The Brexit Withdrawal Agreement: Overview and First Observations

Professor Steve Peers, University of Essex

Introduction

The recently agreed Brexit withdrawal agreement may turn out to be dead on arrival, or at some point not long after. Nevertheless, if the agreement is ratified, it is the basis on which the UK would leave the EU – unless the two sides agree to amendments to the text.

Since the agreement is both complex and legalistic, but also the subject of intense political debate, there’s a need for a summary and explanation of the text for non-lawyers. This blog post aims to do that by outlining the structure of the agreement and the main content of each part of it in turn. It does not aim to be exhaustive, but only to give a broad indication of what the agreement entails.

Throughout the blog post, I’ve scattered the answers to ‘key questions’ which have been raised about what the withdrawal agreement does. Let’s start with this one:

Key question: Does the withdrawal agreement cover the long-term relationship between the UK and EU after Brexit?

No. The withdrawal agreement governs only the details of leaving the EU, not the long-term relationship between the UK and EU. On that issue, alongside the text of the agreement, there was also initially a draft outline of the political declaration on the future UK/EU relationship. This is the basis for negotiations that would take place after Brexit day. This declaration has since been finalised. Finally, there’s also a joint report on the state of negotiations to date.

This is (hopefully) the first in a series of several blog posts on the withdrawal agreement. I aim to come back to three specific issues in more detail: the ‘backstop’ relating to the Irish border; the role of the CJEU; and the future relationship. (Update, Dec 8: the blog post on the future relationship is here).

Also, for more detailed discussion on parts of the withdrawal agreement, I’ll refer to some blog posts on the March draft of the agreement – although, as I point out below, that draft has been amended on a number of issues.

Key question: What if the withdrawal agreement is not ratified?

Unless the two sides agree to amend it and then ratify the amended text, in principle the alternatives are the UK leaving the EU without any withdrawal agreement, or staying in the European Union. However, it remains to be seen if remaining by revoking the Article 50 notice of withdrawing is even legally possible: the Wightman case, now pending before the CJEU, will clarify this issue. A general election might be held, but that will not in itself change the options available. The Brexit date could be delayed, but both the UK government and the EU27 Member States (acting unanimously) would have to agree to this.

The EU Commission has issued preparedness notices setting out its view on what would happen if the UK leaves the EU without a withdrawal agreement. The UK government has also issued its own no deal notices. Last week, the Commission published a communication explaining what it would do in the event of no deal. In particular, it would suggest unilateral EU measures to keep basic aviation services to the UK (less than current aviation service). It would not propose issuing extra lorry permits, so access to the EU by UK commercial drivers would be highly curtailed. There would be no ‘adequacy decision’ on flows of personal data from the UK to the EU, so digital services would be somewhat more complicated.

It also proposed to waive visa requirements for short-term visits by UK citizens to the EU (for 90/180 days). This will not address residence for longer periods, which will be subject to EU and national laws on non-EU citizens, which are more restrictive than EU free movement rules. (For more details of the law applicable to UK citizens visiting or staying in the EU after Brexit, see my discussion here).

Structure of the withdrawal agreement

Part One of the withdrawal agreement (Articles 1-8) sets out the “Common Provisions”. These deal with basic issues like definitions and territorial scope.

Part Two of the withdrawal agreement (Articles 9-39) deals with citizens’ rights, for those EU27 citizens in the UK, and UK citizens in the EU27, before the end of the transition period.

Part Three of the agreement sets out “separation provisions” (Articles 40-125), ie the exact details of phasing out the application of EU law in the UK at the end of the transition period.

Next, Part Four sets out the rules on the transition period (Articles 126-132), during which substantive EU law still applies to the UK for two to four years after Brexit day.

Part Five concerns the financial settlement (Articles 133-157).

Part Six sets out “Institutional and Final Provisions” (Articles 158-185). This includes rules on CJEU jurisdiction over EU27 citizens, dispute settlement, making further decisions to implement the agreement and rules on entry into force and dates of application.

In addition there are three Protocols to the agreement (these have the same legal force as the main treaty), on Irish border issues,  UK bases in Cyprus, and Gibraltar.

Part One: Common Provisions

The main legal issue in Part one is the legal effect of EU law for the UK. The agreement must have the same legal effect for the UK as EU law does for EU Member States – including the principles of direct effect (meaning that the agreement as such can be enforced in national courts) and implicitly supremacy (meaning that national law which breaches the agreement must be disapplied by national courts). The UK must also pass an Act of Parliament to give effect to the Agreement; this is above and beyond the planned ‘meaningful vote’ on the agreement in Parliament.

