the Supreme Court. At the start of the Supreme Court hearing in January 2007 F.H. Ltd. submitted copies of inter alia the certificate referred to in Article 54 of the Brussels I regulation. Later the same day the Latvian Supreme Court quashed the October 2006 order of the Regional Court, and ordered the recognition and enforcement of the Cypriot judgment. In doing so, the Supreme Court held that under article 36 of the Brussels I regulation a foreign judgment “may under no circumstances be reviewed as to its substance” (para 34 of the judgment, citing the January 2007 judgment of the Latvian supreme court).


The applications

Avotiņš then filed complaints against Latvia and Cyprus before the ECtHR. The application against Cyprus was rejected, due to being too late (see para 97 of the judgment, referring to a ECtHR decision of 3 March 2010). However, his application against Latvia was filed within the time-limits.

In his application against Latvia, Avotiņš argued that the Latvian Supreme court had infringed his right to a fair hearing, by recognizing and enforcing the Cypriot judgment which in his view was defective as it had been given in breach of his right to a defence. Several third parties intervened in the latter case, including the European Commission, which provided a lengthy submission on the applicability of the Bosphorus presumption to the case and the compatibility of Brussels I regulation with ECHR article 6 (the right to a fair trial).

The ECtHR’s introductory remarks

The judgment of the ECtHR, which was adopted by a majority of sixteen votes to one (with two judges appending a joint concurring opinion), opens with the premise that ECHR article 6 is applicable to the execution of foreign final judgments. According to the Court (para 98 of the judgment):

a decision to enforce a foreign judgment cannot be regarded as compatible with the requirements of Article 6 § 1 of the Convention if it was taken without the unsuccessful party having been afforded any opportunity of effectively asserting a complaint as to the unfairness of the proceedings leading to that judgment, either in the State of origin or in the State addressed.

The ECtHR then noted that it had “never previously been called upon to examine observance of the guarantees of a fair hearing in the context of mutual recognition based on European Union law” (para 98). However, the ECtHR added it had “always applied the general principle” that a request for recognition and enforcement of foreign judgments cannot be granted without the court examining the request “first conducting some measure of review of [the foreign] judgment in light of the guarantees of a fair hearing”.

Does the Bosphorus presumption apply?

Following these initial remarks, the ECtHR went on to consider whether and to what extent the Bosphorus presumption was applicable to the case. It did so over ten pages (paras 101-127), making this probably the longest treatment of this famed presumption by the ECtHR to day.

First, on the scope of the Bosphorus presumption, the ECtHR confirmed the principles laid down in its previous by referring to the summary of that case-law in paras 102-104 of its judgment in the Michaud case. From that case-law it follows that the substantive protection of human rights in the area of EU law that the Brussels I regulation belongs to is equivalent. In particular, this is confirmed by article 52(3) of the EU’s Charter of Fundamental Rights, which stays that the Charter has to be interpreted consistently with ECHR rights that correspond to it. The fundamental condition for applying the Bosphorus presumption was thus fulfilled.

Next, it follows from the ECtHR’s case-law that two further conditions must be satisfied for the Bosphorus presumption to apply. These are (1) the “absence of any margin of manouvre” on the part of the domestic authorities implementing an EU law obligation, and (2) the “deployment of the full potential of the supervisory mechanism” provided for under EU law. Applying these principles to the present case, the ECtHR first found that the Latvian Supreme Court did in fact not have any margin of manoeuvre in this case. In coming to this conclusion, the ECtHR pointed to the CJEU’s case-law on the relevant provisions of the Brussels I regulation, which “did not confer any discretion on the court from which the declaration of enforceability was sought” (para 106 i.f.).

The ECtHR’s discussion of the second condition, the deployment of the full potential of the supervisory mechanisms under EU law, was much more extensive. The Latvian Supreme Court had not requested a preliminary ruling from the CJEU regarding the interpretation of the relevant provisions of the Brussels I regulation. However, this was not decisive for the ECtHR, which stated (para 109):

this second condition should be applied without excessive formalism and taking into account the specific features of the supervisory mechanism in question. It considers that it would serve no useful purpose to make the implementation of the Bosphorus presumption subject to a requirement for the domestic court to request a ruling from the CJEU in all cases without exception […].

Following this statement, the ECtHR referred to cases where it has found that ECHR article 6 require domestic apex courts to give reasons when they refuse to refer questions to the CJEU for a preliminary ruling, “in light of the exceptions provided for by the case-law of the CJEU” (para 110). However, the ECtHR was quick to add that the review conducted in those cases differs from that in the present case, where “it examines the decision not to request a preliminary ruling as part of its overall assessment of the degree of protection of fundamental rights afforded by European Union law” (para 110).

