Halliburton Company v Chubb Bermuda Insurance Ltd & Ors

This is the judgment of the Court.

This appeal raises issues of importance in relation to commercial arbitration law and practice. The specific issues upon which the judge gave permission to appeal may be summarised as follows:

The second of those issues gives rise to the consideration of two further general issues, namely:

On 20 April 2010 there was an explosion and fire on the Deepwater Horizon oil rig in the Gulf of Mexico, when a well which was in the process of being plugged and temporarily abandoned suffered a blow out (“the incident”).

BP Exploration and Production Inc (“BP”) was the lessee of the rig. Transocean Holdings LLC (“Transocean”) was the owner of the rig and had been engaged by BP to provide crew and drilling teams. The Appellant (“Halliburton”) provided cementing and well-monitoring services to BP in relation to the temporary abandonment of the well.

Both Transocean and Halliburton purchased liability insurance on the Bermuda form from the First Respondent (“Chubb”). It appears that the material policy terms were the same. Halliburton’s insurance policy provided coverage of US$100 million excess US$500 million. It was governed by New York law but provided for arbitration in London by a tribunal consisting of three arbitrators, one appointed by each party and the third by the two arbitrators so chosen; in the event of disagreement between the arbitrators as to the choice of the third, the appointment was to be made by the High Court.

Following the incident, numerous claims were made against BP, Halliburton and Transocean by the US Government and corporate and individual claimants. The US Government claims were for civil penalties under various federal statutes. The private claims for damages were pursued through a Plaintiffs’ Steering Committee (“PSC”). Many of the claims were consolidated into a single ‘Multi District Litigation’.

Following a liability trial in the Federal Court for the Eastern District of Louisiana, judgment was given on 4 September 2014 holding the apportionment of blame to be BP 67%; Transocean 30% and Halliburton 3%. Shortly before judgment, Halliburton concluded a settlement of the PSC claims against it in the sum of approximately US$1.1 billion. Following the judgment, Transocean settled the PSC claims for some US$212 million and paid civil penalties of about US$1 billion to the US Government.

Halliburton made a claim on its liability insurance against Chubb. However, Chubb refused to pay Halliburton’s claim, contending amongst other things that Halliburton’s settlement of the claims was not a reasonable settlement, and/or that Chubb had reasonably not consented to the settlement.

Halliburton commenced arbitration by appointing N, the third respondent, as its arbitrator on 27 January 2015 (“reference 1”). The fourth respondent, P, was appointed on behalf of Chubb. The identity of the third arbitrator could not be agreed and so an application was made to the High Court for appointment of a third arbitrator. Following a contested hearing, in which a number of candidates were put forward on both sides, Flaux J appointed M, the second respondent, as the third arbitrator by an order of 12 June 2015. M was Chubb's preferred candidate. Halliburton’s main objection to Chubb’s candidates, including M, was that they were English lawyers and this was a policy governed by New York law. Halliburton did not seek to appeal against that order.

Prior to expressing his willingness to be appointed, M disclosed that he had previously acted as arbitrator in a number of arbitrations in which Chubb was a party, including appointments on behalf of Chubb, and that he was currently appointed as arbitrator in two pending references in which Chubb was involved.

Halliburton served its Statement of Claim in reference 1 on 18 September 2015. Chubb served its Statement of Defence on 11 December 2015.

In December 2015 M accepted appointment by Chubb through Clyde & Co, who were also Chubb's solicitors in reference 1, in relation to an excess liability claim arising out of the incident made by Transocean under its liability insurance policy with Chubb (“reference 2”). The same manager, Mr Trimarchi, was responsible for monitoring the claims made by both Transocean and Halliburton on behalf of Chubb and took the decision to refuse the claim in each case.

Prior to his acceptance of this appointment, M disclosed to Transocean his appointment in reference 1 and in the other Chubb arbitrations which had been disclosed to Halliburton. Transocean raised no objection. M did not, however, disclose to Halliburton his proposed appointment by Transocean.

In August 2016 M accepted appointment as a substitute arbitrator in another claim made by Transocean against a different insurer on the same layer of insurance (“reference 3”). This proposed appointment was also not disclosed to Halliburton.

In references 2 and 3 there was an order for a trial of a preliminary issue which was potentially dispositive of the claims, if decided in favour of the insurers. It involved construction of the policy terms on undisputed facts relating to the exhaustion of underlying layers by reference to the fines and penalties paid by Transocean. The preliminary issue was heard in November 2016.