All references to EU law require it to be interpreted in accordance with the normal rules of EU law. As for case law of the CJEU, references to EU law must be ‘interpreted in conformity with’ CJEU case law delivered before the end of the transition period. However, the rule is different for CJEU case law after the end of the transition period: in that case, the UK courts and authorities are only required to have ‘due regard’ for the case law. Note that none of this gives jurisdiction to the CJEU as such as regards the UK (although other parts of the Treaty give the Court some jurisdiction).

References to EU law include that law as amended until the end of the transition period, while references to Member States include the UK except when they refer to having voting or representation rights on EU bodies etc.

Part Two: Citizens’ rights

This part will mostly not apply until after the end of the transition period, since free movement of people will continue during that period. In principle, it provides that EU27 citizens in the UK before the end of that period (and UK citizens who are in the EU27 before the end of that period) will retain the same rights as those who arrived before Brexit day. To that end, it requires the two sides to keep applying EU free movement legislation to the people concerned, including legislation on social security coordination and the recognition of qualifications.

Some aspects of their legal status will change, however: the UK or EU27 Member States may require them to apply to prove their right to stay on the territory. The UK in particular plans to implement this, by means of a ‘settled status’ scheme. The risk is that some people will not have the documentation to prove their right to stay. Some categories of people currently covered by EU law (such as UK citizens returning to the UK with non-EU family members, or UK children in the sole care of one non-EU parent) will not be covered by the withdrawal agreement, so their position will be up to UK law. Similarly a number of EU27 citizens who would not qualify for status because they do not meet the ‘comprehensive sickness insurance’ requirement will be covered by a unilateral waiver of this rule by the UK, not the withdrawal agreement.

Also, the rules on family reunion in EU free movement law (which are more favourable than those under national law or EU law on non-EU families) will only apply where the family relationship existed before the end of the transition period, or the family member was legally resident in the same State then. If the citizens commit a criminal offence after the end of the transition period, national rules on expulsions will apply – and they may be more stringent than EU free movement rules on this issue.

The CJEU will have jurisdiction to rule on how the rules apply to EU27 citizens in the UK, on the basis of requests from UK courts, for eight years after the transitional period ends. There will also be an independent monitoring body in the UK with power to bring court cases on their behalf.

For further details, see my annotation of a previous version of the citizens’ rights Part of the treaty, which is only a little different from the final version.

Key question: Does the withdrawal agreement end free movement of people?

Yes, free movement ends at the end of the transition period, unless the UK and EU decide to sign a separate treaty as part of the future relationship extending free movement in the future. Currently the UK government opposes this idea. The ‘backstop’ relating to Northern Ireland, if it ends up applying, does not include free movement of people, but only the continuation of the UK/Ireland common travel area, which is more limited.

The withdrawal agreement also ends free movement for UK citizens already in the EU27, unless (again) a separate treaty as part of the future relationship addresses this issue.

Part Three: Separation provisions

This part tells out exactly how EU law ceases to apply at the end of the transition period, for a list of different issues. It was mostly agreed in March, with a few more Articles agreed in June, and the rest agreed since. The biggest difficulties were over geographical indications (ie, protection for those who call their product ‘Parma ham’) and what happens to cases pending before the CJEU on Brexit day.

Key question: Is the UK bound to EU law on geographical indications forever?

No, but it would be guaranteeing the continued protection of geographical indications for products which were protected at the end of the transition period.

More precisely, this part sets out rules for ending the application of EU law as regards thirteen issues: goods placed on the market; ongoing customs procedures; ongoing VAT and excise procedures; intellectual property protection; police and criminal law cooperation; cross-border civil litigation; personal data; public procurement; Euratom; judicial and administrative procedures; administrative cooperation; privileges and immunities; and other issues, such as the European Schools.

Key question: Does the CJEU have jurisdiction in the UK indefinitely?

No. Some have misread Article 89 of the withdrawal agreement, which refers to CJEU jurisdiction as regards cases coming from the UK, to mean that the UK will be covered by the Court’s jurisdiction forever. In fact, Article 89 simply confirms that the Court will have jurisdiction for cases sent from UK courts before the end of the transition period.

After that point, UK courts can send the CJEU cases only in limited contexts: for eight years after the transition period ends as regards EU27 citizens in the UK, or the protocols on Northern Ireland (in part) and Cyprus. Also, the Court has jurisdiction after this point to decide on how to interpret EU law if a dispute about the withdrawal agreement goes to arbitration (see Part Six).

Furthermore, the Commission can sue the UK in the CJEU for failure to implement EU law correctly for four years after the end of the transition period (though this can only concern alleged breaches which occurred before the end of the transition period as an EU Member State, or under the transition period Part of the agreement). Finally, the Commission can sue the UK to enforce State aid and competition decisions which were based on proceedings which started before the end of the transition period, but concluded afterward. 