For those reasons, the ECtHR found that “whether the fact that the domestic court hearing the case did not request a preliminary ruling […] is apt to preclude the application” of the Bosphorus presumption “should be assessed in light of the specific circumstances in each case” (para 111). It then pointed to the relevant circumstances at play in the present case: Avotiņš “did not advance any specific argument concerning the interpretation” of the relevant provisions of the Brussels I regulation, and he did not request that the Latvian Supreme Court should ask the CJEU for a preliminary ruling (para 111). Since there was thus norequest for a preliminary ruling, the fact that the Latvian Supreme Court did not ask for a preliminary ruling was not “a decisive factor” (para 111). Consequently, the ECtHR found that also the second condition for the application of theBosphorus presumption was satisfied.

Was the protection of ECHR rights “manifestly deficient”?

A finding that the Bosphorus presumption applies is not the end of it, however, since that presumption can be rebutted if the protection of the rights laid down in the ECHR was “manifestly deficient” in the present case (para 112). In the opening paragraph of this part of the judgment, the ECtHR points to the fact that the Brussels I regulation is based on the principle of mutual trust, and affirmed the importance of this principle in EU law (para 113):

The Court is mindful of the importance of the mutual recognition mechanisms for the construction of the area of freedom, security and justice referred to in Article 67 of the TFEU, and of the mutual trust which they require.

Nevertheless, the ECtHR soon went on to stress that the “methods used to create that area must not infringe the fundamental rights of the persons affected by the resulting mechanisms, as indeed confirmed by Article 67(1) of the TFEU” (para 114). This statement was immediately followed by some key critical remarks (para 114, emphasis added):

However, it is apparent that the aim of effectiveness pursued by some of the methods used results in the review of the observance of fundamental rights being tightly regulated or even limited. Hence, the CJEU stated recently in Opinion 2/13 that “when implementing EU law, the Member States may, under EU law, be required to presume that fundamental rights have been observed by the other Member States, so that …, save in exceptional cases, they may not check whether that other Member State has actually, in a specific case, observed the fundamental rights guaranteed by the EU” […]. Limiting to exceptional cases the power of the State in which recognition is sought to review the observance of fundamental rights by the State of origin of the judgment could, in practice, run counter to the requirement imposed by the Convention according to which the court in the State addressed must at least be empowered to conduct a review commensurate with the gravity of any serious allegation of a violation of fundamental rights in the State of origin, in order to ensure that the protection of those rights is not manifestly deficient.

By thus requiring domestic courts to presume the observance of fundamental rights by other member states, as the EU law principle of mutual trust requires, the domestic courts are “deprived of […] discretion in the matter, leading to automatic application of the Bosphorus presumption” (para 115). Although it is a bit difficult to discern exactly what the ECtHR is alluding to here, it is hard to disagree that the nature of the mutual trust principle creates a paradoxical situation (para 115 i.f.); a twofold limitation of the domestic court’s review of the observance of fundamental rights, due to the combined effect of the presumption on which mutual recognition is founded and the Bosphorus presumption of equivalent protection.

However, despite these apparent limitations on domestic courts when the principle of mutual trust is at play, the ECHR, which is a “constitutional instrument of European public order”, nevertheless requires of them to ensure that there is no manifest deficiencies (para 116, emphasis added):

"Accordingly, the Court must satisfy itself […] that the mutual recognition mechanisms do not leave any gap or particular situation which would render the protection of the human rights guaranteed by the Convention manifestly deficient. In doing so it takes into account, in a spirit of complementarity, the manner in which these mechanisms operate and in particular the aim of effectiveness which they pursue. Nevertheless, it must verify that the principle of mutual recognition is not applied automatically and mechanically […] to the detriment of fundamental rights – which, the CJEU has also stressed, must be observed in this context […]. In this spirit, where the courts of a State which is both a Contracting Party to the Convention and a Member State of the European Union are called upon to apply a mutual recognition mechanism established by EU law, they must give full effect to that mechanism where the protection of Convention rights cannot be considered manifestly deficient. However, if a serious and substantiated complaint is raised before them to the effect that the protection of a Convention right has been manifestly deficient and that this situation cannot be remedied by European Union law, they cannot refrain from examining that complaint on the sole ground that they are applying EU law."

The test laid down in the final sentence of the quoted paragraph is a tough one. It is therefore no surprise that Mr. Avotiņš was unable to meet its criteria.

