09 March 2016

Why the EU-Turkey Deal is Legal and a Step in the Right Direction

A sense of urgency and desperation had been building up among decision-makers in Brussels and national capitals over recent weeks. The acting Council Presidency warned that ‘we’re running out of time’, while German and Austrian politicians openly considered the move towards a domestic ‘Plan B’. In short: the Common European Asylum System (CEAS) and the Schengen Area were on the brink of collapse – and the effective closure of the Western Balkans route showed that there was little time left to thwart a vicious circle which would be a serious throwback to both EU integration and effective refugee protection. This danger has been averted for the moment: the EU managed to buy some time in order to find out whether the ‘deal’ with Turkey will deliver.

Pro-refugee NGOs were quick to castigate the deal for falling foul of the EU’s on legal standards and for being an anti-humanitarian solution, in particular insofar as forced returns to Turkey are concerned. Academics also present a critical outlook reiterating the legal criticism or criticising the EU for burden-shifting.

This blogpost will present a different perspective, although the critique highlights a number of valid concerns: the deal with Turkey distracts from the EU’s internal shortcomings; many poorer countries accommodate greater numbers than the EU; some actors, such as Viktor Orbán, support the plan, since they reject refugee protection as a matter of principle. But these caveats do not unmake the legal and conceptual value of the approach pursued by the EU: mass-influx scenarios require international cooperation.

The Territorial Dilemma of Refugee Protection

The international refugee regime as we know it today was designed in the post-war period when the idea of state sovereignty was all the rage. This explains why the Refugee Convention focuses on the protection of refugees within state territories. Its obligations kick in once a refugee has reached a safe haven; the Convention stays largely silent on the travel route. This may sound abstract, but it has direct implications for our topic.

Much of the recent criticism stems from the assumption that it is illegitimate to expect refugees to seek shelter elsewhere. That is incorrect. Ever since 1951, the Geneva Convention has emphasised that ‘the grant of asylum may place unduly heavy burdens on certain countries, and that a satisfactory solution … cannot therefore be achieved without international co-operation’ (Recital 4). It remains one of its principal shortcomings that the cooperation objective remained a declaration of intent. Attempts to remedy this by means of the ‘Convention Plus Initiative’ largely failed – notwithstanding enthusiastic support by international academics and the relentless efforts of UNHCR.

That is why the EU-Turkey deal is much more than burden-shifting. It is a building block for an international solution to a transnational problem. Many NGOs and academics may strive for an optimisation of refugee protection within the countries in which they are rooted, but the objective of the Geneva Convention can ultimately not be realised by a few countries alone – especially if we keep in mind that the alternative to an agreement between the EU and Turkey might be border closures across Europe and the collapse of the CEAS. That is not to say that the forthcoming agreement should not be scrutinised carefully, both politically and legally. But the underlying idea should be applauded – as the Executive Committee of UNHCR recognised indirectly almost thirty years ago:

‘The phenomenon of refugees … who move in an irregular manner from countries in which they have already found protection, in order to seek asylum or permanent resettlement elsewhere, is a matter of growing concern…. Where refugees … move in an irregular manner … they may be returned to that country if i) they are protected there against refoulement and ii) they are permitted to remain there and to be treated in accordance with recognized basic human standards…’

The Promise of EU-Turkey Cooperation

Political negotiations with Turkey present the EU with a number of problems which are specific to the European project. The EU has always taken pride in presenting itself as a normative avant-garde which overcomes the nation state internally and promotes world peace and human rights externally. It appeared as a normative power, in contrast to more realist players like the US or Russia. This idealism may always have masked underlying self-interests, but that does not unmake the conceptual challenge the EU-Turkey deal presents for the EU’s self-image. Negotiations with an autocratic regime and the return of Syrians to Turkey reveal unequivocally that the EU pursues (legitimate) self-interests apart from the worldwide optimisation of human rights.

Yet, we should be careful not to mistake the ‘new realism’ of the EU’s external migration policy as a slippage back into nationalistic paradigms, since the outline of the deal bears the signature of Angela Merkel much more than the footprint of Viktor Orbán. The EU will spend six billion euro on advancing the living conditions of more than 2.5 million refugees in Turkey (that is more than two thousand euro per capita); it will resettle a significant number of Syrians, albeit only in return for Turkey taking back similar numbers; the Western Balkans route may ‘have come to end’, but the EU will activate massive additional resources to support Greece and to prevent the CEAS from collapse.