It is possible that the future relationship will provide for continued EU/UK cooperation on some of these issues, in which case some of these separation clauses either will not come into practical effect at all, or will only apply in part. For instance, the EU and UK might end up agreeing in a new security treaty on another form of fast-track extradition instead of the European Arrest Warrant (EAW), and may therefore choose to agree different rules on transition from the EAW system to that new fast-track system.

Part Four: transition period 

This is a short part of the withdrawal agreement, yet it has the biggest effect: it keeps substantive EU law in place in the UK until the end of 2020.  It was agreed by March, except the negotiators have now added a new Article allowing for a possible extension.

The key elements of the transition period are that EU law (including new EU law) applies to the UK, except in areas covered by UK opt-outs (such as the single currency and justice and home affairs law; in the latter case, the UK retains part of its power to opt-in to new proposals on a case-by-case basis). There are special rules on external relations: for instance, the EU will notify non-EU countries that the UK should still be regarded as covered by EU free trade agreements, or other types of EU treaties between the EU and non-EU countries (the non-EU countries aren’t obliged to agree to this). The current rules on the allocation of fisheries catches can’t be changed to benefit either the EU or UK fishing fleets.

However, the UK will not be represented on any EU institutions or bodies – including on the CJEU, which will continue to have its usual jurisdiction regarding the UK during the transition period. The UK will only be consulted on new EU measures as a special exception. In one area – foreign and defence policy – the UK can refuse to apply new EU measures if it has fundamental objections to them, and the withdrawal agreement foresees an early treaty between the EU and UK that will replace the transition period rules.

For a detailed annotation of the transition period clauses – except for the extension clause which has now been added – see my earlier blog post.

Key question: Which EU laws does the transition period cover?

I’m often asked if the transition period covers a particular EU law like driving licences or aspects of the free movement of people, because readers can’t find a specific reference to that law in the withdrawal agreement. The simple answer is that the transition period covers all laws applying to the UK except a handful of exclusions, so the drafters didn’t bother with a list of those EU laws which are covered by it.

The EU laws which are not covered by the transition period are those about the European citizens’ initiative (on that process, see the case law discussed here) and also voting and standing for office in the European Parliament and local government. The UK (or its devolved legislatures) can, if it chooses, unilaterally continue to let EU27 citizens vote in local elections, and EU27 countries can choose to do the same for UK citizens. Also, as noted above, EU foreign and defence policy might not fully apply to the UK for the whole of the transition period.

This brings us to the new clause on extension of the transition period. It’s possible for the Joint Committee set up by the withdrawal agreement – which works by the mutual consent of the EU and UK – to decide by July 2020 if the transition period will be extended, for a period of up to a date to be decided (reportedly that date could be the end of 2022).

In that case, the transition period rules continue to apply to the UK for that extended period, except there would have to be an ad hoc negotiation on how much the UK pays into the EU budget during the extended period.  There are also special rules on agricultural support.

Key question: Can the UK be forced to stay in the transition period indefinitely?

No – and no. First of all, any extension of the transition period has to be agreed jointly, as noted already. Secondly, any extension won’t be indefinite, since the negotiators will add a final possible date for extension when they agree the final text of the withdrawal agreement.

On the other hand, the UK might theoretically end up in the backstop relating to Northern Ireland indefinitely. Although the withdrawal agreement says that this arrangement must be temporary, unlike the transition period there is no final date to end it and the UK cannot unilaterally end it at a certain date. However, the backstop is more limited in scope than the transition period, as it concerns only some EU laws (mainly on external trade, customs, and goods regulation, plus some limited application of EU laws on labour, the environment, state aids and competition in its Annexes). In particular, the backstop does not concern the free movement of people or services, or contribution to the EU budget.

Part Five: financial settlement 

This part incorporates the earlier agreement that the UK takes part in the EU’s spending until the end of the current budget cycle (end 2020), which matches the end of the transition period (unless that period is extended). As noted above, though, if the transition period is extended, the UK and EU will negotiate a separate EU contribution to the EU budget. It also includes UK payments to the budget incurred because the EU often makes financial commitments in one year and then pays them out in later years (the system known as reste à liquider). Furthermore, it includes continued payments to ‘off-budget’ EU spending such as commitments to developing countries, until the current versions of those programmes expire. For a detailed discussion of the finances of the deal, see reports from the OBR, the National Audit Office and the Institute for Government.

Key question: Has the UK agreed to pay £39 billion in return for nothing?

No. First of all, this begs the question: the EU regards the financial settlement as money which is payable in any event regardless of any future relationship, and took the view that the future relationship could not be formally negotiated before Brexit day anyway. Even on its own terms, the argument that ‘£39 billion is paid for nothing’ is wrong, since about half of of the money relates to the UK still applying substantive EU law during the transition period (see the OBR report), during which the UK still has its current access to EU goods and services markets (and vice versa). Anyway, a free trade deal could not have been negotiated in the time available, even if that had been legally possible (again: the EU says it was not).

