Wednesday 16 March 2016

The draft EU/Turkey deal on migration and refugees: is it legal?

Steve Peers

In the last week, there has been considerable legal controversy over the planned EU/Turkey agreement on refugee issues. I commented (together with Emanuela Roman) in general on the relevant points last month, but now we have a leaked draft text of a final deal. (See also today's Commission communication on the deal, which adds a lot of important detail). This is a good moment to comment specifically on this draft, just before the summit meeting due to finalise it. 

I have underlined the full leaked text below, and added annotated comments on each part of it. I will update this blog post if necessary in light of the final deal (if there is one).

The agreement will be formulated as an EU-Turkey statement. It will take as its basis the principles set out in the statement of 7/3/2016 while adding the following elements:

a) Since the agreement will take the form of a ‘statement’, in my view it will not as such be legally binding. Therefore there will be no procedure to approve it at either EU or national level, besides its endorsement by the summit meeting. Nor can it be legally challenged as such. However, the individual elements of it – new new Greek, Turkish and EU laws (or their implementation), and the further implementation of the EU/Turkey readmission agreement – will have to be approved at the relevant level, or implemented in individual cases if they are already in force. I will come back to the implications of this below.

b) The March 7 EU/Turkey statement is still applicable. As a reminder, it provided that: ‘all new irregular migrants’ reaching the Greek islands from Turkey would be returned to Turkey, with the EU covering the costs; there would be a ‘one-for-one’ resettlement of Syrians from Turkey by the EU, for every Syrian readmitted by Turkey; the aim was to lift short-term visa requirements for Turkey by June 2016; the existing €3 billion in EU and Member State funds committed for Syrian refugees in Turkey would be spent more quickly, with a decision on ‘additional funding’; the EU and Turkey would ‘prepare for the decision’ on opening new chapters in the accession process; and the EU and Turkey would work toward a de facto ‘safe zone’ in part of Syria. The statement also included some commitments on restoring Schengen, but they aren’t affected by the draft full deal.

As regards accession to the EU, note that: there are 35 ‘chapters’ to be negotiated; only one chapter has been closed to date in a decade of accession talks; a commitment to prepare for opening a chapter does not close a chapter, or even mean that a chapter will be opened (any Member State can still block this); it takes years to negotiate chapters; and there are many political obstacles to approving Turkish accession, which requires national government and parliament approval in every Member State (and possibly referendums in some).   

1.       On returns to Turkey: a) This will be a temporary and extraordinary measure which is necessary to end the human suffering and restore public order, b) Migrants arriving in the Greek islands will be duly registered and any application for asylum will be processed by the Greek authorities in accordance with Directive 2013/32/EU. Migrants not applying for asylum or whose application has been found unfounded or inadmissible in accordance with the said directive will be returned to Turkey, c) Migrants having been returned to Turkey will be protected in accordance with the international standards concerning the treatment of refugees and respecting the principle of non-refoulement.

a) How temporary exactly? b) This makes clear that the EU’s asylum procedure directive will apply to those who reach the Greek islands, as legally required. Note that the text does not refer to Greek waters: but the Directive explicitly applies to them too. It does not apply to international or Turkish waters. It is not clear what is planned as regards those intercepted before they reach the Greek islands.

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

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

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

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

Furthermore, again as discussed in the previous post, many NGOs argue that refugees are not always safe from mistreatment in Turkey itself, although no one argues that all of them are mistreated there.  Equally Turkey allegedly returns some people (but clearly not all of them) to unsafe countries, and the March 7 deal explicitly plans for a ‘safe zone’ in Syria.  Such a zone is conceivable in theory, but whether it would indeed be safe would have to be judged when and if it happens; and it may become less (or more) safe in light of events. To address these issues the procedures Directive says that the asylum-seeker must be able to argue that ‘the third country is not safe in his or her particular circumstances’. Everything will then turn on the assessment of an argument along these lines.

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

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

2.    On resettlement based on 1-for-l principle: a) Priority will be given to Syrians who have not previously entered the EU irregularly, b) On the EU side, resettlement under this mechanism will take place, in the first instance, by honouring the commitments taken by Member States in the conclusions of Representatives of the Governments of Member States meeting within the Council on 22/7/2015. [Any further need for resettlement will be carried out within the limits and in accordance with the distribution set out in [relocation decision of 22/9/2015 - non-allocated places]. c) Should the number of returns exceed the numbers provided for by these commitments, this agreement will be subject to review.

