Taking the ‘Union’ out of ‘EU’: The EU-Turkey Statement on the Syrian Refugee Crisis as an Agreement Between States under International Law
Source: Ejil Talk
April 20, 2017
Almost one year after its conclusion, the Court of Justice of the European Union (CJEU) has eventually made clear the real nature of the ‘so-called’ EU-Turkey Statement.
The ‘Statement’ is a document that was primarily aimed at preventing irregular migrants reaching the EU from Turkey, and established a resettlement mechanism based on the transfer of one vulnerable Syrian from Turkey to the EU “for every irregular Syrian being returned to Turkey from Greek islands”. The case was brought by three asylum seekers who arrived in Greece by boat and risked being returned to Turkey pursuant to this Statement if their request for asylum was rejected. They asked the Court to annul what they identified as an “agreement concluded between the European Council and the Republic of Turkey” (see CJEU, Orders of 28 February 2017, Cases NF v European Council, T 192/16; NG v European Council, T-193/16; NM v European Council, T-257/16).
According to the CJEU, the ‘EU-Turkey’ Statement is a non-EU agreement. In fact, it is a European agreement between EU Member States and Turkey, which was made at the margin of the European Council’s meeting held in March 2016. As such, according to Article 263 of the Treaty on the Functioning of the European Union (TFEU), the CJEU lacks jurisdiction to review its legitimacy, especially in relation to the provisions set out for the conclusion of international treaties by the EU (similarly, CJEU, 30 June 1993, Parliament v Council and Commission, C-181/91 and C-248/91.).
This expected (?) conclusion (see S. Peers here) raises more questions than it answers. After a brief analysis of the CJEU’s order at least two points deserve attention. Firstly, were all aspects of the Statement duly considered in order to exclude the possibility that this is an agreement of the EU with a third country? Secondly, in light of customary international law of treaties, is a different reading of the EU’s involvement possible? In fact, it seems that in the Orders being discussed here, the CJEU has clearly departed from the application of the customary international law of treaties. Hence, from an international law perspective, the nature and the effects of the ‘EU-Turkey’ agreement are still open to debate.
An Agreement of ‘an International Organization’ or an Agreement of its Member States?
Since the beginning of the cooperation with Turkey in the field of migration, the EU has played a role as an independent actor. As a subject of international law, in all negotiations the EU has expressed its own autonomous willingness as well as its Member States’ readiness to reach an agreement to solve the Syrian refugees ‘crisis’ and to reduce irregular migration to Europe. Not only did the EU use its international role to speed up this cooperation, it also included this new relation with Turkey in the broad context of Ankara’s path towards EU membership. Significantly, in all meetings between the EU Heads of State or Government and Turkish authorities, the discussion was focused on the “relations between the EU and the Republic of Turkey”. While special attention was afforded to the protection of the external borders “of the Union”, an active involvement of the President of the European Council was envisaged (i.e. an institution of the EU – Article 15 of the TEU – and not an organ at Member States’ disposal). Moreover, readmission of migrants not in need of international protection to Turkey was supposed to be based on the Agreement between the EU and that country signed in 2013 (OJ 2014 L 134, p. 3).
In this context, the ‘EU-Turkey’ Statement (or more correctly, the EU Member States-Turkey agreement?) was revealed to the world on the European Council’s website as Press Release No. 144/16 with the express recognition that the EU and the Republic of Turkey had decided to end irregular migration headed to the Union. Until the CJEU’s order, no doubts had emerged on the Statement as an international agreement despite its peculiar form (see here). Turkey’s counterpart seemed clear as well. Indeed, the applicants identified the Statement as an act attributable to the European Council that, in its capacity as an EU institution, bound the Union to specific international commitments in the field of migration.
