In this comment I discuss the proposals for a new settlement between the UK and the EU. As I will show, the precise nature of the draft agreement is unclear. This legal instrument raises difficult issues of both EU and public international law and could potentially cause serious uncertainty or even a constitutional crisis. Press reports have missed this legal complexity. Ministerial statements have been silent about it. While the issues require a much fuller treatment than I can give here (and the substantive issues of EU law have already been expertly discussed in a blog post by Professor Steve Peers) I still hope that even a preliminary discussion of some of the relevant constitutional issues may cast some light on matters of considerable public interest.
The basic facts are as follows. The President of the European Council Donald Tusk has published his draft proposals for the settlement between the EU and the United Kingdom. A draft of the ‘Conclusions’ of the European Council has also been informally published. These proposals will be discussed at the next meeting of the European Council on 18 and 19 February in Brussels. It is widely reported that if there is an agreement, the most likely date for the United Kingdom’s referendum would be June 23 of this year.
In the United Kingdom this has been a matter of great political and constitutional significance and the government has been very active. In preparation for the EU referendum the European Union Referendum Act 2015, received the Royal Assent on 17 December 2015. Following the recommendation of the Electoral Commission, the question of the referendum according to s. 1(4) will now be: “Should the United Kingdom remain a member of the European Union or leave the European Union?” The two alternative answers that will appear on the ballot papers will be: “Remain a member of the European Union” and “Leave the European Union”. The Act introduces also detailed financial and campaigning restrictions, supplementing those of the Political Parties, Elections and Referendums Act 2000. On February 1 the government commenced the remaining provisions of the Act by way of the European Union Referendum Act 2015 (Commencement) Regulations 2016.
On 5 February the Electoral Commission published its overview guidance on the conduct of campaigns. There are in place some very strict rules on campaign finance and a highly regulated process. There will be a ten week period of campaigning. If the vote is for “Remain”, then the new settlement will take effect between the UK and the EU.
The legal nature of the settlement
I now turn to the legal nature of this settlement. The settlement is not an EU Treaty in the usual sense. As is well known, the relations of the UK with the EU are a matter of EU treaties. These are constituent instruments of an international organisation and are therefore subject to special rules of international law (for example, under the Vienna Convention on the law of Treaties unless the constituent instrument says otherwise, a reservation requires the acceptance of the competent organ of that organization). International law, however, is almost never deployed in European Union matters. Pieter van Kuijper has calculated that there are only about forty cases where the Court of Justice deploys public international law. As is well known the EU treaties have created a highly integrated legal order, which the Court of Justice calls a ‘new legal order’ and which has direct effect and primacy in all domestic legal orders. It is applied under the supervision of the Court of Justice of the EU and the courts of the member states, which nevertheless enjoy “procedural autonomy”.
This legal order has produced voluminous case law on every aspect of EU law. This is a very different legal architecture to that of an ordinary international organisation. A feature of EU law is that its Treaties can only be changed according to Article 48 TEU. The Court has explicitly said that this is an exclusive procedure in Case 43/75 Defrenne v SABENA (II)  ECR 456, at pars. 57-58 where the Court said that: “the Resolution of the Member States of 30 December 1961 was ineffective to make any valid modification of the time-limit fixed by the Treaty. In fact, apart from any specific provisions, the Treaty can only be modified by means of the amendment procedure carried out in accordance with Article 236”.
So when other member states refused to amend the treaties in response to Mr. Cameron’s demands for a renegotiation of the UK’s position, they effectively ended any such possibility. In the end it was accepted that a formal treaty amendment was not possible. Hence, the proposed settlement does not follow the procedures of treaty amendment envisaged by Article 48 TEU. It is a decision that will be reached within the European Council. The settlement, however, will not be a decision of the European Council either. If it were, it could not have had the desired effect. Under the Treaties the European Council does not have “law-making powers” (see Article 15(1) TEU).
Mr Tusk has proposed, therefore, a very subtle arrangement which aims to satisfy both the UK government and its EU partners. His plan is for the member states to reach a “Decision” in their capacity as representatives of states and governments in international law, not in their collective identity as the European Council. This will be an intergovernmental agreement.
There is some precedent for this arrangement. What is being proposed is similar to what happened with Denmark in 1992 and Ireland in 2009, when both countries held referendums that rejected the Maastricht and Lisbon Treaties respectively (something well discussed in a post by Philip Moser QC). They then proceeded to strike separate agreements allowing for some special arrangements. The Danish arrangement, was part of the process of ratification of the Maastricht Treaty by Denmark and similarly, the 2009 Decision on “the Concerns of the Irish People on the Treaty of Lisbon” was related to the ratification of that Treaty by Ireland. Ireland’s concerns became part of the Treaties by way of a formal amendment at the same time as Croatia joined.