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

On point a) it is open to Member States to prioritise resettlement on whatever criteria they like. Obviously the intention here is to deter people from attempting unsafe journeys via smugglers; whether that would work depends on the numbers who might be resettled. That is addressed by point b), which refers to the remainder of the 23,000 people that the EU committed to resettle from non-EU countries last year, and possibly (note the square brackets) another 18,000 who were originally going to be relocated from Hungary, but weren’t because the Hungarian government refused. These numbers clearly fall far short of the 2 million-plus Syrians estimated to be in Turkey. Point c) only undertakes to review the deal if the original modest numbers are reached. While the Hungarian government has reportedly been objecting to the idea of resettlement, note that this country didn’t commit itself to accept any resettled refugees last year, and so would not have to take any more people under this deal.  Whether other countries decide to resettle people is up to them. The Hungarian government resents interference in its own migration decisions; it does not and should not have any say in the resettlement decisions of other States.

3)   Turkey will take any necessary measures to prevent new routes for illegal migration opening up out of Turkey and into the EU.

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

4)   Once the irregular crossings between Turkey and the EU have come to an end, the Voluntary Humanitarian Admission Scheme will be activated. EU Member States will contribute on a voluntary basis to this scheme.

This scheme is set out in a Commission Recommendation from December, as discussed in detail here. Note that this would not apply until irregular crossings have stopped. This seems rather utopian - although the Commission paper talks about substantial reductions as an alternative.

5)   The EU and Turkey will further speed up the disbursement of the initially allocated 3 billion euros and ensure funding of additional projects before the end of March. Furthermore, the EU will decide on additional [X] billion for the period [Y] for the Turkey Refugee Facility.

The amount of additional money from the EU and its Member States is still open for negotiation. Note that this money is not, as is widely assumed, simply handed over to Turkey for unnamed nefarious purposes; legally speaking it is only intended for projects that assist the Syrian refugee population. Today's Commission paper lists how the money will be spent, starting with a contract to provide food aid to over 700,000 Syrians. Of course everyone should keep a beady eye on developments to ensure that the money is all spent as intended.


Overall this draft tries to address the two main legal concerns about the March 7 ‘deal’. It makes clear that the EU asylum laws will apply to those who reach Greece (subject to the caveat about what happens to those intercepted in Greek waters), and that Turkey will have to meet the relevant standards when taking people back. The key legal question will therefore be how these commitments are implemented in practice.

The main legal route to challenging what happens should be by asylum-seekers through the Greek courts. Those courts could refer questions to the CJEU about EU asylum law (the CJEU could fast-track its replies). Alternatively if the asylum-seekers have gone through the entire Greek court system, they could complain to the European Court of Human Rights.

What about the ‘deal’ itself? As I said at the outset, it is not binding so cannot be challenged as such. Its individual elements are binding and so their legality (or the implementation of them) can be challenged separately. On this point, it would be possible for the European Parliament or a Member State to challenge in the CJEU one particular legally binding element: the decision on the EU’s position on the EU/Turkey readmission treaty. That won’t directly affect the Greece/Turkey readmission deal, which is the key element in returns to Turkey in practice; but any ruling the CJEU might make would obviously be relevant to that latter deal by analogy.

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

Photo credit:


  1. It is not possible to challenge anything in the Greek courts because lawyers and courts are on a long-term strike. Moreover, the backlog of cases is such that even without the strike the delays in court cases extend to many years.

    Greece is almost a failed state, and the EU seems determined to guarantee its complete failure while also using that to avoid their legal obligations to manage a mass refugee flow.

    It is the eurozone technique yet again -- obviously a successful formula, in the eyes of northern Europe.

    1. If there is no effective access to the Greek courts at all, then an asylum-seeker facing expulsion can go direct to the European Court of Human Rights, because no domestic remedy would be possible. More generally, the 'EU' is not something done *to* its Member States, it's primarily done *by* them. The failure to relocate sufficient numbers of asylum-seekers from Greece is a failure by Member States. The EU doesn't stop them doing this; just the opposite, it obliges them to take a certain number (as most of them agreed). The relocation laws would at least help to manage the mass refugee flow a little, but again it's *Member States* that aren't applying them.

    2. My analogy with the eurozone management is that these agreements are not made within the normal framework of EU law and are not automatically justiciable. In a sense, it is a reversion to the former intergovernmental agreements of the 1980s -- and a very serious backward step for the EU.

