Friday 5 February 2016

The EU, Turkey and the Refugee Crisis: What could possibly go wrong?

Steve Peers* and Emanuela Roman, University of Palermo and Vrije Universiteit Amsterdam**

*Emanuela Roman wrote the sections on ‘safe third country’, ‘super-safe third country’ and ‘first country of asylum’. Steve Peers wrote the rest of the main text and the annexes.

The key non-EU country in the EU’s ongoing refugee crisis is Turkey: the host of over 2 million Syrian refugees, and a transit country for many asylum-seekers. An increasing number of them have been making the journey from Turkey to the Greek islands, leading to a significant rise in the number of would-be asylum-seekers in the EU over the last year. Tragically, many have died making this crossing.

To address these issues, the EU and Turkey reached a deal in November with a number of different elements. The main aim was to improve the position of Syrian refugees in Turkey (reducing the ‘push’ factor which results in more people planning to leave), and to return to Turkey those who did not need international protection. But, according to the latest Frontex statistics, most of the people arriving from Turkey do need international protection: about 90% of those arriving in Greece in December were from Syria, Iraq and Afghanistan, countries with high recognition rates for asylum claims.

It follows that returning to Turkey only those who don’t need international protection would only make a small dent in the numbers coming to the EU. Many politicians, in reaction to a portion of public opinion, would like to reduce those numbers far more. So last week, a further plan emerged: to return to Turkey those who do need international protection, on the grounds that they already had such protection in Turkey – or if not, they should seek it there. This would entail designating Turkey as a ‘safe third country’. The plan would entail a nearly immediate return to Turkey of any would-be asylum-seekers who reached the Greek islands.

The following analysis looks first of all at the main elements of the overall EU/Turkey deal on the refugee crisis, then in more detail at the new plan to return asylum-seekers to Turkey.

The EU/Turkey deal

The EU/Turkey deal contains concessions on both sides. The EU side has agreed: the opening of another chapter in the EU/Turkey accession process, and preparatory work by the Commission on further chapters; a proposal to lift Schengen visa requirements for Turkey by October 2016, if Turkey meets the requirements of the EU’s ‘roadmap’ on visa liberalisation; and a ‘Refugee Facility for Turkey’, totalling of €3 billion for Turkey. Subsequently, the Commission adopted a Recommendation on an EU ‘humanitarian admission scheme’ of Syrian refugees from Turkey (discussed in detail by Laura Robbins-Wright here).

In return, the Turkish side agreed: to readmit non-Turkish nationals to its territory, from June 2016; to apply a previously-agreed plan on the status of Syrian refugees in Turkey; and to prevent non-EU citizens from leaving.

There are also agreements in both sides’ mutual interest: to hold regular summits (at least twice a year); and to ‘upgrade’ the EU/Turkey customs union, to include services and investment, with negotiations to start before the end of 2016.

Several elements of the deal should be clarified further. It’s sometimes claimed that the deal has ‘fast-tracked’ Turkish membership of the EU. This is clearly not the case. Turkey applied for EU membership in 1987, and negotiation talks opened 18 years later, in 2005. In the eleven years since, the EU and Turkey have agreed only one of the 35 ‘chapters’ of issues being discussed in the negotiation. Before last year’s deal, they had opened another 13 chapters; the deal raised that number to 14. There is no date to open talks on the remaining 20 chapters; the deal is expressly without prejudice to EU Member States’ position on the planned ‘preparatory work’. Overall, the deal means that the enlargement negotiations will now move at the pace of a turtle, instead of a snail.

The prospect of Turkish accession to the EU also faces a number of obstacles on the EU side: objections by many Member States (including possible referendums), and misgivings by the European Parliament (which must approve accession treaties). Nor would accession lead immediately to free movement of Turkish citizens to the EU. Recent accession treaties have provided for waiting periods of seven years before workers from the new Member States can move to the older Member States, and the UK, which has a veto over accession treaties, insists that future enlargements must provide for even longer waiting periods.