As shown by the Orders, the nature of the ‘EU-Turkey’ Statement was controversial within the CJEU itself. The Court asked the European Council, the Council and the Commission whether the meeting of 18 March 2016 had led to “a written agreement” and how the obligations described in the Statement were negotiated and eventually reached. Interestingly, the European Council denied that, on that occasion, the EU concluded an agreement or a treaty as provided by Article 218 TFEU or in the sense of Article 2(1)(a) of the Vienna Convention on the Law of Treaties of 23 May 1969. It deemed the Statement to be the “fruit of an international dialogue” between Turkey and EU Member States during an “international meeting”. For the Commission, in turn, the Statement planned the adoption of “future” measures showing only a political commitment rather than generating any legally binding effect (also the EP Legal service adopted the same approach in 2016). However, for the Commission, this engagement seemed to directly involve the EU insofar as the President of the European Council and the President of the Commission took part in this political’ negotiation.
Taking the EU out of the EU-Turkey Statement
To reach its conclusion that the EU-Turkey Statement was not an international agreement of the EU, the CJEU relied on the need to understand whether “an EU institution sought to conclude the international agreement at issue”. However, in the attempt to identify the authors of the agreement, the Court gave an overriding weight to these purported authors’ views rather than to the Statement’s content.
From an international law perspective, in adopting such an approach the CJEU has disregarded the customary international law on the interpretation of treaties (see Article 31 of the Vienna Convention), thus failing to interpret the Statement “in accordance with the ordinary meaning to be given to [its] terms”. While the Court has already relied on these rules on other occasions (for example, CJEU, 21 December 2016, Council v Frente Polisario, C-104/16 P, at 86), the reasoning seems aimed at taking the EU out of the ‘EU-Turkey’ agreement, thus preventing the agreement being reviewed by the Court itself. It is evident from the 18 March 2016’s Press Release No. 144/16 that the Statement is the result of a meeting between the ‘Members of the European Council’ and Turkey. After that meeting, it was ‘the EU’ that agreed a list of mutual obligations with that third country. While these expressions imply a proactive role of the Union, the European Council contrasted this evidence with a need for simplification in order that ‘the general public’ would understand the press release. However, there is nothing complex in the use of the term ‘Heads of State or Government of the EU countries’. Despite this, the CJEU was nonetheless ready to accept the inappropriateness of the language used by the EU press service to change the nature of the Statement from an agreement of the EU to an ‘indefinite’ commitment between States. At the same time, relying on some preparatory documents (e.g. the note of the ‘Working Programme of the Protocol service’ of the Directorate-General ‘Administration’ of the Council), the Court also took the European Council out of the picture. In its view, the Statement had not been decided ‘during the meeting of that institution’ but at the end of the parallel ‘international summit’ organised to discuss immigration issues with Turkey. As such, the reasoning is completed by the attempt to make a distinction between the EU Member States as a ‘group of States’ and the organisation to which they belong.
As a result, the European Council did not adopt ‘a decision to conclude an agreement with the Turkish Government in the name of the European Union’. Being irrelevant for the purpose of identifying its jurisdiction, the CJEU did not specify whether the commitments that emerged during the 2016 meeting were legally binding or not . To this end, it only stated that, “even supposing that an international agreement could have been informally concluded”, the Statement is an agreement between States.
Possible Alternative Readings?
In contrast to EU institutions’ views and following the CJEU’s supposition, several elements related to the ‘context, the object and the scope’ of the Statement support the view that, during the meeting of the 18 March 2016, an international agreement was concluded and this bound (also) the Union.
While it is well known that an international agreement may also take the form of verbal notes or a press release (see ICJ, Maritime Delimitation and Territorial Questions between Qatar and Bahrain, 1 July 1994), the Statement did not reach a general compromise on how to manage migration movements between Turkey and the EU. Neither did it postpone the negotiation of a detailed agreement in a subsequent occasion. Thus, the EU institutions’ belief that the Statement was simply a political commitment does not correspond to the nature of the obligations finally agreed. The reason lies in the Statement’s content and to the ordinary meaning of the language used therein. The Statement was indeed intended to produce legal effects for the parties thanks to the identification of a few operational points.