On 10 November 2016 Halliburton learned of M’s appointment in references 2 and 3. On 29 November 2016 Halliburton’s US lawyers, K & L Gates, wrote to M, referring to the International Bar Association Guidelines on Conflicts of Interest in International Arbitration (“the IBA Guidelines”) concerning the continuing duty of disclosure of potential conflicts of interest, and asking for clarifications and explanations.

M replied by email on 5 December 2016, explaining in outline how he came to be appointed in references 2 and 3. He stated that he had not made disclosure to Halliburton at the time of those appointments because it did not occur to him at the time that he was under any obligation under the IBA Guidelines to do so, an explanation which is accepted as truthful by both parties. He further stated as follows:

“I do not think and did not think that the above circumstances put any obligation upon me to make any disclosure to you or your clients under the IBA Guidelines. However, I appreciate, with the benefit of hindsight, that it would have been prudent for me to have informed your clients through your firm, and I apologise for not having done so.

It is correct that all three References arise from the Deepwater Horizon incident, but it is not the case, as you suggest, that they raise the same or even similar issues. The two claimants, Halliburton and Transocean, as I understand it, performed very different roles and the issues were totally different and, so far, beyond matters which are public knowledge, my only involvement in the Transocean cases, has concerned the issue of construction argued by Counsel in two 2-day Hearings, without any evidence save as to the circumstances of the making of the relevant insurance contract. I have received no information which would not be shared by my co-arbitrators in the Halliburton case.

Both you and your clients have my assurance that during the period of about 20 years during which I have practised as a full-time international commercial arbitrator, I have at all times remained independent and impartial and will continue to do so.

That said, I readily acknowledge that it is important that both parties in arbitration should share confidence that the dispute will be determined fairly on the evidence and the law without bias.

I do not believe that any damage has been done but, if your clients remain concerned, I would be prepared to consider tendering my resignation from my appointment in the two Transocean cases if the results of the determination of the preliminary issues of construction, which are likely to be issued shortly, does not effectively bring them to an end.”

Halliburton responded repeating its concerns about M’s impartiality and suggesting that he resign, to which Chubb was not prepared to agree. M responded further by email of 15 December 2016 in which he stated:

“It is in accordance with my duty to both parties that my response seeks to take into account what I believe to be the best interests of both.

I do not think that it would be helpful to either party for me to continue the debate as to whether or not, by accepting appointment in the two Transocean arbitrations, I was in breach of any duty to Mr Birsic's clients by failing to disclose the fact, and presumably, giving them an opportunity to object. I would merely add that, even if the IBA Guidelines did apply (and I think Mr Payton is probably right in his view that they did not) I remain unpersuaded that I was in breach of them. However, I have accepted in my earlier letter that, with the benefit of hindsight, it would have been prudent for me to have made disclosure to avoid any sense of a lack of transparency on my part.

In relation to the other points raised in Mr Birsic's letter I can only repeat that neither him nor his clients need have any fear that I will have learned anything in the course of the Transocean arbitrations which could be of any relevance in the Halliburton case. The points so far considered relate only to preliminary issues of construction as to the attachment point, and I learned nothing about the facts of the incident and its consequences which is not public knowledge and which would not be well known to my co-arbitrators …

Putting the above to one side, the current potion [sic] is clearly unsatisfactory, to say the least. I repeat that I believe it is of fundamental importance that both parties should have confidence in the impartiality of the members of the Tribunal, and in particular the chairman, and, if my first letter together with what I have added above does not both put Mr Birsic's and his clients’ minds at rest, there is what seems to be a total impasse between the parties, to both of whom I owe an obligation.

Mr Payton wishes me to remain as chairman and for the hearing to go ahead. But if I were to decline Mr Birsic's invitation to resign, I have little doubt that an application would be made to the court to remove me which may well take some time to resolve….

… were the decision left to me to be determined in accordance with my own self-interests, I would resign. I have no wish to continue to serve as chairman in a tribunal in a case in which one of the parties, through its legal team, has expressed serious doubts as to my impartiality. Furthermore, as you may know, I plan to retire later this year and would not wish that my long career as an international commercial arbitrator which has spanned over three decades should end with my being the subject of a debate in the Commercial Court as to whether I have behaved improperly.

However, as I have already indicated, I have duties to both parties: by accepting the Court's appointment as chairman, I undertook to continue to serve in that capacity until I had completed the task, unless prevented by circumstances beyond my control and I would, I think, be in breach of those duties were I simply to resign in the face of strong opposition from one party.