Nevertheless, the Court of Justice has not recognised these Decisions as part of the EU treaties. This is understandable because they were not agreed and ratified in the same way. The Court referred to the Denmark Decision in one case (Case C-135/08, Rottman in 2010) where it said that it “had to be taken into consideration” as an instrument for the interpretation of the Treaty, but nothing more than that. Reflecting this reality the Draft decision now states that it is merely a “clarification” which “will have to be taken into consideration as being an instrument for the interpretation of the Treaties”. It also notes that a future treaty amendment will incorporate the “substance” of the section on economic governance, presumably following further intergovernmental negotiation. So it is clear that this is not a formal amendment of the EU treaties under the EU’s own legal order.
Is it binding in international law? The European Council’s Legal Opinion
The key then to the legal status of the settlement is its status in international law. This is the view taken by the recently released (by the House of Commons European Scrutiny Committee) Opinion of the Legal Counsel of the European Council, dated 8 February.
This Opinion is a relatively short document (about nine pages) without much detail. There may be other opinions available, but I have not seen them. It concludes that the decision is “an instrument of international law by which the 28 Member States agree on a joint interpretation of certain provisions of the EU Treaties and on principles and arrangements for action in related circumstances” (par. 4). It then states that it is an agreement according to Article 31 of the Vienna Convention on the Law of Treaties, which provides certain principles of interpretation for treaties, namely that they are to be interpreted “in good faith, in accordance with the ordinary meaning to be given to the terms in their content and in the light of its object and purpose”. Article 31 also provides (in its paragraph 2) that agreements at the time of a treaty’s conclusion regarding its interpretation are part of the relevant “context” and (in paragraph 3) that subsequent agreements regarding interpretation are to be “taken into account”.
The Opinion then refers to Article 11 of the Vienna Convention, which determines that there is no particular form for creating binding treaties. The Opinion notes that the UK Decision requires no formality and that member states will be bound merely by their “common accord”. The legal effect therefore is taken to be that the Decision will be “legally binding in international law for the Member States”. The Opinion then adds that this legally binding agreement “does not amend the EU Treaties, which can be achieved only following the specific procedures provided for this purpose by the Treaties themselves” (par. 13). So for the Opinion the Decision is a treaty in international law but one which does not amend the EU treaties. This is not because of its contents but because of its method of coming into being.
I believe that the Opinion is obscure on two important matters. The first is on the distinction between two different senses of ‘agreement’. The Vienna Convention introduces a distinction, which the Opinion fails to note. The first and obvious sense of ‘agreement’ is that of Treaty which is what Article 11 of the Vienna Convention addresses. A second sense, however, is of what we could call an interpretive agreement, which need not be a Treaty in the strict sense. Declarations or conclusions reached in the context of various diplomatic conferences or meetings could well be such interpretive agreements. In that sense they will be legally relevant or even binding in that they will assist in the interpretation of an existing treaty between the same parties. Article 31(3) of the Vienna Convention makes that clear when it says that “(a) any subsequent agreement between the parties regarding the interpretation of the treaty or the application of its provisions” shall be taken into account alongside other matters and the relevant context.
By definition such an interpretive agreement is not a new treaty between the parties on the same subject matter, because if it were it would be amending the earlier treaty. So it is possible that an interpretive agreement may either be a treaty about a different subject matter that touches on the earlier treaty or simply a new and informal agreement of the parties over some ambiguous issue of interpretation of the existing treaty, exhibiting no intention by those parties to be formally bound by it.
Why does this matter? Because it allows for a sense of “legally binding” which falls short of being a treaty. This is the sense suggested, for example by the ECJ, when it mentioned the Denmark agreement. As it stands, the Opinion gives the impression that there are only two possibilities: that the agreement is a treaty and is legally binding or that it is neither. It suggests that for it to be legally binding the draft Decision should be either an EU Treaty (under Article 48 TEU) or a public international law treaty (under Art. 11 of the Vienna Convention).
Under the Vienna Convention, however, there is a third possibility. The UK Decision may well be an agreement which is not a treaty, but may still be legally binding in the sense of being relevant to the interpretation of a Treaty under Article 31(3) as a subsequent agreement between the parties regarding interpretation (and obviously not under Art 31(2) in the “context” of the conclusion of a treaty, for there is none here). Such agreements are admittedly rare events but are clearly possible. It is in this spirit, I think, that the Court of Justice considered the Denmark agreement in Rottman as merely relevant to interpretation, but not part of the treaties. Such an interpretive agreement does not amend, by definition, the rules it is aimed at interpreting.
