“Does the principle of the autonomy of the EU legal order allow for a Member State to revoke the notification of its withdrawal from the EU?”

By Manolis Perakis

The question whether it would be lawful for a Member State to revoke the notification of withdrawal from the EU before the two-year lapse (laid down in the third paragraph of Article 50 TEU) has, clearly, vital political, economic and social implications. Even though it cuts to the core of the philosophy governing the “ever closer Union” and the role that States and private individuals play in it, it’s also a matter to which the provision itself does not give a definite answer. Moreover, there is no case law issued on the matter by the CJEU that could contribute to the interpretation of the provision, while the UK Supreme Court seems to have posited the irrevocability of a withdrawal notification in the famous judgment issued in the case of R (Miller) v Secretary of State for Exiting the European Union [UKSC 2016/0196], foregoing the opportunity to use the preliminary reference mechanism. This absence of relevant CJEU case-law is expected to change after the Inner House of the Court of Session in Scotland decided on 21.09.2018 to refer a relevant preliminary question (see O. Garner’s analysis),

Part of the literature expressing interesting and strong legal arguments has suggested the answer to the above question to the positive (e.g. P. Craig, S. Peers, O. Garner, A. Sari). Arguing the contrary and attempting a contribution to the academic debate, this post’s point of view is based on the fundamental principle that the EU legal order constitutes an “autonomous legal system”, which is governed by its own rules enacted by its own institutions and interpreted by its established Court (C-26/62, Van Gend en Loos). It is, therefore, argued that the legal lacuna regarding the provision of revoking the withdrawal notification, leads, according to the interpretation of the provision set out in Article 50 TEU – in line with the letter and spirit thereof – to the conclusion that permitting such a revocation would contradict the principle of autonomy, regardless of whether it is unilateral or initiated upon consensus.

The present post is divided into two parts. In the first part I approach the interpretation of Article 50 TEU through its letter and spirit. In the second part I develop my argumentation concerning the critical role that the fundamental principle of autonomy should play when attempting to find the true meaning of the provision and to fill the legal gap concerning the right to revoke the withdrawal notification.

I. The provision’s letter and spirit

i. The provision’s wording

The wording of the provision set out in Article 50 TEU lays down a framework of definitions and concepts, based on which one can convincingly argue that the correct interpretation of the article leaves no room for for an allowed revocation of the withdrawal’s notification.

More specifically, the irreversible nature of the Member State’s decision to withdraw is grammatically established from the first paragraph of the provision, by using the term “decides” instead of the terms “wishes” or “intends” that can be found in other provisions of the Treaties. Moreover, this decision, taken by the withdrawing Member State, is not subject to the approval or rejection on the part of the EU institutions or of the other Member States – as in cases such as those regarding tax measures and provided for in Article 65(4) TFEU – and no time limit for its revocation is set out in the provision.

Besides, the fact that the phrase stipulating that the Member State in question “shall notify the European Council of its intention” ends with the word “intention”, should not mislead. This specific term does not grammatically indicate a simple plan or wish. Indeed, in the same phrase it is stipulated that the Member State, which notifies its intention to exit from the EU, is the one that “has decided to withdraw”, while the meaning of the term “decision” of that Member State is defined in the immediately preceding paragraph of the Article.

Furthermore, the notification of the Member State’s decision to exit from the EU, initiates a process that, as follows from the letter of paras 2 and 4 of the provision, ascribes to it a special status, which could only be described as “transitional”. Thus, on the one hand, the second paragraph stipulates that the withdrawal agreement shall be negotiated in accordance with the procedure of Article 218(3) TFEU, which is set out to regulate negotiations between the Union and third countries or international organizations, and on the other hand, in the fourth paragraph, the Member State is characterized as “withdrawing” and thus forfeits the right to participate and vote in the relevant discussions of the European Council and the Council concerning the particular process. Consequently, the reversal of this process and the return to the initial status of full membership would require taking specific steps and actions that would have been stipulated and set out in detail in the said provision, should the EU’s Constitutional Legislature wished for them to be allowed, especially given that such a suggestion was actually set forward during the drafting of the corresponding provision of the Constitutional Treaty and was obviously rejected.