As for the visa waiver, it should be noted that it will apply to (short-term) visas to visit the Schengen states. Therefore it will not ever apply to the UK and Ireland; and does not yet apply to Romania, Bulgaria, Croatia or Cyprus. It’s the EU’s usual practice to offer visa facilitation (fast-track issue of visas, with reduced fees) and then a full visa waiver to neighbouring States which have agreed a readmission treaty. As the text of the EU/Turkey deal points out, that waiver is dependent in each case upon the third State fulfilling a list of conditions set out by the EU (for the Commission’s most recent report on Turkey meeting those conditions, see here).

It’s the readmission deal – the quid pro quo for the visa waiver – that is central to the issue of the refugee crisis. The EU/Turkey readmission treaty has applied since October 2014. It applies to Turkish citizens straight away, but Turkey (like many other states signing up to readmission treaties) negotiated a delay before it applies to nationals of other countries. That’s a three year delay, so it was due to expire in October 2017. However, in light of the perceived migration and refugee crisis, the EU was not willing to wait that long until it called upon Turkey to accept third-state citizens back onto its territory.

Finally, the ‘Refugee Facility for Turkey’ aims to reduce the ‘push’ factor which leads to departures from Turkey to the EU. According to the Decision establishing the fund – which Member States finally agreed to recently – the Fund will assist refugees and host communities. No further details are given.

Before we look at the substance of the law on returning people to Turkey, it’s useful to make some general observations on EU policy in this area. Some critics of EU asylum policy argue that it should be more like the Australian policy. In fact, in some ways the EU is moving towards such a policy, as we’ll see. But – leaving aside for a moment the desirability of the EU adopting an ‘Australian’ asylum policy – there are legal, political, geographical and practical limits to doing so.

In a nutshell, Australia intercepts vessels of asylum-seekers on the high seas (international waters) and arranges for the asylum-seekers to be detained and their applications processed in other countries, which Australia considers to be ‘safe’. Furthermore, Australia resettles confirmed refugees directly from third countries (about 13,000 a year), as its contribution to sharing the burdens of the countries which host those refugees (the large majority of refugees live in developing countries).

There’s no legal obligation upon the countries which Australia considers ‘safe’ to take the asylum-seekers and process them. The UN Convention on Refugees (the ‘Geneva Convention’) imposes no such obligation (on the international law framework for ‘safe third country’ rules, see discussion here). Even if it did, some of the countries concerned haven’t ratified that treaty. So Australia has to talk those countries into accepting the people concerned. They won’t accept unless Australia pays most of the costs.

How does this compare with the EU? First of all, the numbers are hugely different: 18,000 people arrived illegally by sea in Australia in 2012-13, whereas over a million potential asylum-seekers arrived in the EU last year. We’re comparing apples and orchards here. There are simply more people wanting to apply for protection in closer vicinity to the EU, as compared to Australia, and the distance to travel is shorter. Furthermore, there are no ‘high seas’ between Greece and Turkey, so interception raises different legal issues. Once would-be asylum-seekers reach Greek waters, EU law says they can apply for asylum, and Greece is also bound by the European Convention on Human Rights (ECHR) as regards them.

That brings us to the next distinction. The Australian constitution has weak human rights protection, and that country’s High Court has recently ruled in favour of the offshore detention policy. In contrast, EU countries are governed by the case law of the European Court of Human Rights (ECtHR), which is open to individual complaints and can give binding rulings, often critical of national policies in this area. While Australia has signed up to the International Covenant on Civil and Political Rights (ICCPR), and allows individuals to make complaints to the United Nations Committee which interprets that Covenant, the Committee’s rulings aren’t binding. When the Committee rules against Australian policy on asylum-seekers – which it often does – Australian politicians in effect throw the ruling on the barbecue.

So given the large numbers concerned and the legal constraints, if the EU wants non-EU countries to take back non-citizens of those countries who have made their way to the EU, it needs to offer a lot to them to convince them to do so. At first sight, it may seem overly generous for the EU to offer money to Turkey, waive visa requirements and accelerate the accession process a little, in order to secure Turkey’s cooperation as regards refugees and migrants. But Turkey is under no obligation now to take back non-citizens or to restrain them from leaving. Even after it agrees to accept returns of non-citizens to its territory, it could end its obligations at any time by denouncing the readmission agreement – if it is willing to accept the re-imposition of visa requirements by the EU as a consequence. The financial assistance, if spent as intended, will also reduce the 'push' factor for migration of Syrian refugees towards the EU.