In these circumstances, might I venture to propose to the parties that, even now, they put aside their differences to the extent of concentrating their attention on trying to agree upon a mutually acceptable replacement chairman who would be available for the hearing, without spending further time on argument, and applications to the Court.

Were they to do so, I would gladly resign. If that does not occur, I fear that I would have no alternative but to leave my fate in the hands of the Court.”

On 21 December 2016 Halliburton issued a Claim Form seeking an order pursuant to section 24(1)(a) of the Arbitration Act 1996 (“the Act”) that M be removed as an arbitrator.

On 4 January 2017, M responded by email to further questions in relation to the overlap between the references, stating that he was “unaware that there were any common issues”. By an email of 5 January 2017, K & L Gates requested that M clarify whether he had seen any document in which Chubb or any other respondent in references 2 or 3 set out similar defences to those pleaded in reference 1. M did not respond to that email. On 10 January 2017 Halliburton obtained from Chubb the release of the pleadings in reference 2. These revealed that the pleadings relied on by Chubb in the two references were substantially similar, specifically that the settlement was not reasonable and Chubb had reasonably withheld its consent. There was an additional defence in reference 2, which was the subject of the preliminary issue determination.

Halliburton’s arbitration application was heard by the High Court on 12 January 2017, and on 3 February 2017 Mr Justice Popplewell delivered his judgment dismissing the application.

On 1 March 2017 the tribunals in references 2 and 3 issued awards deciding them in Chubb’s favour on the preliminary issues of policy construction. The effect of this was that the references were brought to an end, and the tribunal was not required to consider any issues relating to the reasonableness of the settlement.

On 5 December 2017, the tribunal in reference 1 issued its Final Partial Award on the merits (“the Award”), deciding in Chubb’s favour. One of the arbitrators, the third respondent N, issued “Separate Observations” in which he wrote that he was unable to join in the award as a result of his “profound disquiet about the arbitration’s fairness”, explaining that:

“…arbitrators who decide cases cannot ignore the basic fairness of proceedings in which they participate. One side secured appointment of its chosen candidate to chair this case, over protest from the other side. Without any disclosure, the side that secured the appointment then named the same individual as its party-selected arbitrator in another dispute arising from the same events. The lack of disclosure, which causes special concern in the present fact pattern, cannot be squared with the parties’ shared ex ante expectations about impartiality and even-handedness.”

The judge considered and addressed the three elements of M's conduct which were relied upon as giving rise to an appearance of bias: (1) his acceptance of the appointments in the Transocean arbitrations; (2) his failure to disclose those appointments to Halliburton; and (3) his response to the challenge to his impartiality.

In relation to element (1), the judge rejected the suggestion made that M’s appointment in references 2 and 3 involved him being given a secret benefit by Chubb in the form of the remuneration he would earn from the arbitration. He pointed out that “the duty to act independently and impartially involves arbitrators owing no allegiance to the party appointing them. Once appointed they are entirely independent of their appointing party and bound to conduct and decide the case fairly and impartially”. He also found that the appointment conferred no immediate benefit in terms of his fees, observing that the appointing party does not undertake to bear those fees and that the tribunal as a whole would decide who ultimately is to bear them, in the light of the course of the arbitration and the result.

The judge also rejected the contention that the overlap between the references was a matter of concern since it meant that M would learn information during the course of the Transocean references which was relevant to the issues in the Halliburton arbitration, and available to Chubb but not to Halliburton. He observed that it is “a regular feature of international arbitration in London that the same underlying subject matter gives rise to more than one claim and more than one arbitration without identity of parties”. He considered that this was desirable because (1) “parties should be free to appoint their chosen arbitrator in accordance with the procedure agreed in the arbitration clause in fulfilment of the contractual bargain”; (2) arbitrators are often chosen for their particular knowledge and expertise, but frequently comprise a limited pool of talent, and it is “undesirable that parties should be unnecessarily constrained in their ability to draw on this pool if there are multiple arbitrations arising out of a single event or overlapping circumstances”; and (3) the principle of finality is served “if the tribunal is already familiar with the background to and uncontroversial aspects of the subject matter of the dispute.”