      Concerning the relocation laws, these are clearly an abject failure. If they are not working by now, what chance of such relocations working in the future? Is the Commission actually taking any MS to the CJEU over this failure, or is there no actual breach of law?

    3. EU asylum law is certainly justiciable, and this deal can't somehow override it. That's why there are attempts to at least make a gesture in the direction of the EU laws. I agree, the Commission should start challenging Member States over the relocation laws. They are drafted to avoid giving individual rights so it is difficult or impossible for individuals to invoke them.

    4. If Greece is 'almost a failed state' I would suggest that Mr Baldwin-Edwards leaves it immediately. Moreover he should get his facts correctly: the courts are not on strike, only lawyers are abstaining from courts complaining about proposed persion reforms.

      Konstantinos Magliveras

    5. In reply to Mr Magliveras, I suggest that he learn to distinguish between empathy and hostility. As for getting facts correct, the fact is that the courts cannot operate when the lawyers are on strike. Do you actually have anything substantive to say on the refugee crisis?

  2. Great assessment, thanks!
    According to the Procedures Directive the application of the "Safe Third-Country Concept" is subject to "rules requiring a connection between the applicant and the third country concerned". Won't it be difficult in practice to establish such "connection" for each and every (Syrian) migrant arriving in Greece from Turkey?
    1) A mere transit through Turkey is in my opinion not enough to establish said connection; 2) Turkey does not really keep track of every refugee living in private accomodation on its territory (and that concerns the very large majority of its 2+ mio refugees). Therefore, in practice it will be difficult to prove said connection in every single case.
    Or am I missing a point? - I haven't read about this issue in your blog post of 5 February either...
    In any case, to be continued tonight!

    1. I think at least some Member States would be prepared to argue that mere transit is enough. If this is correct, that should be easy to prove, since it is self-evident that those arriving in the Greek islands came from Turkey.

    2. Agreed, they probably would, since it is convenient in the present situation to argue accordingly. Hungary for example is regarding Serbia as a "safe third country" for every person transiting. I just don't think that the ECJ would uphold such an interpretation... Also, UNHCR and ECRE were always of the opinion that mere transit is considered as an insufficient link.

    3. Mirza C‑695/15 PPU may be of some interest here.

    4. I will comment on the Mirza case when I have a moment. But that judgment doesn't analyse the content of the 'safe third country' clause at all. It just says that Member States can freely apply it even if they are the responsible Member State under Dublin. It was rather cheeky of the Hungarian court not to ask the substantive question of whether Serbia is safe.

  3. Hello,

    Thanks for the detailed analysis of the draft plan.
    Every applicant has the right to an effective remedy as stipulated in the article 46 of procedures directive. As the majority of persons going to Greek island will lodge an asylum application, they will also have the right to appeal the decision of inadmissibility concerning their asylum application. However, article 46 (6) b allows the Greek courts to decide whether the applicant is entitled to stay in the Greek territory pending the outcome of the remedy. This shortcut for sending applicants to Turkey is further qualified by the following paragraph 7. In minimum, paragraph 7 gives 1 week time to applicants (along with the obligation to provide translation and legal assistance to the applicants).
    I am of the opinion that the decisions to send back the applicants to Turkey will contradict the abovementioned articles of the procedures directive, since the conditions in Greek islands are far from ensuring that the related procedural safeguards are implemented effectively. Do you think that such a shortcoming will lead to a backlog of cases before ECHR and CJEU and the failure of the plan in medium or long term?

    1. That's not quite right. Article 46(6)(b) of the procedures Directive allows for that shortcut only as regards *some* inadmissibility cases. It applies to 'first country of asylum' cases, but *not* to 'safe third country' cases. Also the rules about a minimum one-week time only come into play if Greece applies a border procedure. I don't know whether Greek law counts all this as a border procedure or not.
      In order to cause a backlog at the CJEU cases have to get there in the first place. They can only get there if Greek courts send questions to the CJEU; individuals can't go directly.
      The ECtHR is a different question. Applicants can go there if they have exhausted Greek remedies, or never really had access to them. They will need a lawyer to file the claims though, so if they have no effective access to a lawyer then I can't see how that works.
      The CJEU can fast-track cases and decide within a couple of months. The ECtHR usually takes longer. All this depends on how many people are willing to fight their corner and how effective their access to a lawyer is.

  4. This comment has been removed by a blog administrator.