Part Six: Institutional and Final Provisions

First of all, as noted above (in Part Two), the CJEU will have jurisdiction to rule on how the rules on citizens’ acquired rights apply to EU27 citizens in the UK, on the basis of requests from UK courts, for eight years after the transitional period ends. There will also be an independent monitoring body in the UK with power to bring court cases on their behalf. The two sides might agree to wind up the monitoring body at the end of the same time period.

Secondly, the CJEU will have jurisdiction, after the end of the transitional period, over the references to EU law in the financial settlement part of the agreement, in references from national courts or as regards Commission infringement actions brought against the UK.

Next, there’s a Joint Committee set up to oversee and implement the agreement. It will meet at least once a year, and there are a number of sub-committees dealing with specific issues like Northern Ireland and citizens’ rights. It can take certain decisions to add to the agreement – such as an extension of the transition period – but for all these decisions the EU and UK must both agree.

There are detailed rules on dispute settlement, providing for arguments about the agreement to go to a panel of arbitrators. However, if the arbitrators have to decide an issue of EU law when settling the dispute, they will have to ask the CJEU to give a ruling. This is unavoidable, since CJEU case law insists that the EU and its Member States cannot be bound by an interpretation of EU law other than the CJEU’s (see, for instance, CJEU Opinion 1/92).

The Northern Ireland backstop rules out the arbitration process (including the CJEU) from much of the ‘level playing field’ rules on labour law, environmental law, tax, and competition and state aids law (see below). However, the CJEU has jurisdiction over part of the backstop, as well as the Protocol on bases in Cyprus.

Key question: Does the CJEU have jurisdiction over the entire agreement?

No. The Court has its usual jurisdiction under the transition period, and following that specific jurisdiction over EU27 citizens’ rights and EU law referred to in the financial settlement, as well as the protocols on Northern Ireland (in part) and bases in Cyprus. But the arbitrators can only refer disputes over EU law to it, not disputes over the entire agreement – although a lot of the agreement does refer to EU law. Also, some key points in the backstop are outside the arbitrators’ or the CJEU’s jurisdiction.

Note that this part of the agreement differs a lot from the dispute settlement clauses as the Commission proposed them in March (see discussion here) – which included powers (tilted toward the EU) for both sides to impose sanctions on each other. The final version looks a lot more like traditional international dispute settlement, and is presumably a concession by the EU to the UK.

Finally, the last provisions of the main withdrawal agreement set out ‘boilerplate’ rules: confirming that the three Protocols and nine Annexes are binding; setting out the authentic languages of the text and the depositary; and setting the date of entry in force (30 March 2019). The withdrawal agreement applies from that date, except the parts on citizens’ rights, separation provisions, dispute settlement and the three Protocols mostly apply from the end of the transition period (with specified exceptions which apply immediately). There’s also a commitment to negotiate on the future relationship, referring to the separate joint declaration on that issue, ‘with a view to’ agreeing those texts by the end of the transition period ‘to the extent possible’.

Protocol on Irish border issues

The Northern Ireland Protocol starts with three new Articles compared to the March draft, on UK territorial integrity and emphasising that the Protocol is meant to be temporary. Also there’s a link back to the possible extension of the transition period mentioned above (Part Four). A new review clause was added, but both sides would have to agree to disapply the protocol partly or wholly.

As for the substance of the Protocol, it refers to equality rights and the common travel area between the UK and Ireland, as already agreed (these issues were never controversial).

Key question: Does the common travel area continue the free movement of people between the UK and EU?

No. The absence of border checks makes it impossible to refuse entry to people at the border with Ireland, but that does not mean any EU27 citizens crossing that border (besides Irish citizens) have the right to reside or work in the UK (and vice versa for people crossing into Ireland, besides UK and EU citizens). There is UK legislation on liability for employers, landlords et al who hire or do business with people who are not authorised to work or reside that will be relevant here.

Then there’s the UK-wide customs union backstop, part of which is specific to Northern Ireland. There are a number of elements of this. This part of the Protocol links to Annexes on: a) trade in goods between EU/UK/non-EU states; b) customs cooperation; and c) a ‘level playing field’, which means some degree of continued harmonisation of law relating to tax, the environment, labour law, state aid, competition, and public companies/monopolies.

However, this falls short of the obligations of EU Member States; there are limited obligations to keep up with new EU legislation and CJEU case law; and as noted above (Part Six), the arbitration rules (including CJEU jurisdiction) mostly do not apply to this ‘level playing field’. There’s a lot of EU law that wouldn’t apply to the UK – most notably the free movement of persons, services and capital, and contributions to the EU budget. So while the backstop would still commit the UK to a chunk of EU law on trade in goods, and in a limited way to some law in the ‘level playing field’ areas, the continued application of EU law would be much less than under the rules on the transition period.