None of the above is wrong per se. To the contrary: UNHCR has always called for more money, invited states to embark upon resettlement and to build a functioning pan- European asylum system. It does not come as a surprise, therefore, that UNHCR ‘welcomed’ various aspects of the forthcoming deal, while expressing its ‘concern’ about the procedural aspects of returns to Turkey (note that it did not condemn the returns as such, but rather asked for procedural safeguards). Experts on refugee law may dislike cross-policy issue linkages with the EU promising Turkey visa liberalisation in return for taking back asylum seekers, but to do so is standard practice in international relations.

Legal Pitfalls of the Safe-Third-Country Concept

The concept of ‘safe third country’ is well established in international custom and compatible with the Geneva Convention as a matter of principle. It is unproblematic, therefore, that the EU enshrined it in Article 38 of the Asylum Procedures Directive. It is less certain, however, whether Turkey meets the statutory threshold, not least since it ratified the Geneva Convention with a geographical limitation. In a blogpost building on an email exchange with myself, Steve Peers and Emanuela Roman presented an argument against Turkey as a safe third country and it may be worthwhile, therefore, to list the counter-arguments in order to allow readers to make up their mind.

Firstly, the systemic structure of the Directive distinguishes between ‘super-safe’ third countries (Article 39) and countries being just ‘safe’ (Article 38). Full ratification without geographical limitation is explicitly prescribed for the former (Article 39(2)(a)), while safe third countries must only grant protection ‘in accordance with’ (French: conformément; German: gemäß) the Geneva Convention (Article 38(1)(e)). Turkey can only qualify for being a safe third country under Article 38 which, in contrast to Article 39, requires an individual assessment of any asylum claim before a rejection.

Secondly, the ordinary meaning of ‘in accordance with’ is to follow or to obey a rule. To do so does not require prior ratification, neither semantically nor logically — as the example of the EU itself shows, since the EU never ratified the Geneva Convention and is not bound, therefore, qua international treaty law, only via Article 78(1) TFEU and international custom in accordance with ECJ case law. That is not to say, crucially, that the EU need not act ‘in accordance with’ the Geneva Convention. To the contrary: it is obliged to respect the Convention without having ratified it. In the Annex I of its original proposal on the predecessor provision of today’s Article 38, the Commission explicitly foresaw that also states not having ratified the Convention can be qualified as safe if their laws and practices are ‘in accordance with the abovementioned principles’.

Thirdly, the terms ‘refugee’ and ‘refugee status’ employed in Articles 35 and 38 Directive 2013/32/EU do not necessarily require prior ratification of the Convention either, since they are technical terms which are defined explicitly in Article 2(j), (d) to refer, ‘for the purposes of this Directive’, to third-country national who are recognised as being persecuted for the reasons listed in the Convention. Of course, state can afford this protection to someone without prior ratification. UNHCR does the same.

Finally, the drafting history, described meticulously by Steve Peers, does not support a different reading. To the contrary, the Council merged the original Commission proposal, which had explicitly foreseen that ratification was not required, into a general provision which remained deliberately vague on the question of ratification, while ensuring that the substance of the Convention must be fully complied with (the German proposal to explicitly prescribe full ratification was not followed). The British and Spanish demand that alternative means of protection should be sufficient does not support a different reading, since it was aimed at a lower threshold than the current statutory prescription in Article 38, in particular insofar as socioeconomic rights are concerned.

On the basis of the above-mentioned arguments, it is far too simple to claim that the EU-Turkey deal manifestly falls short of the EU’s own standards. A different position can be defended, in line with the opinion of the European Commission which stated explicitly in a recent communication that the safe third country concept does not require full ratification of the Convention — for as long as its requirements are met in practice. That, however, will have to be assessed carefully in relation to Turkey.

Applying the Criteria to Turkey

From the very beginning, cooperation between the EU and Turkey has served the laudable aim of improving the living conditions of Syrians and other refugees in Turkey. UNHCR recognises that Turkey now ‘provides protection and assistance for asylum seekers and refugees, regardless of the country of origin’ on the basis of the international protection law of April 2014. On 15 January 2016, Turkey adopted a regulation giving Syrians under temporary protection access to the labour market and effective support for an enhanced access to education is a priority of the EU-Turkey Action Plan. All this is highly relevant for European safe third country concept, since education and economic agency define the rights of refugees under the Geneva Convention.

Of course, there are reasons of concern, including reports about rejections at the border and/or indirect refoulement by means of harsh treatment. That could potentially fall foul of the non-refoulement obligations under Article 38(1)(c), (d) Directive 2013/32/EU, although a final conclusion would require a thorough assessment of both the facts and the law. One crucial legal question will be the degree of security necessary for the qualification as a safe third country, also considering that Article 38(2)(c) allows each applicant to rebut the presumption of security on an individual basis. This shows that 100% of security are not required (unlike for ‘super safe’ countries under Article 39).