Returning people to Turkey

Turkish citizens

Turkey is already obliged to accept back its own citizens under the readmission agreement with the EU. Turkish citizens in the EU might have rights to stay under the EU’s immigration and asylum legislation, or under the EU/Turkey association agreement. But they have no general right of free movement to the EU, and there is no prospect of it being extended to them before (or indeed well after) EU membership. So those Turkish citizens in the EU without such a right to stay, including failed asylum-seekers, must be returned.

It should be noted that the EU Commission has proposed to list Turkey as a ‘safe country of origin’ for asylum purposes, meaning that asylum claims by Turks would be fast-tracked (but not rejected automatically). This is a rather dubious suggestion (for the reasons set out here), and it is not yet known whether it will be accepted.

Non-Turkish citizens: Irregular migrants

Turkey is not obliged to accept back any non-Turkish citizens until it brings forward the relevant obligations in the EU/Turkey plan – as it has agreed to do so. If someone has never applied for asylum, and has no other ground to stay, EU Member States will then be able to return them to Turkey, if the Member States can prove that the person was previously in Turkey. It should be sufficient to show that the person concerned crossed from Turkey to a Greek island.

This is equally the case for failed asylum-seekers, if the person concerned has failed on the merits. In other words, if the non-Turkish citizen has not convinced the Member State’s authorities (or the courts on appeal) that he or she faced persecution or serious harm, that person could be sent back to Turkey once that country extends the scope of the readmission deal.

The more difficult question – which is the subject of the new plan – is whether asylum-seekers who have come via Turkey can be rejected and returned to Turkey on the grounds that their applications are inadmissible. Let’s be clear what that means: those applications would not be rejected on the basis that the person wasn’t a genuine refugee, but that he or she either (a) could have applied for protection in Turkey or (b) already had protection there. The former is the ‘safe third country’ concept; the latter is the ‘first country of asylum’ concept. There are detailed definitions of each concept in EU law, in the Asylum Procedures Directive. We will consider those definitions in turn.

Is Turkey a ‘safe third country’?

On Thursday 28 January 2016, Diederik Samsom, leader of the Dutch Labour Party, announced in an interview with the newspaper De Volkskrant (followed by an interview on the nightly TV programme Niewsuur) a Dutch proposal for a new plan to radically reduce the number of migrants and asylum-seekers entering the EU from Turkey. The proposal was immediately baptised ‘the Samsom Plan’.

The plan would have the support of Dutch PM Mark Rutte and would also receive support by a number of EU Member States, among which Germany, Austria and Sweden. The idea is to offer Turkey the resettlement of 150,000 to 250,000 refugees per year from Turkey to the EU countries that voluntarily agree with the plan. This resettlement would presumably be on the basis of the Commission Recommendation on humanitarian admission from Turkey, referred to above. In exchange for this, Turkey would have to accept the return of all migrants and asylum-seekers who cross the Greek-Turkish border irregularly. According to Mr Samson, these people would have to be very rapidly returned from Greece to Turkey by ferry-boat, and it would be Turkey’s responsibility to deal with their reception and asylum application.

This new plan is based on the assumption that Turkey can be considered a ‘safe third country’ – a non-EU country where an asylum-seeker can apply for asylum and be granted access to asylum procedures and reception conditions in line with international and EU law.

The Samsom Plan, however, does not seem to come from Mr Samsom’s mind. The European Stability Initiative (ESI), a think tank specialised in Southeast Europe, presented a very similar proposal in October 2015. The original plan, based on the fundamental idea of considering Turkey a safe third country, was called ‘Merkel Plan’, because initially Germany alone would have the main role in the resettlement scheme with Turkey. The original plan was then further developed (becoming ‘Merkel Plan 2.0’) and a ‘coalition of the willing’ (including the Netherlands) was gathered around Germany. From October 2015 to January 2016 the ESI presented this proposal in different countries across Europe, but it was only following Mr Samsom’s interview, that the now renamed Samsom Plan burst into the public debate.