The judge observed that: “generally, the fact that an arbitrator may be involved in an arbitration between party A and party B, whose subject matter is identical to that in an arbitration between party B and party C does not preclude him or her from sitting on both tribunals.” He considered that this was borne out by the duty enshrined in section 33 of the Act which requires an arbitrator to decide the case by reference to material available to the parties to the particular reference and by authority, citing the Court of Appeal decision in AMEC Capital Projects Ltd v Whitefriars City Estates Ltd. [2005] 1 WLR 723. He accordingly concluded that:

“29.

The informed and fair-minded observer would not therefore regard M as unable to act impartially in the reference between Halliburton and Chubb merely by virtue of the fact that he might be an arbitrator in other references arising out of the incident, and might hear different evidence or argument advanced in another such reference. The objective and fair-minded assessment would be that his experience and reputation for integrity would fully enable him to act in accordance with the usual practice of London arbitrators in fulfilling his duties under section 33 by approaching the evidence and argument in the Halliburton reference with an open mind; and in deciding the case, in conjunction with the other members of the tribunal, in accordance with such material, with which Halliburton will have a full and fair opportunity to engage.”

In these circumstances, his conclusion was that there was nothing in the acceptance of the Transocean appointments by M which gave rise to an appearance of bias against Halliburton, even if the issues which had to be decided in the references were identical or substantially overlapping, which he found they were not.

In relation to element (2), the judge held that his conclusion on element (1) meant that there was nothing to disclose: “If a particular circumstance does not give rise to any justifiable concerns as to an arbitrator's impartiality, then his failure to disclose that circumstance cannot, without more, give rise to any equivalent concern.”

He further held that even if disclosure ought to have been made, the failure to do so did not give rise to a real possibility of apparent bias against Halliburton. M’s unchallenged explanation in correspondence was that he did not do so because it did not occur to him that there was any obligation to do so. The judge held that: “even if such honest belief were mistaken (which it is not), the fair-minded observer would not think that it would raise a real possibility of apparent bias”.

In relation to element (3), the judge went through each of the complaints made about M’s response to the challenge to his impartiality and rejected them. He concluded that M “dealt with the challenge in a courteous, temperate and fair way, demonstrating commendable even-handedness. His response would only serve to reinforce the confidence any fair-minded observer would have in his ability and intention to continue to conduct the reference fairly and impartially.”

The grounds of appeal are:

Under English law an arbitrator’s duties are governed by the Act. The Act imposes a duty of impartiality.

The general principles set out in section 1 of the Act include that the object of arbitration is to obtain the fair resolution of disputes “by an impartial tribunal”.

The duty of impartiality is reflected in section 24 of the Act which provides that the court has power to remove an arbitrator on the grounds:

“(1)(a) that circumstances exist that give rise to justifiable doubts as to his impartiality …”

It is also reflected in section 33 of the Act which imposes a duty on arbitrators to “act fairly and impartially as between the parties”.

Although lack of independence may give rise to justifiable doubts of impartiality, the Act deliberately did not include this as a separate ground for removal, as explained in the DAC Report on the Arbitration Bill in the commentary on Clause 24. In particular, it was there pointed out that there would be no point in including lack of independence as an independent ground if it covered cases which did not give rise to justifiable doubts as to impartiality and that there was “no good reason for including “non-partiality” independence as a ground for removal”. It was also noted that “there may well be situations in which parties desire their arbitrators to have familiarity with a specific field, rather than being entirely independent”.

Section 24 has been held to reflect the common law test for apparent bias, namely whether the fair-minded and informed observer, having considered the facts, would conclude that there was a real possibility that the tribunal was biased.

This is an objective test and is not to be confused with the approach of the person who has brought the complaint. It involves taking a balanced and detached approach, having taken the trouble to be informed of all matters that are relevant – see, for example, Helow v Secretary of State for the Home Department [2008] 1 WLR 2416 at [2]-[3] per Lord Hope.

With one caveat, both parties accepted that the judge had directed himself correctly as to the law at [16] of his judgment. That caveat was Halliburton’s contention that the judge failed to have sufficient regard to the risk of unconscious bias. That risk, which the judge recognised (see, for example, at [62]) does not affect the relevant legal test for apparent bias, which he correctly set out. It provides an example of how bias may act, or appear to act, on the mind, but it is not part of the test for whether there is bias. We accept, however, that it is a relevant risk for the fair-minded and informed observer to take into account.

At the heart of Halliburton’s appeal is its contention that the judge failed to have proper regard to the unfairness which may arise where an arbitrator accepts appointments in overlapping references with only one common party.