This is evident for the EU. In fact, these legal effects were not directed (only) at Member States as sovereign entities but also to the Union, which has been called upon expressly to implement the agreement. In other words, Member States aimed to bind the EU as the international organisation directly involved in the management of the “migratory crisis” through the European Council. It is no coincidence that, as reported by the EU Commission, in its meeting held on 15 December 2016, the European Council “reiterated its commitment” to the “EU-Turkey” Statement in order to grant its “full and non-discriminatory implementation” and “endorsed the Joint Action Plan” elaborated to this end. It is also significant, for example, that Greek authorities refer to the Statement as a different legal basis for carrying out individual returns to Turkey (in parallel with its own bilateral readmission protocol with Turkey and to the 2013 EU-Turkey readmission agreement).
Having regard to the customary international law of treaties and to the (rare) possibility that the Orders will be challenged, at least two alternative readings may be outlined.
A first possibility is connected to the effects that an international treaty produces on third parties. Firstly, even if we accept the CJEU’s analysis of the authors of the Statement, the agreed obligations involved the EU. To begin with, the Statement builds on the joint action plan that Turkey and the EU had already agreed in October 2015. All costs for implementing the agreement are covered by EU funding, i.e. one of the tools through which the Union acts as an independent and autonomous actor. In addition to the regulatory framework already in place to financing return operations, the EU is also involved on the ground through its Agencies that act in the Union’s name. Finally, the EU Commission is regularly reporting on the implementation of the Statement, showing that it produced legal biding obligations for the EU and that the Union is participating on the ground along with Member States.
As is known, under customary international law of treaties a treaty cannot ordinarily produce effects on third parties. If we accept that the Statement is an agreement between States and certainly not a trilateral international agreement (i.e. Member States-Turkey-EU) the EU was not obliged to respect the measures agreed therein. Instead, having regard to its implementation, EU institutions acted as if it was. They did in the belief that the obligations contained in the Statement bound the Union. Consequently, by adopting what may now be identified as unilateral actions, they have created at least a legitimate expectation in Turkey that the EU was legally committed to fulfil the 2016 agreement. Put this way, any (additional) public declaration aimed to accept expressly obligations contained in the Statement, in line with the prevailing international law principles on unilateral acts, might have been deemed unnecessary. As a result, the Statement may have produced legal effects on the Union. If it cannot be reviewed per se, measures adopted to respect the obligations contained therein are acts of the EU and might nonetheless be challenged by asylum seekers.
A second possible reading is connected to the competences in the field of migration. In the attempt to involve the Union, its Members States used the European Council to reach the agreement with Turkey. Thus, instead of acting outside the European Council, the Heads of State and Government acted in their capacity as Members of an international organisation in the framework of one of its organs and adopted obligations for the same organisation outside the procedures established by the EU Treaties.
In fact, Member States “acted collectively” in the framework of the European Council because they did exercise competences that are shared with the EU. In this kind of negotiations, their involvement “as States” is necessary because readmissions and returns are dependent on their decisions, while the EU is (still) unable to implement this part of the agreement.
The CJEU disregarded this complex system of competences between the EU and its Member States. In fact, it sets a clear division in “the capacity to act” while a similar clear separation of powers in dealing with the migration “crisis” does not really exist. This aspect is evident in the Statement’s ambiguous language. If the Union was not directly involved via the European Council, there was no need for Turkey to meet all Member States. A meeting and an agreement with those Member States willing to set up specific obligations in this field would have sufficed. Hence, instead of acting outside the EU institutional framework, Member States may have used the European Council ultra vires.
Both readings show that alternative perspectives were available to the CJEU. Its formalist approach instead seems to validate a new practice followed by Member States in deciding outside the EU Treaties but for the EU itself. As a result, from now on the EU risks to be increasingly unable to exercise fully its competences because it is obliged to operate within the ‘framework’ set up in parallel by the “Heads of State and Government” of its Member States. Is the EU still a Union?