What is the legal definition of a ‘safe third country’? Article 38(1) of the Asylum Procedures Directive sets out a series of legal requirements that need to be met in order for a third country to be considered ‘safe’ for asylum-seekers:
a)      life and liberty shall not be threatened on account of race, religion, nationality, membership of a particular social group or political opinion;
b)      there shall be no risk of serious harm (consisting of: death penalty; torture or unhuman or degrading treatment; or a serious threat to the applicant’s life due to indiscriminate violence in situations of conflict, as defined by Article 15 of the Qualification Directive, ie the concept of ‘subsidiary protection’);
c)       the principle of non-refoulement (non-return to an unsafe country) shall be respected; and
d)      the possibility shall exist for the applicant to claim refugee status and to receive protection in accordance with the  Geneva Convention.

Article 38(2) establishes also several procedural guarantees. The safety of a third country must be always assessed on a case-by-case basis in order to check whether the notion is applicable to the particular circumstances of the individual applicant concerned. Moreover, the applicant must be guaranteed the right to challenge the application of the safe third country concept to his or her case, based on the fact that that country may not be safe in his or her particular circumstances. The question is whether an extremely rapid procedure as the one envisaged by Mr Samsom, would allow for a case-by-case examination of the individual circumstances of each asylum-seeker arriving in Greece from Turkey.

A further, more practical, question concerns who would be responsible for these procedures. Considering the difficulties faced by the Greek authorities in managing the current migrant flow and the established deficiencies of the Greek asylum system, it is hard to believe that the Greek authorities (despite the assistance provided by Frontex and the European Asylum Support Office, EASO) would be able to implement a systematic readmission plan as far-reaching as the one foreseen by Mr Samsom. As mentioned above, such plan must indeed be based on the individual consideration of each asylum application, otherwise it could amount to  a violation of the prohibition of collective expulsions entailed in Article 4 Protocol 4 of the ECHR. Such practice has been condemned by the European Court of Human Rights (ECtHR) in the famous Hirsi case, where the Grand Chamber found Italian authorities responsible for violating the ECHR, because they returned a group of Eritrean and Somali migrants intercepted on the high seas back to Libya without granting them the possibility to apply for asylum. The same rule obviously applies to asylum-seekers who enter the territorial waters or land on the territory of a Member State.

In addition, according to Article 46 of the Procedures Directive, asylum-seekers have the right to refer to a national court the decision to consider their application inadmissible pursuant to Article 33(2). They can stay on the territory during their initial application, and apply to a court to stay during this appeal. The Grand Chamber of the ECtHR unanimously confirmed in Hirsi (and reiterated in following case law – see for instance, Sharifi and others v Italy and Greece and Khlaifia and others v Italy) that return is only possible after the asylum-seeker has been able to claim asylum before a national authority, and to stay on the territory at least until the first instance decision on the application was made. However, it is even more doubtful that the accelerated procedure proposed by Samsom would allow for asylum-seekers to challenge the decision to return them to Turkey in front of a judicial authority and in the respect of all due procedural safeguards under the Directive and the ECHR.

However, besides the procedural issues, the crucial question here is more substantive: can Turkey be considered as a safe third country? Does Turkey comply with the requirements established by the Procedures Directive?

First, Turkey ratified the 1951 Geneva Convention and its 1967 Protocol, but maintains a geographical limitation for non-European asylum-seekers, thus recognising refugees originating only from Europe (i.e. from countries which are members of the Council of Europe). The geographical limitation provides the first barrier to accessing asylum in the country. Moreover, Syrians represent a particular case. They were at first received as ‘guests’ and then subject to a temporary protection regime, formalised by a Regulation on Temporary Protection only in October 2014 (for more details, see the updated AIDA Country Report on Turkey). The basic idea behind the temporary protection regime is to host Syrians until the conflict is over and then possibly let them return to their country of origin. As such, Syrians have a right to reside in the country but are denied the prospect of a long-term legal integration. They have access to limited rights compared to asylum-seekers in the ‘normal’ procedure, in particular as concerns access to education for children (on this point, see for instance, Human Rights Watch report) and access to employment. Although in January 2016, the Turkish government adopted a decision aimed at ensuring that Syrians can enter the labour market, the effects of this new regulation are yet to be seen in practice. Most importantly, Syrians in Turkey do not have access to refugee protection in its full sense, as enshrined in the Geneva Convention. For the reasons set out in more detail in the annex to this blog post, it is arguable that the ‘safe third country’ clause can only be interpreted as applying to countries which have ratified and fully apply the Geneva Convention.