The Protocol also contains provisions on the UK internal market, as well lists of specific EU laws that apply in Northern Ireland: product regulation, VAT and excise tax, agriculture and the environment, a single electricity market,  and state aids. There’s also a vague reference to other North/South cooperation. Finally, the institutional provisions of the Protocol provide that EU bodies, including the CJEU, have competence to apply or interpret the provisions of the Protocol that are specific to Northern Ireland.

Protocol on UK bases in Cyprus

This Protocol confirms that the bases in Cyprus remain within EU customs territory after Brexit, and EU regulations on goods, including agricultural and fisheries laws, still apply. EU law on excise taxes and VAT also continues to apply. Goods supplied to the staff on the bases are exempt from customs and taxes, and the UK and Cyprus may agree further rules on social security coordination. There are rules on checks at the border of the bases area, and a general obligation to cooperate to prevent fraud. Finally, the EU institutions, including the CJEU, have competence to apply and interpret EU law referred to in the Protocol.

Protocol on Gibraltar

First of all, this Protocol provides for the UK and Spain to cooperate on workers’ rights as regards the Spain/Gibraltar crossing. Next, it retains the status quo on access to aviation, unless the Joint Committee decides differently. It also contains general provisions on cooperation on tax and fraud, environmental protection and fishing, and police cooperation.

Assessment

The desirability of the withdrawal agreement can only be assessed in light of the alternative options available. Whether it’s possible to end the Brexit process at all by revoking the notification under Article 50 remains to be clarified by the CJEU, as noted above. If that’s not possible, Brexit could be averted only by extending the period of negotiations under Article 50 in order to amend the Treaties to permit revocation of the withdrawal notice, but that is a stretch.

Amendment of the text of the withdrawal agreement is technically possible, but the EU has signalled that it would not be willing to consider any significant amendment to the text. Holding a general election obviously appeals to anyone who wants a change of government in any event, but does not change the fundamental issues relating to the withdrawal agreement or the Brexit process as such.

Objecting to the withdrawal agreement because of objections to the separate future relationship declaration makes little sense. The EU always took the view that it would not and could not negotiate the future relationship formally before Brexit day, and there would anyway not have been enough time to negotiate a full future relationship in two years (unless the UK decided to stay in the EEA as an off-the-shelf model, which would still have required some separate issues to be negotiated). Furthermore, since the future relationship declaration is not binding, the nature of that future relationship is still all to play for.

The options as regards the withdrawal agreement are therefore to support it, to overturn the Brexit process (if that turns out to be legally possible), or to leave the EU without a withdrawal agreement, and therefore without a trade agreement. Each of these three options may individually lack a majority in Parliament, but the no deal option – although it probably commands the least support – is the default if one of the two other options does not command a majority. A no deal outcome – damaging UK exports to their largest market, leaving the position of UK citizens in the EU27 and EU citizens in the UK less secure, disrupting the UK security relationship with the EU, significantly limiting flights and commercial lorry transport with the EU, and raising barriers to transfers of data from the EU to the UK – is manifestly not in the UK’s interest, and no responsible politician should support it. The choice is therefore between the withdrawal agreement or reversing the Brexit process (if possible) as a means to avoid this end.

For supporters of Brexit, the withdrawal agreement may be a disappointment, applying substantive EU law to the UK for the transition period and potentially a proportion of EU law to the UK via the backstop after that. From this perspective, it’s unfortunate that the possibility of staying in the EEA – giving the UK a judge on the EFTA Court and consultation rights on EU proposals, limiting the effect of EU law in the UK, and possibly reducing the scope of EU law applicable to the UK at an earlier point – wasn’t explored from the outset as at least a partial substitute for the transition period.

However, given that new Member States have a transition period of several years when they join the EU (as the UK did when it joined the previous EEC) and usually gear up alignment with EU law for several years before joining, it was unrealistic to imagine ending the UK’s application of substantive EU law would happen all in one go. The withdrawal agreement recognises the unavoidable complexity of the process, phasing out membership via different stages while still offering options for a future relationship – subject to the need to avoid checks on the Irish border.

Whichever of these two routes is ultimately chosen – managed withdrawal or reconsidered reversal – it’s important to find a way of avoiding the worst-case scenario of a no deal Brexit: the project of greedy capitalists, seedy jingoists, tweedy fantasists and needy narcissists.

Barnard & Peers: chapter 27

Photo credit: BBC

Preliminary references and investment tribunals: is the Luxembourg Court extending a helping hand?