The reason why this is so important has to do with the second obscure point in the Opinion. The Opinion takes the view that a public international law treaty between the same parties over the same subject matter (e.g. economic governance, sovereignty and free movement of workers) cannot change EU law, because EU law provides for a definitive process for its amendment. It is not explained why or how this is the case. Yet again the Vienna Convention gives a different answer. The process does not matter. Under Article 30 (3) a new treaty amends an earlier one between the same parties over the same subject matter. It says that: “When all the parties to the earlier treaty are parties also to the later treaty but the earlier treaty is not terminated or suspended in operation under article 59, the earlier treaty applies only to the extent that its provisions are compatible with those of the latter treaty”. So just like equal Acts of Parliament, the later one impliedly repeals the earlier one. It is a matter of the Vienna Convention, but also a matter of simple logic.
The Opinion does not address this point at all. Perhaps it works under the well-known theory of the uniqueness and “autonomy” of EU law. On some occasions the Court of Justice has spoken of this autonomy of EU law from public international law. The doctrine, however, has been challenged by all national courts. For example, the German Constitutional Court in its OMT judgment considered the EU treaties as normal treaties of public international law – and will meet in a few weeks to examine if it is going to declare the OMT programme unlawful in Germany. The UK Supreme Court assesses EU law on the basis of its own “constitutional instruments”. So I am not sure that the certainty exhibited by the Legal Opinion on the autonomy of EU law is justified.
If the Opinion is wrong about this, then whether the new treaty (the Decision) amends the EU treaties or not will be a matter of substance and will depend on what the new treaty says. Its formal status would be sufficient. Everything will turn on whether what it says is actually compatible with the EU treaties. Treaties are all equal, as we have seen. This opens the possibility that the EU treaties will be amended by this Decision almost by accident, even if this was not what the parties intended. If the provisions they bring about are incompatible with the earlier treaties, then the earlier provisions will have to be set aside. The only safeguard they seem to have put in place (in the leaked draft ‘Conclusions‘) is a declaration that “the content of the Decision is fully compatible with the Treaties” is not sufficient to prevent this from happening. This declaration is not in the text of the Decision but in its interpretive ”context” and it therefore follows that the safeguard of Article 30(2) of the Vienna Convention does not apply. This Article provides: “When a treaty specifies that it is subject to, or that it is not to be considered as incompatible with, an earlier or later treaty, the provisions of that other treaty prevail”. There is no such statement in the body of the Decision, but only in an accompanying declaration. But even if the statement became part of the body of the Decision, it would still be a matter of interpretation in what sense earlier provisions “prevailed” over the new provisions. If the later ones are more specific, a court may find there was no conflict, but a further development of the law. Whether the earlier provisions survive will depend on what the later provisions actually said.
Why does it matter?
So my view is that the Legal Opinion has simplified the legal questions far too much and has opened up the possibility of a new constitutional crisis. I will now explain how it may unfold. As far as I can see the possibilities for a “legally binding” instrument regarding the UK settlement are not two, but four:
I. EU Treaty
II. Public International Law Treaty
III. “Context” Agreement Under 31(2)(b) which was made between all the parties ‘in connexion with the conclusion of a treaty’
IV. Subsequent Agreement Under 31(3)(a) which was made subsequently between the parties regarding the interpretation of the treaty
The Ireland and Denmark precedents most likely fall under III and were “Context” Agreements, because they were created at the time when a EU treaty was being ratified. The UK draft decision must fall under IV as a Subsequent Agreement, because no EU treaty is currently under negotiation.
In my view, the settlement as drafted is not intended to produce legal obligations in the way of a binding treaty, under I or II. The draft Decision calls for a future EU amendment that will achieve treaty change after further negotiation. So the parties are saying that they are not ready to bind themselves by way of a treaty. There is no agreement to bring these changes now. It is thus merely an interpretive agreement that is aimed at outlining future policy on some things and is guiding interpretation about existing commitments wherever there is ambiguity, following the Ireland and Denmark examples. It is legally binding only in the weak sense that, as a matter of international law, it provides material that is relevant to the interpretation of existing treaties under Article 31 of the Vienna Convention.
Why does all this legal technicality matter? Getting it wrong, and taking this to be a treaty, matters for a constitutional reason and for a political reason. I start with the political reason. The Government has given no indication of these legal complexities. The Prime Minister stated before Parliament that:
“These changes will be binding in international law, and will be deposited at the UN. They cannot be changed without the unanimous agreement of every EU country—and that includes Britain. So when I said I wanted change that is legally binding and irreversible, that is what I have got.”
The Prime Minister seems to suggest that the United Kingdom considers this to be a legally binding treaty, without question. There is no suggestion that it may be an agreement regarding the interpretation of pre-existing treaties. It is only treaties that are deposited with the UN. I am not sure that this description meets with the agreement of the other parties, however. The leaked draft of the Conclusions only includes a declaration to the effect that “this Decision is legally binding, and may be amended or repealed only by common accord of the Heads of State or Government of the Member States of the European Union”. Of course, the UK cannot make this a treaty unilaterally. If the other parties, having received their own legal advice, consider the settlement to be an interpretive agreement alone, which is legally binding only in the sense given by Article 31, then they may wish to say so clearly. If so, the Prime Minister will not have achieved a new treaty. If they do not say so clearly, it will be a matter for interpretation. If both sides insist, there can be no agreement. This ambiguity may still prevent a deal being reached.