Secondly, Turkey should respect the principle of non-refoulement, a prohibition on returning a person to a place where he or she faces a risk of persecution, torture, or inhuman or degrading treatment. However, several reports suggest that Turkey has engaged in refoulement and push-back practices throughout the years 1990s and 2000s. In particular, in November and December 2015, Human Rights Watch and Amnesty International denounced an increase in deportations, push-backs, arbitrary detentions and physical violence against asylum-seekers trying to cross the Turkish southern border coming from Syria or Iraq, or trying to enter Greece from Turkey, either by land or sea. This increase would coincide with the period leading up to and after the signing of the above-mentioned EU/Turkey deal.

Thirdly, in Turkey asylum-seekers and migrants in general, face a number of obstacles which may increase their risk of serious harm. In particular, Turkey has a record of treating asylum-seekers and refugees harshly in detention: episodes of torture or inhuman or degrading treatment have been reported by NGOs (Global Detention Project and Amnesty International among others) and condemned by the ECtHR in a series of judgments (see for instance, Abdolkhani and Karimnia v Turkey and the recent SA v Turkey, judgement of 15 December 2015). Furthermore, with reference to serious harm due to indiscriminate violence in a situation of conflict, the internal conflict between Turkey and the Kurdish rebels, which has escalated during the last year, may pose threats to the lives of asylum-seekers and refugees in the southeast of the country.

Therefore, it seems that Turkey does not fulfil many of the requirements for designation as a safe third country under the Procedures Directive. Even though the Samsom Plan does not mention this option, it seems now interesting to consider whether a similar policy could theoretically be implemented based on two different concepts entailed in the Procedures Directive, the concepts of ‘European safe third country’ and ‘first country of asylum’. The latter will be examined more in detail.

Is Turkey a ‘super-safe’ third country?

This wasn’t mentioned above, but the Asylum Procedures Directive also provides for a special category of ‘European safe third country’, which has been dubbed (sarcastically) the ‘super-safe’ third country concept. In this case, a Member State could have ‘no, or no full’ consideration of an asylum application – as foreseen by the Samsom Plan. The legal requirements for a country to be considered a ‘European safe third country’ are set out in Article 39(2) of the Directive:

a)      the ratification and full implementation of the Geneva Convention without any geographical limitation;
b)      the existence of an asylum procedure prescribed by law; and
c)       the ratification and full implementation of the ECHR.

Even though Turkey has in place an asylum procedure prescribed by law (Law on Foreigners and International Protection adopted in April 2013 is Turkey’s first-ever national legislation on asylum), is a party to the ECHR (even though one of the parties with the highest number of condemnations by the ECtHR for violations of this treaty) and has ratified the Geneva Convention, as mentioned above, it maintains a geographical limitation to the application of the Geneva Convention, excluding non-European asylum-seekers from the refugee status. For this reason Turkey could not even be considered a ‘European safe third country’.

Could Turkey be considered a ‘first country of asylum’?

Could then the notion of ‘first country of asylum’ apply to Turkey? Could asylum-seekers possibly be returned from Greece to Turkey based on the fact that Turkey is their first country of asylum?

Article 33(2)(b) of the Procedures Directive foresees the possibility for a Member State to deem an asylum application inadmissible if it considers a non-EU country to be a first country of asylum for a particular applicant. Article 35 establishes that a third country can be a first country of asylum in two cases:

a)      if the applicant has been recognised as a refugee in that country and can still avail himself or herself of that protection; or
b)      if the applicant otherwise enjoys sufficient protection in that country, including benefiting from the principle of non-refoulement.