Hannes Lenk, PhD Candidate at the University of Gothenburg

The relationship of arbitral tribunals with the Court of Justice of the European Union (CJEU) has been the subject of a long-lasting juridical struggle. The current position is as simple and pragmatic as it is controversial. Commercial arbitration tribunals are not considered to be a ‘court and tribunal of a Member State’ within the meaning of the Article 267 TFEU and, thus, unable to refer questions to the CJEU on matters of interpretation of EU law.  At the same time, it is an open secret that questions of EU law do arise during arbitration proceedings, and there is an inherent risk that tribunals get it wrong—at least sometimes. In commercial arbitration these shortcomings might be addressed through the indirect involvement of domestic courts and the CJEU at the recognition and enforcement stage of arbitral awards. A similar possibility might not exist in investment arbitration and for some time now the question of whether or not investment tribunals are entitled to request preliminary references from the CJEU has been simmering under the surface of a deeply politicised debate on investor-state dispute settlement provisions in currently ongoing negotiations for deep and comprehensive trade and investment agreements with, inter alia, Canada and the US. A recent opinion of Advocate General Wathelet might break new ground in this debate and prepare the field for future judicial dialogue.

Commercial arbitration: from Nordsee to Eco Swiss

'Article 267 TFEU is an important instrument for cohesion and coherence in the judicial system of the European Union (EU), including domestic courts as ‘ordinary courts of the EU legal order’ (Opinion 1/09, para. 80). By way of establishing a judicial dialogue, the preliminary reference mechanism guarantees that individuals have their rights under EU law enforced in domestic courts, and assures a uniform interpretation and application of EU law in all Member States. Notably, the decision to request a preliminary reference is generally within the discretion of domestic courts, which are obligated to refer questions only in limited circumstances, i.e. in instances where the case is pending before a domestic court of last instance.

Article 267

1. The Court of Justice of the European Union shall have jurisdiction to give preliminary rulings concerning:

(a) the interpretation of the Treaties;

(b) the validity and interpretation of acts of the institutions, bodies, offices or agencies of the Union;

2. Where such a question is raised before any court or tribunal of a Member State, that court or tribunal may, if it considers that a decision on the question is necessary to enable it to give judgment, request the Court to give a ruling thereon.

3. Where any such question is raised in a case pending before a court or tribunal of a Member State against whose decisions there is no judicial remedy under national law, that court or tribunal shall bring the matter before the Court.

4. If such a question is raised in a case pending before a court or tribunal of a Member State with regard to a person in custody, the Court of Justice of the European Union shall act with the minimum of delay.'

However, Article 267 TFEU includes a significant procedural limitation. In order to to request a reference from the CJEU the judicial body must be covered by the concept of ‘any court or tribunal of a Member State’. The CJEU has historically interpreted this concept restrictively. In Dorsch Consult the CJEU clarified the characteristics that need to be taken into account. Accordingly, a ‘court of tribunal’ is any judicial body that exercises judicial functions, i.e. that is (a) established by law, (b) a permanent institution, (c) with compulsory jurisdiction, (d) whose procedure is inter partes, (e) applying rules of law, and (f) acts independent of other branches of government. Applying these criteria to a commercial arbitration tribunal, the CJEU subsequently declared in Nordsee that despite “certain similarities between the activities of the arbitration tribunal … and those of an ordinary court”, the tribunal in question was not a ‘court or tribunal’ within the meaning of the preliminary reference procedure.

Arbitral tribunals are, therefore, left without guidance on the interpretation of EU law where this becomes relevant during the arbitration proceedings. From an EU law perspective, the adverse effect of incorrect interpretation and application of EU law in commercial arbitration is mitigated by the indirect involvement of domestic courts, and by association the CJEU. Indeed, domestic courts play an important role in supporting the arbitral tribunal upon request, as well as in the recognition and enforcement of arbitral awards. In Eco Swiss the CJEU emphasized that domestic courts are generally required to assess the compatibility of arbitral awards with EU public policy and may request a preliminary reference from the CJEU to that end. The award in Eco Swiss was considered a violation of EU competition rules (now Article 101 TFEU), which, according to the CJEU, constitutes a ‘fundamental provision which is essential for the accomplishment of tasks entrusted to the [Union]’ (para. 36). The CJEU furthermore clarified that it is to be considered part of public policy in the meaning of Article V(1)(c) and (e), and II(b) of the New York Convention on the Recognition and Enforcement of Foreign Arbitral Awards.  Consequently, arbitral awards that are irreconcilable with EU public policy are unenforceable within the territory of the Member States under domestic and international law.

Investment arbitration: commercial arbitration in disguise or something else entirely?

This approach is open for much criticism, not least because the concept of EU public policy remains notoriously undefined. “In light of a constant referral of additional tasks upon the European Union over the last twenty years, it has become increasingly difficult to identify those provisions which may be regarded as fundamental for those tasks.” (Basedow, p. 373). Nonetheless, from an EU law perspective it appears to provide a pragmatic procedural solution that reserves the involvement of the CJEU and assures the application of EU law in domestic courts. Particularly in the area of investment arbitration it has been suggested that invoking EU public policy at the enforcement stage might prevent some of the most controversial awards from gaining legal effect within the EU legal order. But investment arbitration is in many ways different from commercial arbitration. The vast majority of proceedings is governed by the rules of the International Centre for Settlement of Investment Disputes (ICSID). Article 54 ICSID provides for the automatic recognition and enforcement of awards, excluding domestic courts from any involvement in the review of ICSID awards vis-à-vis public policy. To make matters worse, non-ICSID awards are not seldom enforced outside the territory of the respondent state. Enforcement of the controversial Micula award, for instance, is currently sought in the US.

Gaffney and Basedow have recently advocated the view that investment tribunals should be able to request references under Article 267 TFEU. It is in this context noteworthy that the CJEU in Nordsee conceded to the possibility that an arbitral tribunal might fall within the scope of Article 267 TFEU, provided that the tribunal derives its jurisdiction not exclusively from party autonomy of the disputing parties, but instead involves the exercise of state authority to the extent that it can be considered an institution of the state. This view was later confirmed in Ascendi, a request from the Tribunal Arbitral Tributário in Portugal. The CJEU observed that Portuguese law provides for the resolution of tax disputes through arbitration, which also regulates the functioning and constitution of the tribunal. “[The Tribunal’s] jurisdiction stems directly from the provisions of Decree-Law No 10/2011 and is not, as a result, subject to the prior expression of the parties’ will to submit their dispute to arbitration”, the CJEU concluded (para. 29). Not unlike the Tribunal Arbitral Tributário, investment tribunals are an alternative dispute settlement system provided for in law, i.e. the the underlying investment agreement, which constitute a “non-transient element of [the domestic] judicial system” (Basedow, p. 379-380).

The idea to construe investment tribunals as ‘court or tribunal’ for the purpose of Article 267 TFEU is not merely a scholarly endeavor to square the circle. There are signs from within the CJEU that this might present an acceptable solution to the problem of integrating investment tribunals in the EU legal order. In his recent Opinion in Genentech, a preliminary ruling from the Cour d’appel de Paris concerning the notion of EU public policy in the recognition and enforcement of arbitral awards, Advocate General Wathelet presented his well-balanced and carefully drafted view on the relationship of arbitral tribunals with the CJEU. On the outset the AG simply confirms well established case law along the lines of Nordsee and Eco Swiss.

'Referring to the system for reviewing the compatibility of international arbitral awards with EU law through the public policy reservation […] the Court has held that arbitral tribunals ‘constituted pursuant to an agreement’ are not courts of the Member States within the meaning of Article 267 TFEU. Consequently, they cannot refer questions for a preliminary ruling. It is therefore for the courts of the Member States, within the meaning of Article 267 TFEU, to examine, if necessary by referring a question for a preliminary ruling, the compatibility of (international or domestic) arbitral awards with EU law where an action is brought before them for annulment or enforcement, or where any other form of action or review is sought under the relevant national legislation.'

Much more powerful considerations are hidden in the footnotes. Here the AG addresses the situation of investment tribunals explicitly to which, in his view, different considerations should apply.

Footnote 34

'Based on this case-law, the arbitral tribunals hearing cases within the framework of the Convention on the Settlement of Investment Disputes Between States and Nationals of Other States (ICSID) could be regarded as being able to refer questions to the Court for a preliminary ruling. See, to that effect, [Basedow], p. 376 to 381. Since the number and size of investment arbitrations raising questions on the application of EU law are increasing, particularly in the field of State aid, the possibility for arbitral tribunals to refer questions for a preliminary ruling could help to ensure the correct and effective implementation of EU law.'

Indeed, these remarks are merely obiter dicta and the CJEU is unlikely to address any of this in the final judgment, other than—probably—following the AG on substance. The opinion, nonetheless, sends a strong signal to investment tribunals, i.e. an explicit invitation to request preliminary references from the CJEU on matters of interpretation of EU law.

Helping hand or last straw

In the best of cases AG Wathelet’s opinion would be construed as a helping hand from Luxembourg to find an amicable solution to the current conflict, based on judicial dialogue and mutual comity. However, investment tribunals have thus far refused to engage with the CJEU on questions of EU law. Or put differently, it does not appear as if investment tribunals consider EU law to be of actual relevance to arbitration. In Oostergetel and Laurentius, for instance, the tribunal acknowledged that there is “absence of any conclusive position of the [CJEU]” on the relevant issues of EU law, but subsequently rejected the respondent’s request to refer a question to the CJEU with the help of a domestic court (para. 109). The investment tribunal in Micula rejected concerns raised by the Commission to the effect that the award, if rendered, were unenforceable under EU state aid law; plainly ignoring the resulting conflict.