In practice, it will be Greek authorities and courts which will have to assess whether Turkey meets that threshold, since the CEAS does not (yet) establish the option of supranational designation of safe third countries (a Commission Proposal wants to establish a common minimum list of safe countries or origin which, however, would apply only to Turkish nationals applying, not to third-country nationals on transit through Turkey). Of course, the ECJ may provide guidance to Greek courts on how to interpret EU law. The recent opinion of Advocate General Kokott of 8 March 2016 on the interaction of the safe third country concept and the Dublin III Regulation indicates that the Court of Justice will not unduly restrict the effet utile of Article 38 Directive 2013/32/EU.

Moreover, the European Court of Human Rights will exercise judicial supervision, although we should be careful not to overstate its impact. It is correct that it prohibits collective returns, but the precise meaning of this prohibition will be clarified in a forthcoming judgment on Lampedusa after the Grand Chamber accepted the Italian request for referral. In practice, the outcome of the case will most probably have little impact on future returns from Greece to Turkey for the simple reason that Article 38(2)(c) Directive 2013/32/EU requires an individual assessment anyway. It is also true that the ECtHR famously disqualified the Greek asylum system to fall short of human rights standards five years ago, but the situation has changed after years of substantial financial and operational support from the EU. This support is now being replicated in relation to Turkey.

The Challenge of Implementation

The enthusiasm prevalent among participants of the EU-Turkey summit may be explained, in part at least, by the degree of desperation before the meeting. Its outcome is a silver lining which indicates that the CEAS may survive the coming months. This is good news in itself and, yet, experience advises caution. Those who have followed the crisis of the CEAS unfolding over recent months may remember that a similar optimism prevailed when the Commission proclaimed its Agenda on Migration or when the Council agreed on the hotspot approach and the relocation schemes last summer. Unfortunately, political declarations and legislative activities do not always translate into success on the ground — and the same may still happen to the EU-Turkey deal.

Firstly, the constitutional structure of the EU means that Greek authorities will be responsible for implementing most aspects of the ‘deal.’ To be sure, they will receive massive operational, logistical and financial support from EASO and Frontex, but it will have to be, formally at least, Greek officers taking decisions. Moreover, each applicant will be able to seize Greek courts in accordance with Article 46(1)(a)(ii) and 46(5), (6) Directive 2013/32/EU before any actual return to Turkey can take place. Previous experience with the Greek asylum system shows that successful implementation is no foregone conclusion; it may turn out that Greek authorities are (again) overburdened.

Secondly, it remains to be seen which categories of people will be covered by the final agreement, since the blueprint which is publicly available indicates that returns to Turkey may only concern Syrians plus those ‘not in need of international protection’, thereby potentially excluding Iraqis, Eritreans or Afghans who, on the basis of an individual assessment, deserve international protection. To be sure, the EU-Turkey readmission agreement covers all asylum seekers without leave to remain, once their application has been declared inadmissible on the basis of the safe third country concept. That outcome, however, presupposes in line with Article 38(4) Directive 2013/32/EU that Turkey is willing to take back Iraqis requiring protection. If that was not the case, the forthcoming agreement would not cover a substantial number of arrivals in Greece.

Thirdly, migration control activities by states have a limited impact and the track record of Greek authorities is particularly nuanced. There will undoubtedly be evasion strategies with migrants trying to bypass returns to Turkey by absconding within Greece and/or moving on irregularly. Given that Article 38 establishes a rather cumbersome procedure, there will be ample opportunities to do so. Of course, the EU may decide to lower the standards for safe third countries in Directive 2013/32/EU by prescribing full compliance with non-refoulement, while demanding only basic socio-economic rights on the basis of a streamlined procedure, but such change would take time.

Fourthly, the deal with Turkey has always been, in part at least, a diversion tactic trying to overcome intra-European difficulties by means of international cooperation. We will learn whether Member States stand ready to overcome internal divisions after a summit which had not been prepared properly at ambassadorial level. The quotas for the resettlement scheme from Turkey remain to be determined and the financial and human resources to support both Turkey and Greece will have to be committed.


Confucius famously said that it does not matter how slowly you go as long as you do not stop. The EU has never been particularly fast, and implementing European rules on the ground has often taken years. However, this crisis may be different. It requires quick results if Member States are to retain their confidence that a European solution is feasible. The massive practical difficulties with the hotspot approach and the relocation scheme indicate that success is far from guaranteed. One can only hope that the EU-Turkey deal will deliver. Failure would herald the demise of the CEAS and may strike a lasting blow to the prospect of effective refugee protection on the European continent.

This blogpost is cross-posted on the EU Immigration and Asylum Law and Policy blog of the pan-European Odysseus-Network of which the author is a member.