Article 35 further specifies that in applying this concept Member States may take into account the legal requirements provided for by Article 38(1) - i.e. those used to establish whether a country is a safe third country. It also states that asylum-seekers ‘shall be allowed’ to argue that the principle cannot apply to their particular circumstances. Furthermore, they also have the right to appeal pursuant to Article 46 of the Procedures Directive (and stay on the territory during the application and at least at the outset of the appeal), as discussed above.

For the reasons set out in the annex to this blog post, option (a) arguably refers only to obtaining status under the Geneva Convention. Therefore Turkey cannot be considered a first country of asylum for a non-European asylum seeker, due to its geographical limitation on that Convention. On the other hand, option (b) might apply. In Turkey, indeed, non-European asylum seekers can, at least theoretically, have access to an alternative form of protection: the so-called ‘conditional refugee status’ (for applicants who would qualify as refugees under the Geneva Convention but who come from a non-European country) or the EU-inspired subsidiary protection. Moreover, as mentioned above, asylum seekers originating from Syria have access to a different form of temporary protection.

These three alternative forms of protection differ in terms of the level of rights their holders benefit from, which in all cases (and in particular in the case of Syrians benefiting from temporary protection) is lesser than the one recognised to ‘European refugees’ (for details on the content of these three alternative forms of protection, see the AIDA Country Report on Turkey). The question is: could these forms of protection be considered as ‘sufficient protection’? How can a Member State establish when protection is ‘sufficient’?

Article 35 provides two reference points, one being strictly mandatory, the other one being optional. The first one is the respect of the non-refoulement principle. Turkey is formally bound to the principle of non-refoulement, being a party to the ECHR and having incorporated the principle into Article 4 of its Law on Foreigners and International Protection as well as in Article 6 of its Temporary Protection Regulation. However, as mentioned above, the country has a historical record of refoulement practices and there are allegations of a recent intensification of push-backs and deportations of Syrians and other asylum-seekers. Therefore, Turkey does not seem to be fully compliant with the principle of non-refoulement in practice. But, in light of the fact that each asylum application must be examined individually based on the specific circumstances of the case, Member States might argue that the risk of non-refoulement could be assessed on a case-by-case basis in order to establish if that particular applicant enjoys sufficient protection in Turkey.

The second reference point mentioned by Article 35 is Article 38(1). In deciding whether a third country can be considered a first country of asylum, Member States may apply the same criteria they use for determining whether that country could be a safe third country. As discussed above, Turkey does not seem to comply with most of the safe third country legal requirements and, on this basis, it might be argued that in general it should not be considered a first country of asylum. However, because this is a ‘may’ clause, Member States have no obligation to apply Article 38(1) requirements to Article 35 cases and can simply ignore the possible link between the two concepts.

Therefore, although the possible application of the concept of first country of asylum to Turkey seems in general rather controversial, Member States might apply it on an individual basis, based on option (b). However, if they decide to do so, Member States’ authorities would have to conduct a case-by-case assessment, taking into due consideration the particular circumstances of each individual applicant in order to determine whether he or she enjoys sufficient protection in Turkey and does not risk being refoulĂ©(e). As discussed above, an individual examination of all asylum claims (including the applicant’s right to appeal against a negative decision) is incompatible with the extremely rapid procedure and systematic readmission mechanism envisaged by the Samsom Plan.

A change in EU law?

The above discussion is based on current EU legislation. It is, of course, possible in principle for the EU to amend that legislation via the usual process, or arguably via means of an ‘emergency’ measure on asylum pursuant to Article 78(3) TFEU. The previous use of Article 78(3), for a ‘relocation’ system, is being challenged by Hungary and Slovakia. (On the latter challenge, see discussion here; and on the general legal issues concerning that provision, see discussion here.) There might be some specific procedural issues about the use of Article 78(3) to establish the Samsom Plan, but the underlying issue is substantive: could EU law be changed (by either means) to set up a ‘return ferries’ process?

The answer is clearly: No. All EU asylum measures are subject to the general rules in Article 78(1) TFEU: ‘compliance with the principle of non-refoulement’, and acting ‘in accordance with the Geneva Convention…and other relevant treaties’. Also, all EU measures are subject to the EU Charter of Fundamental Rights, which must be interpreted consistently with the ECHR (Article 52(3) of the Charter).