Finally, according to Article 50(5) TEU, if a State which has withdrawn from the Union asks to rejoin, its request shall be subject to the procedure followed by any European State that wants to become a member (Article 49 TEU). As is apparent from the letter of the paragraph, the Member State, which has withdrawn, does not enjoy any special status that shall facilitate its readmission to the Union, neither in terms of the time nor of the conditions set for the latter. Therefore, if the priority set by the EU’s Constitutional Legislature, when drafting Article 50 TEU, was to protect the withdrawing Member State from a potentially regrettable mistake, they would have, if anything, set out arrangements for an uncomplicated and swift readmission of the State to the Union, despite the lack of a provision regarding the revocation of the withdrawal notification; however, this was clearly not the case.

ii. The provision’s objective and spirit

From the announcement of the referendum’s outcome in favour of exiting from the EU, and especially following the date of the withdrawal notification (29.03.2017), both the United Kingdom and the EU have been faced with important consequences at all levels of political, economic and social activity, with “Brexit” causing uncertainty for millions of citizens. Furthermore, these consequences have multiplied and intensified during negotiations of the exit process, while at the same time a withdrawing EU Member State is present in all law-making and political activities (with the exception of those related to the withdrawal process), as well as in voting procedures, without having any direct interest in the progress and development of the Union.

Based on the above, it becomes clear why in Article 50(3) TEU the European Union’s Constitutional Legislature limits the duration of this complicated and sensitive period fraught with uncertainty, to a maximum of two years. Moreover, it is also evident why the ability of the withdrawing Member State to revoke the decision to exit from the EU at any time during that two-year period, would be contrary to  the spirit of Article 50 TEU and the purposes set therein.

Indeed, if the ability of the withdrawing Member State to revoke the withdrawal notification is accepted, how would all severed contractual, political and social relations breached because of the incoming withdrawal be re-established; how would Member States and private individuals be compensated for the costs that have been externalized due to the withdrawal notification (as S. Weatherill correctly points out); how would the esteem of the European integration be restored in the eyes of the rest of the world; and how would the broken trust amongst Member States be rebuilt? What would then prevent a strong Member State from exploiting the possibility of exiting from the EU in order to extract the greatest possible benefits and compromises within the Union, while knowing that it can back down at any time (as S. Smismans correctly points out)? And most importantly: what guarantees are provided that the millions of EU citizens, who would have spent months in a state of uncertainty and escalating tension, taking irreversible decisions and altering their life plans, will be compensated and their lives be restored, if, ultimately, it turns out that all these were in vain?

As evident from the hypothetical questions posed above, the EU’s Constitutional Legislature could not have provided for and approved of the possibility that the exit process could be used as a means of extortion or, in any case, at will by the withdrawing Member State, thus leaving unprotected those damaged and affected by the whole process. Therefore, this interpretation is not consistent with the spirit of Article 50 TEU and the purposes set by the Treaties.

The foregoing remarks are also useful when considering the argument that Article 50 TEU, as a provision of the Treaties, should be interpreted as discouraging a Member State’s withdrawal from the EU and the subsequent “shrinking” of the Union; or as allowing a new government to change policy and review its position, if this serves the Union’s integrity; or, finally, as protecting both the EU and the withdrawing Member State against decisions that were made on the basis of incorrect assumptions, by allowing it to rescind its declaration of withdrawal. Indeed, this argument overlooks the fundamental purpose of Article 50 TEU, which is not to discourage a Member State from withdrawing from the EU, if it has so decided in accordance with its political choices and constitutional requirements, but to protect the interests and values of the remaining Member States comprising the Union and of the millions of their citizens, as well as to ensure that there shall be no legal, economic and social chaos amongst the States during the transitional period both before and after the exit from the EU.

II. The autonomy of the EU legal order

i. Allocation of competences and institutional balance

One of the main and most potent arguments of the literature which advocates the possibility of revoking the withdrawal notification, is that, even if the view that the provision of Article 50 TEU does not allow for such a possibility is considered convincing, it should, nonetheless, be accepted that by means of the powerful consensus of all parties involved in the process, i.e. upon agreement between the withdrawing Member State and the other EU Member States, the exit process from the EU should be reversed (e.g. D. Sarmiento, P. Craig, S. Smismans). However, this view of the Member States’ common will, as being fully sovereign and able to effect changes upon concluded agreements, is found in international law and is far removed from the principles, characteristics and nature of the European Union legal order.

In particular, in the EU Court’s reasoning concerning the autonomy of the EU legal order, one can pinpoint many fundamental characteristics, without which its very nature would be compromised, one of them being the allocation of competences between the EU institutions and the Member States (see, inter alia, ECJ Opinion 1/78, ECJ Opinion 1/91, CJEU Opinion 2/13). In the latter the CJEU emphasized that:

164. For the purposes of that review, it must be noted that, as is apparent from paragraphs 160 to 162 above, the conditions to which accession is subject under the Treaties are intended, particularly, to ensure that accession does not affect the specific characteristics of the EU and EU law.

165. It should be borne in mind that these characteristics include those relating to the constitutional structure of the EU, which is seen in the principle of conferral of powers referred to in Articles 4(1) TEU and 5(1) and (2) TEU, and in the institutional framework established in Articles 13 TEU to 19 TEU.

Indeed, a very important differentiating factor between the EU legal order and the international one is the irreversible (except in the case of Treaty amendment), albeit continual and voluntary, transfer of competences from the Member States to the EU institutions, and the nature of the founding Treaties as a Constitutional Charter, which lays down the terms and conditions of such a transfer. Therefore, as a consequence of their voluntary subjection to the obligations of the founding Treaties, the Member States can no longer agree – urgently and consensually – on a temporary derogation from the provisions set out therein, either in the form of non-application, or in the form of an – outside the Treaties’ framework – amendment and addition of rules, especially in processes where the EU institutions play an active role.

Thus, in the EU legal order an agreement amongst all Member States, however powerful and unconditional it may be,  could never lead to a breach of the Treaties’ “constitutional” principles; to the sidelining of EU institutions and the principle of “institutional balance” (C-9/56); and, finally, to the breach of rules governing the allocation of competences between the Union and the Member States. For instance, it would be impossible for the Member States to decide, by common accord and disregarding EU institutions, the amendment of the Treaties without following the procedure of Article 48 TEU; to accept new members to the Union without the convening and action of the institutions provided for in Article 49 TEU; or, to reinterpret the original meaning of the Treaties, usurping the power assigned for this particular purpose to the CJEU (Article 344 TFEU).

For this reason, an interpretation of Article 50 TEU – according to which the EU institutions involved in the procedures laid down by the provision, would be deprived of the powers of review and decision that have been transferred to them through the unanimous agreement concluded to that end by the Member States – would not conform with the Treaties and the fundamental principle of autonomy of the EU legal order. This theory would be valid only if the revocation of the withdrawal notification was also accepted by the participating EU institutions, namely the European Council, the Council and the Parliament. However, such a case lies outside of the EU legal firmament, since no such power is conferred upon these institutions by the Treaties and no governed-by-law procedure is provided for its exercise. Therefore, any such act would be subject to annulment by the CJEU following the Article 263 TFEU procedure as violating the principle of conferral (Article 5 TEU).

ii. The theory on the application of the VCLT

Part of the literature (e.g. A. Sari, C. Closa) has argued that the omission on the part of the European Union’s Constitutional Legislature to provide for the possibility of revoking the withdrawal notification was not intentional, and that in view of this involuntary legal lacuna, the corresponding Articles of the Vienna Convention on the Law of the Treaties (VCLT), and most importantly Article 68, which allows a State withdrawing from an international agreement to revoke its pertinent decision without facing any consequences, should be applied or serve as an interpretative criterion. However, this position is not consistent with the autonomy of the EU legal order and the fundamental principles of EU law.

As a preliminary remark, it is worth mentioning that, as follows from para. 3 of Article 50 TEU, the relationships between the Union and the withdrawing Member State continue to be strictly governed by EU law even subsequent to the address of the notification, and up until the withdrawal process completion. Consequently, during that period within which the notification could have been revoked, the rights and obligations of the withdrawing State are governed by EU and not by international law.

However, this in no way means that the possibility of applying here the main agreement that codifies international customary practices, i.e. the VCLT, does not merit further exploration. Indeed, the EU Court has early on – in the case of Poulsen – expressed the judgment that “the European Community must respect international law in the exercise of its powers”, whilst in the subsequent case of Racke, the Court ruled that “… the rules of customary international law … are binding upon the Community institutions and form part of the Community legal order”. Finally, adopting the same stance as the one taken by the CJEU, the GCEU has remarked, in the case of Opel Austria, that: “the principle of good faith is a rule of customary international law … and is therefore binding on the Community”. At first reading, the aforementioned case law could lead to the conclusion that, eventually, customary international law finds application even within the EU legal order and, therefore, it could influence the interpretation and application of Article 50 TEU, if it can be found codified in widely accepted international agreements such as the VCLT and more specifically its Article 68.

Nevertheless, this conclusion is rather rushed for two reasons.

To begin with, it is worth noting that it is doubtful whether Article 68 VCLT codifies or not customary international law. Indeed, in the case of Espada Sanchez, the CJEU reiterated its settled judgment that “though the Vienna Convention on the Law of Treaties of 23 May 1969 does not bind either the European Union or all its Member States, that convention reflects the rules of customary international law which, as such, are binding upon the EU institutions and form part of the legal order of the European Union”. However, in the case of Brita, which the Court references in its reasoning, it made mention of “a series of provisions in that convention”, whilst an even more restrictive wording is used in the judicial decision issued in the Racke case (“a series of its provisions, including Article 62, reflect the rules of international law”), which is the most representative in relation to the VCLT. Therefore, it becomes clear that the CJEU does not presumptively regard the corpus of the VCLT’s provisions as reflecting customary international law, but each should be independently assessed.

Hence, the question is whether the provision of Article 68 VCLT specifically does, indeed, codify an international custom. It should be noted that, according to the prevailing view in the literature, Article 68 VCLT on the right, on the procedure, and on the requirements set for revoking the notification of withdrawal from an international agreement, neither constitutes a codification of an international custom, nor did it establish such a custom following the conclusion of the VCLT. Besides, the very form, context and letter of the relevant provisions render legally impossible the pro rata application of Article 68 VCLT in the case of Article 50 TEU, since, on the one hand, the former explicitly refers to the “notification” provided for in Articles 65 and 67 VCLT, and on the other hand, it introduces one regulation – amongst many – governing the conclusion, functioning, interpretation, denunciation and termination of an international agreement. Therefore, from a legal standpoint it would be incorrect to isolate Article 68 VCLT and “transfer” it to a completely different legal framework, as that of Article 50 TEU.

Moreover, there is an even more important and non-controversial legal reason to consider invoking a rule of an international custom or agreement in order to interpret or fill a lacuna in a Treaty provision as contrary to EU law.

Indeed, all the aforementioned judicial decisions regarding the application of international customary law in the EU legal order, and all other decisions in this series of EU Courts’ case law, concern – without exception – the interpretation, application and review of the validity of acts under secondary EU law or of international agreements concluded by the Union. They do not concern the validity or interpretation of primary EU law provisions, as these are interpreted exclusively according to the wording, purposes, development and framework of the Treaties.

The said distinction is based on the established principle that the existence of an international custom can be ascertained by the CJEU when codified in widely accepted international agreements, such as the VCLT or the United Nations Convention on the Law of the Sea (1982), and consequently, the international custom is equated with these international agreements in terms of its hierarchical status within the EU legal order, i.e. it takes precedence over the secondary but not the primary EU law. Indeed, in the Kadi case, the EU Court was clear that even the United Nations Charter, a primary example of an instrument codifying customary international law, “would have primacy over acts of secondary Community law, … that primacy at the level of Community law would not, however, extend to primary law”.

As evident from the preceding review of the case law, the proposed invocation of a customary international rule or an international agreement, aiming at the interpretation of a rule or the interpretative filling of a legal lacuna found in provisions laid down in the Treaties and not in secondary law, is not consistent with the autonomy of the EU legal order, as this is perceived and has been set out in detail by the CJEU. Consequently, the reasoning of the argument advocating the filling-in of the regulatory legal lacuna in Article 50 TEU – regarding the revocation of the withdrawal notification – by a corresponding application of provisions of the VCLT or of other international rules, is flawed, regardless of whether this lacuna is intentional or unintentional.