The constitutional reason of getting it wrong is even more significant, however. Let’s assume that the Decision is agreed as it is. Let us also assume that it is taken to be a Treaty by all member states, as the UK wants. Most of its provisions do not compete directly with provisions of the EU treaties. They only say that there should be some law reform in the future. But some provisions do compete with the EU treaty. I shall use as my example one proposition concerning the Eurozone, which requires the following:
“Discrimination between natural or legal persons based on the official currency of the Member State, or, as the case may be, the currency that has legal tender in the Member State, where they are established is prohibited. Any difference of treatment must be based on objective reasons.”
This looks like a clear, if wide-ranging, legal obligation of all member states. All financial regulation is caught by it. It could even have direct effect, if it were part of the EU treaties. If this is part of a new treaty, it could be taken to have effectively supplemented the EU treaties. If so, there are two alternatives about what this means.
If the European Council’s legal Opinion is accepted and EU law is considered to remain intact by this new treaty (either because it is “autonomous” or for another yet unstated reason), this will have created two separate legal regimes about currency discrimination: the one under the free movement of services and capital, under the current EU treaties, and the one created by this new international law treaty. The new rule will have to apply between the parties, but will not be enforced by the ECJ nor will it develop direct effects in domestic courts. Both the ECJ and the domestic courts will have to follow the old rules. But the UK will justifiably claim that the rules have changed. The result is complete legal insecurity.
If the rival view I put forward above about the relations between treaties under Article 30 of the Vienna Convention is accepted, then the new regime will have replaced the framework of the free movement provisions. This will soon, no doubt, become the UK’s position. The old rules will be gone, but the new rules will still not have direct effect. The Court of Justice would not have the power to enforce them, since the new treaty says nothing about the Court of Justice or direct effect.
In either case, the result would be hopeless confusion. No one would be sure about the applicable law on discrimination over currencies. What would the European Court of Justice do? It would have to choose between ignoring the public international law treaty newly created, or accepting it and entirely fragmenting the EU legal order. And if the Court ignored the public international law regime, the United Kingdom would then have the power to enforce the international law obligations by whatever other means were available to it by public international law (and by implementing the treaty domestically, of which see below). European Union law on banking regulation would have become incoherent.
There is also a worst case scenario. This is the scenario whereby the British government implements the new treaty by way of domestic law, i.e. by an Act of Parliament. This will mean that the United Kingdom courts will have to apply the new settlement domestically as law that is superior to all EU law. They would do so not only on the technical legal ground that this will be a new Act of Parliament explicitly repealing the European Communities Act 1972, but also on the substantive ground that as a new legally binding treaty of international law it reflects the desire of all the member states of the European Union to give up uniformity and to create a parallel legal arrangement that supplements the EU treaties outside the EU legal framework. The UK courts will thus have solid constitutional reasons to follow the new settlement, in a claim brought perhaps by a London based financial institution against some banking rule emanating from Brussels.
The EU Court of Justice will then have equally good constitutional grounds to respect the EU treaties and follow the old rules, ignoring the UK’s position. A new constitutional crisis inside the European Union will have been created. And in the process, the procedure of Article 48 TEU will, in effect, have stopped being the exclusive mode for amending EU treaties.
The European Council will meet on 18 and 19 February, with the United Kingdom settlement on the agenda. It is expected that a final deal will be agreed there. I conclude that the only way in which the Draft Settlement can work in practice, is if it is not taken to be new treaty of EU or public international law. If it is such a treaty it will inevitably interfere with the constitutional architecture of the EU. It will create ambiguities that may end up in a constitutional crisis, especially since it opens the way for litigation concerning all types of banking regulation, which matters hugely for the United Kingdom.
In my view, the settlement will only work if it is considered by all parties to be legally binding only as an interpretive agreement under Article 31 of the Vienna Convention, and therefore does not even begin to challenge the treaty framework. This may be politically difficult for the Prime Minister, who has promised a legally binding agreement on the ground that it is a treaty. It is clear that “legally binding as a treaty” and “legally binding as an agreement relevant to interpretation” are not the same thing. But given the alternatives, this is the only reasonable way forward for the European Union. This also follows precedent, because it is the way in which European Court of Justice has spoken of a similar agreement in the past.
Of course, it is in the nature of the European Union as a creature of international law, that the masters of the treaties can decide otherwise. They do have the constitutional power to unsettle the constitutional architecture of the Union, undermine the process of Article 48 TEU and create a hopelessly confused and unstable banking union. It would be a terrible shame, though, if they did so without realising it