It must follow that at the very least, the ECHR case law minimum standards discussed above must apply. So no revised EU law can provide for return of people coming from Turkey without some proper individual consideration of their claim that Turkey would not be a safe country for them; and there must be a right to appeal and stay in the country at least until the first-instance decision is made on this issue. To the extent that the Samsom Plan does not respect this irreducible core of human rights protection, it would be illegal.


Although it is remarkable that Turkey adopted a new comprehensive EU-inspired asylum legislation and is a state party to major human rights conventions, such as the European Convention on Human Rights, the Refugee Convention and the Convention against Torture, the way it has so far implemented its international human rights obligations appears to be still faulty. In particular, the right to asylum in Turkey cannot be considered as ‘fully established’, especially because of the still largely dysfunctional asylum system and the existing inequalities in access to protection and content of protection, which at the present moment are affecting Syrian refugees in particular.

For these reasons, the Samsom Plan proposing the systematic return of all asylum seekers from Greece to Turkey in exchange for increased refugee resettlement in Europe, appears to be not only very difficult to implement (due to both legal and practical obstacles), but also based on the doubtful presumption that Turkey may be (soon) considered a safe third country for refugees and asylum-seekers.

Furthermore, it is unfortunate that the EU and Turkey did not agree to fully apply the Geneva Convention for Turkey, and that there are no mechanisms of accountability in place for the EU institutions to report either in general upon Turkey’s compliance with international human rights standards or in particular to explain exactly how the EU’s money is being spent.

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

**Emanuela Roman is a PhD candidate in Human Rights at the University of Palermo and junior researcher at the Forum of International and European Research on Immigration (FIERI). This article was written during the period she is spending as a visiting researcher at the Vrije Universiteit Amsterdam, Faculty of Law, Migration Law Section. Emanuela would like to thank all her colleagues at the VU Migration Law Section, in particular Theodore Baird, Evelien Brouwer, Thomas Spijkerboer and Hemme Batijes for their precious comments and advise. The sole responsibility for the content of this article lies with the authors.

Annex I
Interpreting the ‘safe third country’ clause in the Procedures Directive

As noted above, Article 38 of the Asylum Procedures Directive says that in a safe third country it must be possible for the applicant ‘to request refugee status and … to receive protection in accordance with the Geneva Convention’. In my view, this can only refer to States which have ratified and fully apply the Convention; therefore it cannot apply to Turkey. I am grateful for a discussion with Daniel Thym on this issue – although I should note that he holds the opposite opinion.

First of all, this interpretation is supported by the legislative history of the text, which is set out in detail in Annex II. The original draft in 2002 made expressly clear that the clause could apply even if a State had not ratified the Convention. During negotiations that text was revised so that the final wording states that it must be possible to get status ‘in accordance with’ the Convention. Attempts by several Member States to make it clear that alternative types of protection besides full Convention refugee status could also trigger the clause were not successful.

Secondly, the ordinary meaning of the words ‘in accordance with’ in English is ‘in compliance with’, although the other language versions are equally valid. This is confirmed by the words ‘refugee status’: the full title of the Geneva Convention is the ‘Convention on the Status of Refugees’. How can one apply for ‘refugee status…in accordance with the Geneva Convention’ if the state concerned has not ratified, or does not fully apply, the ‘[Geneva] Convention on the Status of Refugees’? While the definitions clause refers to Member States as regards the definitions of ‘refugee’ and ‘refugee status’, this logically cannot be intended to apply to Article 38, since that Article only refers to applications made in non-EU states.

Thirdly, the a contrario rule supports this interpretation. Where the drafters of the Directive wanted to refer to the possibility of applying for an alternative form of protection, they did so expressly, as in Article 35(b) of the Directive. Admittedly Article 39, which refers more clearly to the geographical reservation of Turkey as a (failed) condition for the ‘super-safe’ countries rule to apply, points in the opposite direction. But to the extent that these two a contrario analyses simply cancel each other out, the interpretation in line with the legislative history and ordinary meaning should apply.

Similarly ‘recognised as a refugee’ under Article 35(a) of the Directive should be interpreted to refer to the Convention refugee status, in the absence of any indication that any alternative meaning is intended. However, Article 35(b) does clearly provide for an alternative option of designating a state as a ‘first country of asylum’ due to the existence of other forms of protection.

Annex II: Legislative history of the ‘safe third country’ clause

Commission proposal, 2002: explicitly provides in an Annex that a ‘safe third country’ can be a country which has not ratified the Geneva Convention
Council doc 6929/03 – a note indicates that the Council will start work looking at ‘safe’ country concepts. This plan is soon dropped; the Council looks first at Arts 1-22 instead.
Council doc 7214/03 – Annex unchanged at this point, no MS comments
Council docs 10064/03 and 10456/03 – ditto
Council doc 10722/03 – minor amendment to annex to state that Cartagena declaration countries must have a procedure compliant with the principles of the Geneva Convention
Council doc 11108/03 – no change
Council doc 11575/03 – annex shortened a little
Council doc 12281/03 – annex shortened significantly
Council doc 12734/03 – annex is simplified, but still provides for rules (same as in previous text) on when a ‘safe third country’ can be a country which has not ratified the Geneva Convention
Council doc 13369/03 – same text, but Germany now has a reservation linked to the ‘super-safe’ country clause, and Finland says the relevant clause could be deleted
Council doc 13901/03 – unchanged
Council doc 13902/03 – unchanged. The Presidency notes that delegations have inflexible positions on these provisions.
Council doc 14020/03 – ditto
Council doc 14182/03 – issue sent to the Council
Council doc 14330/03 – text unchanged
Council doc 14686/03 – text unchanged. But Spain suggests deleting the annex and having a short description of ‘safe third country’ in the main text, which is vague as to whether the state in question must have ratified the Convention
Council doc 14686/03 add 1 – Presidency proposes to delete the annex and have a short description of ‘safe third country’ in the main text, which only mentions the Convention as regards non-refoulement
Council doc 15153/03 – clause now in the main text, annex deleted. No change re Convention issue. DE still has reservation linked to ‘super-safe’ clause
Council docs 15153/03 rev 1 and 2 – amended to refer to ‘request recognition and be granted protection by that country or by the UNHCR as a refugee’. Spain wants to delete ‘as a refugee’. BE, NL and FI want to add express clause requiring ratification of the Convention and/or observation of the Convention. DE reservation is gone.
Council doc 15198/03 – unchanged
Council doc 15198/03 add 1 – UK wants to delete the whole sub-para
Council doc 6871/04 – redraft adds clause separate from main criteria for ‘safe third country’: ‘…Member States shall have regard to whether the third country has ratified the Geneva Convention…’ when assessing the application of those criteria. The clause in the main criteria now reads ‘request and be granted protection as a refugee in that country’. This deletes the reference to the UNHCR and makes it clear that it must be the country which grants refugee status.
Council doc 6954/04 – unchanged, but UK joined by EL/ES/NL/AT want to add the words ‘or other forms of international protection’ to the criteria
Council doc 7183/04 - unchanged, NL no longer supporting the UK position
Council doc 7184/04 - unchanged
Council doc 7184/1/04 – unchanged. UK and ES now want to change to ‘or another form of status which otherwise offers sufficient protection’ to the criteria. This is similar to final ‘first country of asylum’ clause.
Council doc 7484/04 – due to deadlock, Presidency proposes dropping clause on ‘request and be granted protection as a refugee’ to get a deal.
Council doc 7729/04 – ditto
Council doc 8166/04 – redraft retains ‘request and be granted protection as a refugee’ clause, drops requirement to take into account whether third State has ratified Convention
Council doc 8158/04 – same text and reservation as in 7184/1/04
Council doc 8413/04 – text amended to read:  ‘the possibility exists to request refugee status and, if found to be a refugee, to receive protection in accordance with the Geneva Convention’. UK and Spanish proposal rejected – link to Geneva Convention in fact made explicit instead

Council doc 8415/04 – as before, except the UK seeks to amend to read ‘in accordance with the principles of the Geneva Convention’. This is clearly rejected in the final version.


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

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

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

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

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