What is more surprising, though, is how close he got to doing so. Although the ECtHR found the system of mutual recognition in the Brussels I regulation to be generally compatible with ECHR article 6 (paras 117-119), the ECtHR was skeptical about the Latvian Supreme Court’s interpretation and application of that regulation. Avotiņš had, as mentioned above, argued that the application for recognition of the Cypriot judgment should have been refused. According to the ECtHR he “raised cogent arguments in the Latvian courts alleging the existence of a procedural defect which, a priori, was contrary to [ECHR article 6] and precluded the enforcement of the Cypriot judgment in Latvia”. (para 120 i.f., emphasis added)

Moreover, the ECtHR found that the Latvian Supreme Court applied provisions of the Brussels I regulation that provided for exceptions to the obligation of mutual recognition too mechanically. The details here are quite technical, and concern the determination of the burden of proof – an issue that is not governed by EU law. In its concluding appraisal of the Latvian Supreme Court’s approach, the ECtHR stated (para 121): 'This approach, which reflects a literal and automatic application of Article 34(2) of the Brussels I Regulation, could in theory lead to a finding that the protection afforded was manifestly deficient such that the presumption of equivalent protection of the rights of the defence guaranteed by Article 6 § 1 is rebutted.'

This is as close to a finding of “manifest deficiency” as we have ever gotten in the ECtHR’s case-law – more on that later – but again the specific circumstances of the case came to the rescue. According to Cypriot law Avotiņš had a “perfectly realistic opportunity” of appealing the seemingly final judgment (para 122). That the applicant was unaware of this opportunity did not matter, as he when entering into a loan agreement should have “ensured that he was familiar with the manner in which possible proceedings would be conducted before Cypriot courts” (para 124).

Consequently, the judgment fizzles out with a finding that the protection of fundamental rights was not manifestly deficient, in the specific circumstance of the present case (para 125).


This judgment is notable for at least three reasons. First, it is notable for simple fact that it is the first time the Grand Chamber applies the Bosphorus presumption since Opinion 2/13. The judgment confirms that the presumption is still alive and well, as one could probably expect despite some murmuring from the ECtHR president.

Second, it is notable for being the first case where the ECtHR goes right up to the edge of finding that a “manifest deficiency” in the protection of fundamental rights has occurred, but then backing off at the last second because of a specific feature of the case at hand. As a side note, though, the ECtHR’s reasoning is less clear than one could have hoped for here. Since the burden of proof seems to be key to the outcome of the Latvian Supreme Court’s judgment, and this is an issue that is not regulated by EU law, one might have argued that the Latvian Supreme Court did in fact have some “margin of manoeuvre”. It seems as if it could have complied with both the obligation of mutual recognition and ECHR article 6 by modifying the Latvian rules on the burden of proof. The reason for the lack of clarity on the part of the ECtHR here may be caused by opaqueness of the Latvian Supreme Court’s reasoning; it “tacitly presumed either that the burden of proof laid with [Avotiņš] or that [a remedy against the Cypriot judgment] had in fact been available to the applicant” (para 121).

Third, the case is notable for being the first where Bosphorus presumption takes the principle of mutual trust head on. Particularly because that principle has been elevated to constitutional status by the CJEU over the last couple of years – with Opinion 2/13 as a major catalyst (see, particularly, Opinion 2/13 paras 191-194). The ECtHR’s judgment is wary of the dangers of mechanical application of mutual trust obligations, and reaffirms the principles laid down in Bosphorus. Despite some critical comments, my best guess is that the CJEU will see this judgment as something of an olive branch from the ECtHR. From the CJEU’s perspective the case is indeed welcome, as cases concerning the Dublin Regulation (e.g. M.S.S. v. Belgium and Greece), where the ECtHR have found that EU Member States violated ECHR article 3 by sending asylum seekers back to the first EU country they entered, were not been well received. However, one must not forget that there are important legal differences between cases such asM.S.S. and the present case of Avotiņš. Notably, the Dublin regulation does not – despite myths to the contrary – contain any obligation to send asylum seekers back to the first EU country they entered. In Avotiņš the situation is markedly different: there is seemingly a clear obligation on the Latvian authorities to recognize and enforce the Cypriot judgment. Although, admittedly, the ECtHR’s unclear reasoning concerning the Latvian rules on burden of proof makes this distinction a bit less clear.

Barnard & Peers: chapter 9
JHA4: chapter II:8
Photo credit:
*Reblogged with permission from the Øby-kanalen blog

Friday, 6 May 2016

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

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, the current 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.

Barnard & Peers: chapter 26
JHA4: chapter I:3, chapter I:4, chapter I:5, chapter I:7

Photo credit: