This article was first published on Global Arbitration Review on 1 April 2020.

Simon Rainey QC and Gaurav Sharma of Quadrant Chambers consider the legal and practical implications of the recent Hague Court of Appeal decision reinstating the US$50 billion Yukos awards against Russia.

The Court of Appeal of The Hague gave its much-anticipated judgment in the Yukos set aside proceedings on 18 February 2020. It reinstated UNCITRAL awards worth a combined US$50 billion in favour of the oil company’s former majority shareholders against Russia – overturning the decision of the Hague District Court which had set the awards aside on the ground that there was no valid arbitration agreement. The decision is an important one for all arbitration practitioners, not only for its approach to the investment treaty arguments addressed in the context of establishing the validity of the arbitration agreement but also for some general points of arbitral practice.

In this article, we begin by considering the de novo nature of the Court of Appeal’s review and its divergent approach compared with the UNCITRAL tribunal and the District Court on the key jurisdictional question before it on appeal, namely the provisional application of the Energy Charter Treaty prior to ratification. We then analyse the Court of Appeal’s treatment of Russia’s complaints relating to the role of the tribunal secretary and the damages assessment adopted by the tribunal in the absence of a model presented by the respondent state. Finally, we contemplate the reverberations the Court of Appeal’s decision may have as a matter of international law and relations.

Rehearing versus review

The standard of the Court of Appeal’s review of jurisdiction, as in England and in many other jurisdictions, was that of a rehearing, de novo, of the case for and against arbitral jurisdiction. While the Court of Appeal did not need to address the question of which party bears the burden of proof of the existence or absence of a valid arbitration agreement, it made it clear that its function (like that of the District Court below it) was not to review the correctness of the arbitral decision on jurisdiction, but to determine the issue of construction afresh on the basis of its own review of the position. The rationale was explained as follows:

The fundamental nature of the right to access to the courts entails that answering the question of whether a valid arbitration agreement was concluded is ultimately up to the court and that the court will not exercise restraint when assessing a claim seeking the setting aside of an arbitral award on the grounds that a valid arbitration agreement is lacking.

That accords perfectly with the English courts’ approach to the reassessment of arbitral jurisdiction under section 67 of the English Arbitration Act 1996: see Dallah v Pakistan and many first-instance decisions (although the statements in Dallah were obiter and occasionally applicants reserve the right to contend that that approach is wrong: see, for example, Kyrgyz Republic v Stans Energy).

Two consequences follow.

First, there is (and logically can be) no deference to the arbitral tribunal’s approach, whether on the facts or the law. De novo means what it says: a complete rehearing, necessarily based solely on the court’s own independent assessment. The Hague Court of Appeal accordingly conducted its own exhaustive review of Russian law, the jurisprudence of the Russian courts, the domestic and international practice of Russia, and Russian academic literature. While some commentators debate the merit of completely disregarding the work of the tribunal in this way (for example, Merkin & Flannery, who espouse a more review-based approach), there is no logical basis for deferring to a tribunal who may have no jurisdiction at all. The hard-line English approach in Dallah is echoed strikingly in the Hague Court of Appeal’s approach.

Second, if the hearing is de novo, then the court cannot be fettered by how the case was put before the arbitrators, since it is concerned with its own independent enquiry. This was strikingly illustrated in the case of Yukos where the Hague Court of Appeal upheld arbitral jurisdiction upon a new ground for jurisdiction which had not been argued by the shareholders either in the arbitration or before the District Court. The Court of Appeal dismissed Russia’s arguments that this was impermissible, on the basis that it would be wrong for the court to find there was no jurisdiction where there was, in fact, a valid arbitration agreement, and the actual existence of such an agreement could not depend on incorrect reasoning. The court said it would be “contrary to effective administration of justice in arbitration if an arbitral award would have to be set aside because the arbitral tribunal used a wrong argument for assuming jurisdiction when in fact such jurisdiction does exist”. The Court of Appeal drew no distinction between a new case on the law, building on a previous legal argument, and a new case with new evidence being pleaded for the first time before the court (although it regarded the Yukos shareholders’ new point as one of law in any event). The English approach, at least so far, has been similar and to regard a new evidential case in the same way as a new legal case: see the review of the authorities by Males J in Central Trading & Exports v Fioralba Shipping.

A curiosity of the Hague Court of Appeal’s approach, however, was to distinguish – albeit perhaps tentatively, the position of the party seeking to uphold jurisdiction where this had been determined by the arbitral tribunal from that of the party who seeks to disturb the award and challenge jurisdiction on different grounds. In the former case, “the latter pleaded argument only leads to a conclusion already drawn by the arbitral tribunal” and new cases should be permitted; whereas, in the latter case, the challenging party should be barred from bringing such a challenge because “if that argument had been argued earlier, the arbitral tribunal could have taken a decision on the matter at an early stage, which would, in as much as possible, have prevented unnecessary procedural steps from being taken”. That seems illogical if the court conducts a full rehearing and is concerned to arrive at the correct conclusion, one way or the other, without being tied to the imperfect way in which the case for (or against) jurisdiction was originally framed. The Court of Appeal opened the door by saying that it could not definitely be “ruled out in advance” that a challenger could deploy a new case before the court, but that would depend on the facts, case by case. Why the same would not apply equally to the party seeking to uphold jurisdiction was not explained and is analytically difficult to justify.

The English approach draws no such distinction: de novo applies to each party. There has been a (limited) move, however, towards case management control of what the court might regard as an abuse of process by the way in which a party has held back a point or prejudiced the other by not deploying before the tribunal the case which it now seeks to deploy; but that would, therefore, apply as much to the upholder of jurisdiction as to the challenger: see Central Trading v Fioralba and, earlier, Primetrade v Ythan.

The provisional application of the ECT

The key issue of substance in the Hague Court of Appeal’s decision concerns the provisional application of the Energy Charter Treaty (ECT), and the circumstances in which the ECT’s contracting states would be bound in the absence of ratification of the treaty in domestic law.

The issue turned on the proper construction of article 45 of the ECT, which provides in relevant part as follows:

1. Each signatory which applies the Energy Charter Treaty provisionally in accordance with Article 45(1) and each Contracting Party agrees to apply this Amendment provisionally pending its entry into force for such signatory or Contracting Party to the extent that such provisional application is not inconsistent with its constitution, laws or regulations.

2. Notwithstanding paragraph (1) any signatory may, when signing, deliver to the Depository a declaration that it is not able to accept provisional application. The obligation contained in paragraph (1) shall not apply to a signatory making such a declaration. Any such signatory may at any time withdraw that declaration by written notification to the Depository.

The parties debated the issue extensively in the arbitral proceedings, resulting in an interim award in 2009 which decided this jurisdictional issue as a preliminary point.

In short, Russia objected to the jurisdiction of the tribunal on the basis that the ECT’s article 26 referring disputes between investors and contracting states to international arbitration was incompatible with Russian law, which prohibited the submission of disputes relating to sovereign acts, including claims for damages for expropriation, to arbitration. Moreover, the state argued that article 26 was incompatible with the Russian Constitution, since this required treaties to be ratified by Parliament. This, said Russia, was an emanation of the principle of the separation of powers and thus needed to be respected by the tribunal as a matter of the rule of law, not least because Russia “does not have a dualist legal system” in which “investment treaty protections operate on an entirely separate plane from domestic law”.

The Yukos shareholders, on the other hand, argued that the provisional application of treaties was a well-established principle of international law that created legally enforceable rights and obligations. The plain language of article 45, they argued, provided for the automatic provisional application of the ECT. Russia had not made a declaration under article 45(2) to opt out of the provisional application regime, and as such the ECT applied in full force and effect so far as qualifying investments in Russia were concerned.

The tribunal found in favour of the shareholders, and emphasised that investors were not required to analyse the consistency of every provision of a treaty for consistency with Russian law, but should analyse whether the principle of the provisional application of treaties was consistent with domestic law. In the case of Russia, the tribunal was satisfied that was indeed the case, based in significant part on Russia’s own state practice, including a response to a Council of Europe study in 2001 confirming as much.

In its de novo assessment of the tribunal’s jurisdiction, the Hague District Court disagreed with the approach taken by the Tribunal, finding that the issue was not about the principle of provisional application. Instead, article 45(1) necessitated an examination of each of the provisions of a treaty separately for compatibility with domestic law. The court found that article 26 of the ECT was inconsistent with Russian law, which only permitted the referral of civil disputes to arbitration, and not public law disputes including those between an investor and Russia under a treaty.

Faced with these competing decisions, the Hague Court of Appeal adopted yet a further alternative approach to the interpretation of article 45, which was pleaded in the alternative by the shareholders. The Court of Appeal considered that the key issue was whether the provisional application of any provision of the ECT was incompatible with a rule of national law. It referred to article 23 paragraph 1 of Russia’s Federal Law on International Treaties, which provides:

An international treaty or a part thereof may, prior to its entry into force, be applied by the Russian Federation provisionally, if the treaty itself so provides or if an agreement to such effect has been reached with the parties that have signed the treaty.

The Court of Appeal concluded that this provision contained “no restrictions whatsoever with regard to the (types or categories of) treaty provisions that can be applied provisionally.” Accordingly, the court overturned the HDC’s decision on jurisdiction, and restored the Tribunal’s award.

An important consequence that follows from the Court of Appeal’s approach is that it arguably results in more, and not less, uncertainty as to the proper construction of article 45 of the ECT. That is because the Hague Court of Appeal’s decision represents a meaningful departure from the treatment of the issue by the parties before the tribunal, and by the tribunal itself.

In many ways, the Court of Appeal’s analysis echoes that of the handful of investment treaty tribunals that have considered the issue to date. For instance, the ICSID tribunal in Plama v Bulgaria in 2005 dealt with the issue briefly, finding that the ECT applied in its entirety, including article 26, from the moment it was signed, albeit provisionally and subject to a declaration of exemption from a provisional application under article 45(2).

The SCC tribunal in Petrobart v Kyrgyz Republic in 2005 had to consider the scenario of whether the UK’s ratification of the ECT, which was only expressed as extending to “the United Kingdom of Great Britain and Northern Ireland, the Bailiwick of Jersey and the Isle of Man”, with no mention of Gibraltar, caused the ECT to cease to apply provisionally in respect of investments of Gibraltar’s nationals. It determined that it did not cease to apply provisionally, and thus had jurisdiction to hear the dispute on the merits before it.

While Plama and Petrobart were concerned with the provisional application of the ECT generally, and the circumstances in which it might cease to apply provisionally, the genetic code of the HCA’s decision most closely resembles that of a third case, namely Ioannis Kardassopoulos v Georgia. There, an ICSID tribunal examined the interaction between the first and second paragraphs of article 45 and concluded that they entailed two independent and distinct limbs that were not interdependent. The tribunal found that article 45(1) dealt with inconsistencies between domestic law and provisional application of the ECT prior to its ratification, while article 45(2) dealt with a state’s decision to make a declaration that it wishes to be exempt from the provisional application of the ECT following its signature. Crucially, the point at issue in article 45(1) was whether anything in a state’s domestic law prevented the provisional application of the ECT in its entirety, rather than the provisional application of any of its articles in particular.

Given the significantly differing approaches to the interpretation of article 45 and the provisional application of the ECT in the Yukos case by the UNCITRAL tribunal in its original decision on jurisdiction, the Hague District Court and now the Hague Court of Appeal, the only certainty for the Yukos shareholders and Russia now is that there is no certainty which path the Dutch Supreme Court will take (assuming the issue is put before it in due course, as indicated by Russia).

The tribunal secretary versus the “fourth arbitrator”

A major ground of Russia’s challenge to the awards concerned allegations that the tribunal had improperly delegated some of its responsibilities to the tribunal secretary. Russia relied on a comparison of hours billed by tribunal members and the secretary and detailed semantic and linguistic analysis of the text of the award, arguing it showed that the secretary had “held the pen”. (The claim was buttressed by expert evidence: ultimately a sterile exercise in the case of a lengthy award where there are many redrafts and multiple authorship.) Russia alleged that the secretary had acted as a “fourth arbitrator” and therefore the tribunal had de facto been improperly composed.

The Court of Appeal dismissed the challenge essentially because, even had Russia made good its “textual” case, it could not be established:

• that the secretary had participated in the actual decision-making process which was the sole responsibility of the tribunal members;
• that the tribunal had delegated any portion of the decision-making to the secretary; or
• that the texts of the impugned portions of the award were written before the decision-making process followed by the tribunal and therefore might have influenced the decision-making process: they could equally well have been written after that process had been completed, and to reflect and record the decision-making process which the tribunal members alone had followed.

This corresponds with the English law approach to tribunal secretaries where the focus is on whether it can be shown that the arbitrators had in fact delegated their decision-making responsibility. In the English High Court case P v Q in 2017, the tribunal secretary had not only been used to analyse submissions and draft procedural orders but had also been asked for his views on the merits, although the president had exercised independent decision making. After noting the “considerable and understandable anxiety in the international arbitration community” on the issue, the court posed the test in this way, echoing the Hague Court of Appeal’s yardstick:

The use of a tribunal secretary must not involve any member of the tribunal abrogating or impairing his non-delegable and personal decision-making function. That function requires each member of the tribunal to bring his own personal and independent judgment to bear on the decision in question. (See also Sonatrach v Statoil)

This is of course in line with many institutional rules (see, for example, section XIX of the ICC’s 2018 Note to Parties and Arbitral Tribunals on the Conduct of the Arbitration; and section 8 of the LCIA’s 2017 Notes for Arbitrators).

The Hague Court of Appeal’s decision goes further in certain respects and shows the practical difficulties in mounting a sustainable challenge to an award on “fourth arbitrator” grounds.

First, it rejected an argument that by the tribunal, through the Permanent Court of Arbitration, refusing to give details of the secretary’s activities on the grounds of “confidentiality of the deliberations”, it was necessary to be inferred that the secretary had participated in the deliberations themselves. That makes fishing expeditions to find out material on which to base a putative challenge difficult.

Second, not only did it reject the textual case, but it dismissed Russia’s arguments based on the time spent by the secretary necessarily connoting that he must have strayed beyond the sorts of permissible “secretary” functions involved in the ordering and summarising of the party positions and the relevant legal sources, and must, therefore, have taken on decision making. This is a very difficult argument, and particularly so in big and complex cases. The Court of Appeal had no hesitation in holding that nothing about improper delegation of decision-making could be deduced from comparative hours.

Third, the Court of Appeal rejected the suggestion that there was a bright dividing line between the tribunal secretary drafting memoranda and papers for a tribunal for it to consider in its decision-making (which is permissible, but see below) and the secretary actually drafting decisive parts of an arbitral award (which according to Russia must be written by the arbitrators). In the Court of Appeal’s view, the argument that the decision-making parts of an award, as well as the dispositive part, must be all the tribunal’s own work, their very own words, because otherwise this would be tantamount to the “outright scrapping” of the delegation prohibition applicable to arbitrators, was flawed on two grounds:

• it was not enough to suggest that drafts of findings (still less memoranda and research papers) submitted by the secretary to the tribunal might have coloured or influenced the tribunal’s thinking. This was because “it is up to the arbitrators to check these texts for correctness and completeness” and it could not be shown that this had not taken place; and
• the three members of the tribunal had duly signed the award as theirs; it followed (there being no evidence to the contrary or to suggest that the tribunal was guilty of rubber-stamping without independent thought) “that the arbitrators have decided to assume responsibility for the draft versions of [the secretary], whether in whole or in part and whether or not amended by them”. That was fatal to an argument that the decision-making had been delegated to or shared with the secretary.

While the Hague Court of Appeal made clear that it was not required to pronounce on the correct allocation of tasks between arbitrators and secretaries or administrative assistants in general terms since it was only concerned with Russia’s case that what happened violated the proper composition of the tribunal, the breadth of the court’s reasoning confirms that, absent specific institutional rules to the contrary, “holding the pen” arguments, even in relation to substantive parts of an award, present in practice insurmountable problems for a party seeking to challenge an award.

Fourth, in terms of Russia’s argument that the tribunal had violated its arbitral mandate by using the secretary to write out substantive parts of the award and that this was contrary to an unwritten normative standard in international arbitration, restricting a secretary to preparatory papers or non-substantive drafting tasks only, the Court of Appeal found, unsurprisingly, that there was no such rule or standard. Even assuming that the secretary had drafted parts of the award after decisions had been taken by the tribunal, there was no expectation that these parts had to be written by the tribunal itself:

As long as no concrete party agreements have been made in this respect and the (substantive) decisions are taken by the arbitrators themselves without the influence of third parties, it is left to the discretion of the tribunal to what extent it wishes to use an assistant or secretary for the drafting of the arbitral award.

Will this decision at long last put the “tribunal secretary” issue to rest? The Hague Court of Appeal’s approach, like that of the English court in P v Q, upholds the finality of an arbitration award unless there is real and cogent evidence that something has gone wrong with the decision-making and that it can be demonstrated that the arbitral tribunal has allowed the secretary to decide or to help to decide the case. Arguments based on “necessarily reserved functions” founder in the absence of specific rules of demarcation, while those based on inferences from text and time-keeping are ultimately usually entirely equivocal. The Hague Court of Appeal’s decision may be seen as a welcome corrective to the purist approach and to require challenges on “fourth arbitrator” grounds to be based on concrete facts; that such facts may typically be virtually impossible to expose because of the confidentiality of the deliberations is not a justification for challenges based on inferences. The remedy lies in specific agreement on the role of administrative and other tribunal assistance or the choice of suitable institutional rules.

However, good arbitral practice must be (where there is no applicable institutional guidance) to explain in advance when appointing a tribunal secretary precisely what his or her role and activities will be and, if those change after appointment, given the exigencies of the case, to update the parties accordingly. One may feel a little sympathy for Russia in circumstances where the secretary “was only introduced as an assistant and contact person”. The court held that the tribunal had violated its mandate in that it “failed to fully inform the parties on this point of the nature of [the secretary’s] work” but, given its other findings of non-delegation, that was too minor to impugn the awards.

The “damages discretion” issue

During the arbitration, Russia put in no damages model of its own in response to the model advanced by the shareholders. Instead, it made extensive criticisms of the shareholders’ model which the state (perhaps ambitiously) argued vitiated the model to such an extent that no reliance should have been placed upon it by the tribunal. Faced with the competing arguments, the tribunal did not accept the shareholders’ valuation but assessed their loss based on its own hypothetical calculation, using the materials before it.

That set the ground for an equally ambitious argument before the Court of Appeal that the Tribunal violated its mandate by going outside the permissible limits of available findings on loss and damage and by adopting a new (and flawed) method of calculation on which Russia had had no opportunity to be heard. The fine line between a simple criticism of the tribunal’s decision on the merits (not open to review) and a departure from mandate was (as the Court of Appeal made clear) repeatedly crossed.

But the interest in the Court of Appeal’s decision lies in its emphatic restatement of the fact that once an arbitral tribunal is satisfied that loss of some kind and to some extent has been sustained by the claimant, an international arbitral tribunal then possesses a large measure of discretion as to how to assess what that loss was and what the monetary remedy should be, and its mandate is accordingly very wide. The tribunal was accordingly entitled to “play” with the claimant’s model using its own hypotheses to arrive at its own estimation of the loss, where the tribunal forms the view that it has sufficient evidence before it to make such an estimation, provided only that it does not go beyond what the claimant has itself claimed or outside the materials put before it. A tribunal will stay within its mandate, arising from the boundaries of the parties’ debate, if it bases its determinations on the submissions and evidence (including data) supplied by the parties, the court found.

It hardly helps an argument on “excess of mandate” where the party responding to the claim elects not to put forward its own counter-model as to what the loss was. This is to leave the tribunal with one model, and a range of criticisms which will have to be put together to arrive at an estimation if the tribunal is satisfied that some substantial loss has been or must have been sustained. As the Court of Appeal put it:

it is also important to note that although the Russian Federation has extensively criticised [the shareholders’] damage calculations, it has not itself proposed an alternative valuation of Yukos [...] It is thus foreseeable that the tribunal, which took on board many of the criticisms of the Russian Federation of [the shareholders’] calculations, would calculate the damages itself, as it did in this case, on the basis of the assumptions of [the shareholders] which it considered acceptable and applying the criticisms made by the Russian Federation.

The Court of Appeal’s two-stage approach was orthodox for the quantification for damages in investment treaty arbitration. This corresponds with the realistic approach to the scope of a tribunal’s discretion in damages in English law. While a tribunal may not adopt a wholly different methodology, it may arrive at its own view within the range of possibilities open on the evidence before it.

The practical implications

In addition to the various legal consequences that flow from the Court of Appeal’s decision, there are a number of practical implications that result as well. Undoubtedly the most important among these is the reinstatement of the US$50 billion awards, giving the Yukos shareholders renewed hope in their continuing enforcement efforts – at least pending further appeal to the Dutch Supreme Court by Russia.

With that said, observers will have noted with interest President Vladimir Putin’s declaration of intent at his State of the Union address of 15 January 2020 to introduce changes into the foundational law of  Russia “that would explicitly guarantee the priority of the Russian Constitution in our legal framework”. The President specified that this meant that the “requirements of international law and treaties as well as decisions of international bodies can be valid on the Russian territory only to the point that they do not restrict the rights and freedoms of our people and citizens and do not contradict our Constitution.”

A proper legal analysis of what the president has in mind will have to await further details of the proposed steps themselves. However, on any conception of the interplay between Russian law and international law, any legislative attempt to assure the supremacy of Russian law is likely to be challenging. Nevertheless, the Hague Court of Appeal’s decision can only have had a galvanising effect on Russia’s resolve to implement measures in domestic law to provide greater influence over the effectiveness of international arbitral awards such as that secured by the Yukos shareholders

Ultimately, however, such steps, however wide-ranging as a matter of Russian law, are unlikely to assist Russia in resisting enforcement steps taken by the shareholders abroad, where the relevant domestic courts that are asked to give effect to the award through the New York Convention are likely to begin and end their analysis with the award itself and the decisions of the Dutch courts at the seat of the arbitration.