Gaffney suggested that a lack of guidance on questions of EU law would prompt a domestic court’s responsibility under Article 267 TFEU. However, even domestic courts might be cautious of involving requests for preliminary references in investment arbitration cases. When the award on jurisdiction in Achmea was challenged in May 2012 the Higher Regional Court of Frankfurt decided that, while EU law was raised during the arbitration, the dispute concerned in fact the interpretation of the arbitration clause in the investment agreement and as such fell outside the scope of interpretation of EU law. Ultimately, the final award was challenged before the same court in December 2014. The Frankfurt court recognized that the compatibility of arbitration clauses in intra-EU investment agreements with the Treaties is much debated but refused to refer the question to the CJEU. These cases reflect anything but excitement about the involvement of the CJEU in the arbitration process. Rather than jumping on the invitation from Luxembourg to refer questions, AG Wathelet’s opinion runs the risk of being perceived as a last straw for investment tribunals that are ultimately expected to accept the dominance of EU law and the jurisdiction of the CJEU.

Remaining challenges

However, even if investment tribunals refer questions to the CJEU in the future, a few questions still remain. First, courts or tribunals against whose decision there is no judicial remedy are not only entitled, but, in accordance with Article 267(3) TFEU, obligated to refer questions on the interpretation and the legality of EU law. Although domestic arbitration laws may provide for investment awards to be set aside, it does not prevent the award from being enforced under Article (1)(e) of the New York Convention in another state. Article 52 ICSID provides for an internal procedure for the annulment of ICSID awards on limited grounds, which effectively excludes the involvement of domestic courts. Considering, therefore, that an investment award cannot be appealed or permanently set aside on the basis of wrongful interpretation of EU law, investment tribunals might fit squarely into Article 267(3) TFEU.

The investment court, which was recently incorporated in the Comprehensive Economic and Trade Agreement with Canada (CETA) and the EU-Vietnam FTA, and which is proposed in Transatlantic Trade and Investment Partnership with the US (TTIP), raises similar concerns. Decisions of the Tribunal may be appealed before the Appeals Tribunal, inter alia, on grounds of the wrongful appreciation of domestic law (as a matter of fact). Albeit that the first instance Tribunal is relieved from any obligation under Article 267(3) TFEU, it ultimately shifts this burden onto the Appeals Tribunal. The more fundamental problem in this regard is that the EU Treaties cannot actually obligate investment tribunals to refer questions to the CJEU.

Secondly, decisions of the CJEU under the preliminary reference procedure are binding on the referring court. Without explicit safeguards in the investment agreement, however, investment tribunals are under no obligation to follow the interpretation of the CJEU (Gaffney, p. 13). There is no obvious reason why investment tribunals would refer a question to the CJEU just to subsequently ignore the answer provided. Be that as it may, these two above reservations are likely to affect the essential characteristics of Article 267 TFEU, and the powers conferred thereunder on the CJEU. According to well-established case law of the CJEU, this would adversely affect the autonomy of the EU legal order and consequently violate the Treaty (Opinion 1/09, para. 77-79). An interpretation of Article 267 TFEU that invites arbitral tribunals to refer questions but neither obligates them to do so under Article 267(3) TFEU nor renders answers of the CJEU binding on the referring investment tribunal would, thus, be incompatible with the Treaties.

Third, and perhaps most problematic, are denial of justice cases where the interpretation of domestic law might itself be the reason for an investment dispute. It would be bizarre scenario, indeed, for investment tribunals to request a preliminary ruling from the CJEU on a domestic court’s interpretation of EU law, particularly if the CJEU was involved during the domestic proceedings. Under the EU-Vietnam FTA and CETA, such a scenario could be captured by manifest arbitrariness (e.g. Article 8.10(2)(c) CETA). Gaffney points out a few other challenges such as the steadily growing influx of preliminary references that is already creating a backlog of cases, and which is likely to extend the arbitration process for several month, if not years (p. 14).

Conclusions

Although we are unlikely to see changes in the approach of the CJEU to commercial arbitration anytime soon, AG Wathelet’s opinion amounts to a strong endorsement of the view that investment tribunals are an entirely different story. Whether or not the preliminary reference procedure paves the way for much needed judicial comity between Luxembourg and investor-state tribunals is, however, still very much an open question. 

Barnard & Peers: chapter 10

Photo: ICSID headquarters, Washington DC

Photo credit: icsid.worldbank.org