The axiom of impartiality: Halliburton v Chubb
Back to Arbitration Committee publications
Ruth Keating
39 Essex Chambers, London
ruth.keating@39essex.com
Samar Abbas Kazmi
Atkin Chambers, London
sak@atkinchambers.com
Introduction
The UK Supreme Court’s much-anticipated judgment in Halliburton Company (Appellant) v Chubb Bermuda Insurance Ltd (formerly known as Ace Bermuda Insurance Ltd) (Respondent) [2020] UKSC 48 (on appeal from [2018] EWCA Civ 817) opens with the recognition of an axiom which goes to the heart of justice: ‘it is axiomatic that a judge or an arbitrator must be impartial; he or she must not be biased in favour of or against any party in a litigation or reference.’
After restating the law on the extent to which an arbitrator can, without disclosure and without giving rise to an appearance of bias, accept appointments in multiple references concerning the same or overlapping subject matter with only one common party, the judgment concludes that the arbitrator in question was in breach of a legal duty of disclosure but that, on the facts, he did not need to be removed. With that conclusion, many in the international arbitration community may well wonder whether the ‘axiomatic’ duty has been given sufficient bite.
However, it is important to consider the Supreme Court’s decision in the context of the appeal raised from the Court of Appeal, whose reasoning had raised more doubts than it had offered clarity, thereby causing considerable disquiet in the international arbitration community. Amongst the criticisms of the Court of Appeal’s decision was that a consequence of its reasoning would be that arbitrators could accept multiple appointments in arbitrations with overlapping subject matters without any disclosure, and potentially without meaningful risk of removal. The decision of the Court of Appeal meant that failure to disclose multiple appointments would not necessarily give rise to successful challenges to an arbitrator’s impartiality, as something more was required. This was ‘something’ in the words of the Court of Appeal, ‘of substance’. As to what that meant, this was left unclear by the Court of Appeal. For that reason, the Supreme Court’s decision was eagerly awaited.
As discussed below, the impact of the Supreme Court’s decision cannot be understated for the importance of London as an arbitral seat. In many ways, the Supreme Court’s judgment provides some much-welcomed clarity. However, it raises a number of further issues which will need to be contextualised and considered over time.
Background
Given the interest in this case, the facts of Halliburton Company v Chubb Bermuda Insurance Ltd have been well rehearsed in many articles on the topic. Accordingly, the factual and appellate background are only considered briefly here.
The Appellant, Halliburton, provided cementing and well monitoring services to BP Exploration and Production Inc (BP) in the Gulf of Mexico. Halliburton entered into a liability policy with the Respondent, Chubb Bermuda Insurance Ltd (Chubb). Transocean Ltd (Transocean) also provided services to BP. Those services overlapped with those provided by Halliburton. Transocean was also insured with Chubb.
In 2010, the Deepwater Horizon oil spill occurred. The explosion and fire on an oil rig in the Gulf of Mexico resulted in thousands of civil claims being brought against BP, Halliburton and Transocean. BP in turn claimed against Halliburton and Transocean.
A trial was held in the United States. The judgment apportioned blame between the parties and Halliburton concluded a settlement to agree the amount of damages. When Halliburton sought to claim a proportion of this settlement under its insurance policy, Chubb declined to pay Halliburton’s claim. As a result, arbitration was commenced. Both Halliburton and Chubb selected their own arbitrator. However, the parties were unable to agree the chair of the arbitration. This ultimately resulted in an application to the High Court, in which Chubb’s first-choice candidate – Kenneth Rokison QC – was selected.1
The issue for the Supreme Court arises out of Halliburton’s discovery in 2016 that, following Rokison’s appointment and without Halliburton’s knowledge, Rokison had accepted appointment as an arbitrator in two other references, both of which arose out of the same Deepwater Horizon incident. These were Transocean’s claim against Chubb and a nomination by another insurer to arbitrate another claim by Transocean arising out of the same incident.
In December 2016, Halliburton applied to the court to remove Rokison as arbitrator on the grounds of perceived bias. That application was refused by the High Court. Halliburton appealed against the decision that there were no grounds for removing Rokison under section 24(1)(a) of the Arbitration Act 1996 (the 1996 Act).
The Court of Appeal in turn dismissed Halliburton’s appeal. The Court of Appeal dismissed the appeal as, based on the facts of the case, there was no real possibility that the arbitrator was biased when viewed from the perspective of the ‘fair-minded and informed observer’. The mere fact that an arbitrator accepts multiple appointments in overlapping subject matter with only one common party does not, of itself, give rise to an appearance of bias. Nevertheless, the court also remarked that, in keeping with best practice in international arbitration and as a matter of law, disclosure of the appointments should have been made. Notwithstanding that, the court concluded that non-disclosure alone would not have led the fair-minded and informed observer to conclude that there was a real possibility of bias.
Doubt left by the Court of Appeal judgment
The issue with the Court of Appeal’s judgment was that the law itself was relatively uncontroversial, but the Court of Appeal’s trusting application of those principles gave rise to doubt and concern amongst the arbitration community.
Amongst the aspects in which the Court of Appeal’s judgment was criticised as being confusing and contradictory was the question of whether the non-disclosure itself met the threshold for apparent bias. The Court of Appeal stated that an additional factor was needed, referring to ‘something more’ but did not elaborate on this.
The Court of Appeal rightly highlighted that disclosure should be given of facts and circumstances known to the arbitrator which would or might give rise to justifiable doubts about his impartiality, as at [55]-[71] of its judgment. However, the Court of Appeal’s application of that principle to the particular facts of this case were unhelpful and seemingly contradictory. The court emphasised that if a disclosure that ought to have been made had not been made, that would mean that the arbitrator would not have displayed the ‘badge of impartiality’. The court accepted that, in this case, Rokison ought to have made disclosure to Halliburton at the time of his appointments. This was both as a matter of good practice in international commercial arbitration and as a matter of law. Nonetheless, the court concluded that the fair-minded and informed observer would not conclude that there was a real possibility that he was biased. In this case, this was because the omission was accidental rather than deliberate and that there was a very limited degree of overlap between the proceedings. There was therefore no substance to Halliburton’s challenge. This apparent contradiction in the Court of Appeal’s judgment was one of the main areas of confusion left by the decision.2
A further difficulty in the Court of Appeal’s judgment is the reliance it placed on the omission being accidental as compared to deliberate. This distinction is questionable. A deliberate omission would clearly reflect bias. However, an accidental one can be equally detrimental. An unconscious bias, almost by definition, is accidental and yet can cause no less concern or harm to a party.Further, there is an evidential difficulty faced by a party seeking to establish whether a unilateral act or omission by an arbitrator was ‘accidental’ or ‘deliberate’, when the whole matter is premised on the absence of information. The means by which a party could practically demonstrate whether an omission was accidental or, indeed, deliberate is difficult to envisage.
An additional question which was left outstanding from the Court of Appeal’s judgment was how the fair-minded observer test should take account of the international arbitration context in which this question is asked. Should these expectations, for instance, differ to those in ordinary litigation? This question has been elaborated on by the Supreme Court.
The Supreme Court decision
Halliburton appealed the Court of Appeal’s decision to the Supreme Court. The appeal was heard on 12 and 13 November 2019 by a division of the Supreme Court comprising Lord Reed, Lord Hodge, Lady Black, Lord Lloyd-Jones and Lady Arden.
On 27 November 2020, the Supreme Court handed down its long-awaited judgment.3 With a unanimous judgment, the court dismissed the appeal. The judgment was delivered by Lord Hodge (with whom Lord Reed, Lady Black and Lord Lloyd-Jones agreed) and a brief concurring judgment was delivered by Lady Arden.
Halliburton asked the Supreme Court to rule on two questions, namely:
- whether and to what extent an arbitrator may accept appointments in multiple references concerning the same or overlapping subject matter with only one common party without thereby giving rise to an appearance of bias at [125]; and
- whether and to what extent the arbitrator may do so without disclosure at [132].
Grounds
Turning to the first ground, the Supreme Court held that there may be circumstances in which the acceptance of multiple appointments with overlap with only one common party ’might reasonably cause the objective observer to conclude that there is a real possibility of bias’ (at [152]). The Court held that the objective test of the fair-minded and informed observer applies equally to judges and all arbitrators. The court continued that, in applying that test, it would be wrong to have regard to the particular characteristics of the parties to the arbitration.
The Court added that there is no difference between the test in section 24(1)(a) of the 1996 Act, which speaks of the existence of circumstances ‘that give rise to justifiable doubts as to [the arbitrator’s] impartiality’ and the common law test of whether the fair-minded and informed observer would conclude there is a real possibility of bias (at [55]). However, in applying the test to arbitrators, the Court held it is important to bear in mind the differences in nature and circumstances between judicial determination of disputes and arbitral determination of disputes (at [56]-[63]):
- First, judges resolve civil disputes in courts which are, as a general rule, open. By contrast arbitration is generally conducted in private. Given arbitrators and the parties to an arbitration are generally under a duty of privacy and confidentiality this means there is not a ready means of discovering the existence of multiple appointments. This means a premium is put on frank disclosure.
- Second, an arbitrator is not subject to appeals on issues of fact and often not on issues of law. There are therefore limited powers of review.
- Third, unlike judges, arbitrators are privately paid and so nomination as an arbitrator in a particular case gives the arbitrator a discrete financial benefit.
- Fourth, arbitrators have a wide range of experience. Some may have very extensive experience of arbitration practice while others may have very limited involvement in and experience of arbitration. Moreover, arbitrators in international arbitration come from many jurisdictions and legal traditions, and may have divergent views on what constitutes ethically acceptable conduct.
- The fifth point made by the Court overlaps with the first point. The party which is not common to the various arbitrations has no means of informing itself of the evidence led before and legal submissions made to the tribunal. The Court noted that it is not unusual in commercial litigation for an interested party to instruct its lawyer to sit in on a court case involving other parties which may have a bearing on its interests in a separate action. Such an expedient is generally not available in arbitration.
- Finally, in arbitrations where the parties have, or one party has, an expectation that the party-nominated arbitrator will be predisposed towards it, it is perceived that the person chairing the tribunal, whether appointed by the party-nominated arbitrators jointly or by an appointing institution or the court, has a particular role in making sure that the tribunal acts fairly and impartially.
On the second ground, disclosure, the Supreme Court held as follows:
- Where the hypothetical observer would conclude that circumstances as at and from the date when the duty arose might reasonably give rise to a real possibility of bias, the arbitrator will be under a legal duty to disclose such appointments (at [122]).
- The legal obligation can arise when the matters to be disclosed fall short of matters which would cause the informed observer to conclude that there was a real possibility of a lack of impartiality. It was sufficient that the matters are such that they are relevant and material to such an assessment of the arbitrator’s impartiality and could reasonably lead to such an adverse conclusion (at [116]).
- A question, on which the High Court and the Court of Appeal reached contradictory conclusions, and which was material to this appeal, was whether disclosure is a legal duty in English law or merely good arbitral practice unless the parties submit their dispute to arbitration under arbitral rules which impose a legal obligation (at [74]). The Supreme Court held that the duty of disclosure is not simply good arbitral practice but is a legal duty in English law (at [76]). It is a component of the arbitrator’s statutory obligations of fairness and impartiality.
- A failure to make disclosure does not necessarily lead to a removal of the arbitrator, but is a factor that the fair-minded and informed observer would consider when making a determination (at [155]).
- Equally, if a matter would give rise to justifiable doubts as to an arbitrator’s impartiality, the disclosure of that matter would not, as a general rule, remove this conflict (at [108]).
- The tension between the duty of disclosure and the duty of privacy and confidentiality of arbitrations was considered by the court. The court held that where the information which must be disclosed is subject to an arbitrator’s duty of privacy and confidentiality, disclosure can be made only if the parties to whom the obligations are owed give their consent. Regard must be had to the relevant custom and practice to ascertain whether consent can be inferred (at [88]-[89]).
In her concurring judgment, Lady Arden agreed with Lord Hodge but made a few further points to reinforce or, in some instances, qualify the conclusions reached. She noted that the duty of disclosure is a secondary obligation arising from the arbitrator’s primary duty to act fairly and impartially.
- Unless the arbitration is one where there is an accepted practice of dispensing with the need to obtain parties’ consent to further appointments, the arbitrator should proceed on the basis that a proposed further appointment involving a common party and overlapping subject matter is likely to require disclosure of a possible conflict of interest (at [164]).
- Lady Arden added that, in general, high-level disclosure about a proposed appointment in a further arbitration can be made without any breach of confidentiality by naming only the common party (which may be taken to have consented to disclosure) but not the other parties to the arbitration (at [180]).
The court therefore held that:
- Rokison had breached his legal duty of disclosure in the first reference between Halliburton and Chubb, in respect of the subsequent arbitrations. He should at the time of appointment have disclosed (1) the identity of the common party who was seeking the appointment of the arbitrator (in this case Chubb); (2) the nature of the appointment in the subsequent references; and (3) a statement that the subsequent references arose out of the same incident (at [146]).
- Rokison’s failure to disclose his appointment in the subsequent reference, which was a potentially overlapping arbitration with only one common party, was a breach of his legal duty of disclosure (at [147]). The fair minded and informed observer, if they had considered the question at or around the date of acceptance, may well have concluded that there was a real possibility of bias. However, it was not necessary to express a concluded view on this as that is not the correct time to ask the question (at [147]).
- The court applied the common law test, taking into account the relevant facts known at the time of hearing of the application for removal, and refused to hold that the arbitrator should be removed on the basis that:
- there appeared to have been a lack of clarity in English case law as to whether there was a legal duty of disclosure and whether disclosure was needed;
- the time sequence of the three references may explain why Rokison saw the need to disclose one reference to Transocean but did not identify the need to tell Halliburton about another reference;
- in this case it was likely that there would not be any overlap between the references;
- there was ‘no question’of Rokison having received any secret financial benefit; and
- there was no basis for inferring unconscious bias in the form of subconscious ill-will in response to the robustness of the challenge mounted on behalf of Halliburton (at [149]).
Commentary on the judgment
A number of issues arise out of the judgment of the Supreme Court which will require further consideration.
Fact-specific assessment: in many ways, the judgment does not answer some of the questions practitioners and arbitrators may have sought. However, those unanswered questions must be seen in the context of a judgment which was no doubt intended to be flexible but which also strengthened the duty of disclosure – an area which was left in real doubt following the divergence between the High Court and Court of Appeal.
At various points in the judgment, it is evident that the court was at pains to ensure that its analysis, while being of wider application on the core legal questions, was highly fact-specific. The decision itself, concerning a Bermuda Form arbitration, necessarily may apply more naturally to certain fact patterns than others.4 Furthermore, the court made specific reference to practices or customs within certain markets which do not require disclosure of acceptance of multiple appointments, for example the Grain and Feed Trade Association (GAFTA) and the London Maritime Arbitrators Association (LMAA). These aspects of the court’s decision are likely to vex those hoping for a clearly enunciated principle which could be easily applied across different factual scenarios and contexts.
However, notwithstanding that broadness, an important takeaway from the Supreme Court’s judgment is that the decision develops the law of arbitration by confirming that disclosure is not simply a matter of good practice but is a legal duty.5
Other appointment conflicts: it is important in assessing the judgment and its implications to keep in mind the factual backdrop of the decision. Halliburton concerns multiple appointments with overlapping parties. However, other common potential appointment conflicts which can arise in international arbitration, such as multiple appointments (where the arbitrator is repeatedly appointed by the same party or firm of solicitors) and string arbitration arising out of a chain of contracts, are not directly addressed. Practitioners and arbitrators will therefore have to extract the relevant principles discussed in Halliburton and apply them to the situations in which they find themselves.
Financial benefits: the Supreme Court sidestepped the issue of financial benefit resulting to arbitrators from multiple appointments. This was an issue which the Court of Appeal dismissed. A question arises as to whether this could form the basis of an unconscious incentive, whereby arbitrators avoid unsettling the parties who frequently appoint them. One would think this question then falls to be considered on a case specific basis as part of the overall assessment of ‘the objective observer’. However, this point has been left unaddressed by the Supreme Court’s judgment.
The risk of increased challenges: The risk of increased challenges and the possibility of a rise in personal claims against arbitrators were matters which were referred to by the court. However, in the absence of evidence demonstrating the risk of this occurring, these concerns were discounted. As noted by the court such challenges rarely succeed and some protection is offered by way of section 29 of the 1996 Act, so long as the arbitrator has not acted in bad faith.6
Nonetheless, the outcome may be that parties and arbitrators adapt their approach following this clear enunciation of the duty of disclosure. For example, at [107], Lord Hodge outlined that he agreed with the Court of Appeal’s formulation of the duty of disclosure subject to one qualification, which concerns the words ‘known to the arbitrator’. Lord Hodge continued that an arbitrator can disclose only what they know and is, as a generality, not required to search for facts or circumstances to disclose. However, he did ‘not rule out the possibility of circumstances occurring in which an arbitrator would be under a duty to make reasonable enquiries in order to comply with the duty of disclosure’.
Following from this, it would be unsurprising if arbitral institutions were to outline particularised disclosure rules, where they do not already have them – either by way of amendment to their existing rules or standalone guidance, to specifically address the question of the arbitrator’s duty of disclosure. This is a particular advantage which arbitral institutions could offer by providing clarity to the parties and a private means by which to raise an allegation of bias, as compared with ad hoc arbitrations.
This, like other points raised in the judgment, go some way to offering clarity on the duties of arbitrators and expectations of parties. However, the nature of the fact-specific application of those broad principles has meant that the application of those principles in practice will need to be developed by practitioners and arbitrators. Illustrating that the area of conflicts of interest causes difficulty in a variety of contexts, is the recent decision of the Court of Appeal (handed down on 11 January 2021) in Secretariat Consulting PTE Ltd & Ors v A Company [2021] EWCA Civ 6. The judgment provides yet more guidance from the courts on avoiding conflicts of interest, this time in the context of expert witnesses. Like Halliburton, which the Court of Appeal cited, Coulson LJ highlighted that ‘a conflict of interest is a matter of degree’. This very much reflects the tone of Halliburton: it is one thing for such principles to be enunciated, but how their application is likely to develop is a matter which is far from clear.
[1] It is noted that the arbitrator, Kenneth Rokison, had been unnamed in the decisions in the lower courts. Lord Hodge noted that Rokison had a long-established reputation for integrity and impartiality. However, the protection of that reputation was not a sufficient ground for anonymity, particularly when the courts below had relied on that reputation in their reasoning and the names of the parties to the arbitration had been disclosed. In any event, the challenge in the case involved no assertion of actual bias but relied entirely on an assertion of an objective appearance of bias. There were therefore no good grounds for maintaining the anonymity in the appeal.
[2] This decision perhaps best sits within a general approach of the courts as being non-interventionist and pro-arbitration. This certainly reflects the point made by the Supreme Court at [106] of its judgment that ‘challenges have rarely succeeded’ and at [68] that ‘[o]n other hand, the objective observer is alive to the possibility of opportunistic or tactical challenges’. Nonetheless, that context cannot fully address the apparent lack of consistency between the Court of Appeal’s reasoning and the ultimate outcome reached.
[3] We note that the length of time taken to deliver this judgment was unusually long. There may well have been good reasons for this delay but we respectfully suggest that, given that the eyes of the international arbitration community were on the Court, it may have been desirable to either expedite the decision or, if that was not possible, to acknowledge the length of time in the judgment itself.
[4] Differing, for example, to other arbitrations which may be subject to their own arbitral rules.
[5] Importantly, the judgment has clarified this legal duty and so in respect of international arbitration. For example, similarly in Hong Kong-seated arbitrations, section 25 of the Arbitration Ordinance (Cap. 609) imposes a duty, based on Article 12 of the UNCITRAL Model Law (Grounds for challenge), that: ‘[w]hen a person is approached in connection with his possible appointment as an arbitrator, he shall disclose any circumstances likely to give rise to justifiable doubts as to his impartiality or independence. An arbitrator, from the time of his appointment and throughout the arbitral proceedings, shall without delay disclose any such circumstances to the parties unless they have already been informed of them by him’.
[6] Troublingly, some of the obiter remarks by the court in this regard may lead to new debates and complications. Amongst these is the comment (at [111]) that ‘in cases where the matter is serious but the non-disclosure of that matter, on later examination, does not support the conclusion that there is apparent bias, the arbitrator might, depending on the circumstances, face an order to meet some or all of the costs of the unsuccessful challenger or to bear the costs of his or her own defence’. It is unclear how and when the envisaged costs sanctions may apply.