The essence of that potential unfairness is said to be information and knowledge which the common party acquires, unknown to the other party. The common party may obtain the advantage of, for example, making submissions and adducing evidence that influence the common arbitrator without the participation or knowledge of the other party; sharing with the common arbitrator relevant information that is not shared with the other party; and the opportunity to assess the views of the common arbitrator in one arbitration and to tailor its submissions and evidence accordingly in the other. Given the privacy and confidentiality of arbitration, the possibility of such advantage is likely to be unknown to the other party.

There have been recent judicial observations recognising this as a legitimate concern. In Guidant LLC v Swiss Re International SE [2016] EWHC 1201 Leggatt J was concerned with an application to appoint a third arbitrator in two arbitrations brought by Guidant against Swiss Re companies. Guidant sought the appointment of the same third arbitrator as in its arbitration against Markel, which raised closely overlapping issues. Guidant contended that having a common third arbitrator would reduce costs and delay and minimise the risk of inconsistent decisions. Leggatt J observed as follows at [9]:

“9.

In circumstances where the arbitrations will therefore be taking place separately, it seems to me that Swiss Re has a legitimate basis for objecting to the appointment as the third member of the tribunals in its arbitrations of the same person who is the third arbitrator and chair of the tribunal in the Markel arbitration. If the same person were to be appointed, there would be a legitimate concern that that person would be influenced in deciding the Swiss Re arbitrations by arguments and evidence in the Markel arbitration. Indeed, the likelihood that that would occur is implicit in the very argument which Guidant makes that appointment of the same person would minimise the risk of inconsistent decisions. Swiss Re is not a party to the Markel arbitration and will have no opportunity to be heard in that arbitration or to influence its outcome. Indeed, without a waiver of confidentiality, they will not be privy to the evidence adduced or the submissions made in the Markel arbitration. If the Markel arbitration were to be heard first, the members of the tribunal in that arbitration would form views, without any input or opportunity for input from Swiss Re, from which they may afterwards be slow to resile.”

In the light of these considerations Leggatt J did not appoint Guidant’s requested arbitrator as the third arbitrator in the exercise of his discretionary powers under s.18 of the Act. At the same time, he recognised, however, that the appointment of a common arbitrator did not justify an inference of apparent bias. Thus, he noted that the fact that Guidant had appointed the same arbitrator in all three arbitrations was not a ground upon which disqualification could be sought. As he stated at [10]:

“10.

I accept the submission made by Mr. Tse on behalf of Guidant that the appointment of a common arbitrator does not justify an inference of apparent bias. The fact that the same person has been appointed by Guidant as its arbitrator in the Markel arbitration is not, therefore, a ground on which an application could be made to seek to disqualify him from acting in the Swiss Re arbitrations. Guidant is entitled to choose the same individual as their arbitrator in all three arbitrations, as they have. But conversely Swiss Re, for their part, are in my view reasonably entitled to object to having forced upon them an arbitrator who has already been appointed in the Markel arbitration and about whose involvement in that arbitration they are entitled to feel the concern which I have indicated.”

Leggatt J therefore drew a distinction between the concern which Swiss Re “were entitled to feel” and a concern which would justify an inference of apparent bias.

A similar recognition of legitimate concern was expressed by Fraser J in Beumer Group UK Ltd v Vinci Construction UK Ltd [2016] EWHC 2283. In that case two construction adjudications were heard by the same adjudicator arising from the same underlying dispute, with one party appearing in both. That party advanced mutually inconsistent cases in the two adjudications. All of this was unknown to the other party. Fraser J held that the appointment of the common adjudicator and the conduct of that adjudication with all that involved, in terms of contact and the running of inconsistent cases, without notifying the other party, meant that this was a case of apparent bias. In relation to the contact between the adjudicator and the common party, he observed as follows at [31]:

“31.

If unilateral telephone calls are strongly discouraged (if not verging on prohibited) due to the appearance of potential unfairness, it is very difficult, if not in my judgment impossible, for an adjudicator to be permitted to conduct another adjudication involving one of the same parties at the same time without disclosing that to the other party. Conducting that other adjudication may not only involve telephone conversations, but will undoubtedly involve the receipt of communications including submissions, and may involve a hearing. If all that takes place secretly, in the sense that the other party does not know it is even taking place, then that runs an obvious risk in my judgment of leading the fair minded and informed observer to conclude that there was a real possibility of bias. All of this can be avoided by disclosing the existence of the appointment at the earliest opportunity.”

The legitimate concern identified by Leggatt J and Fraser J was addressed in a recent talk given by Jeffrey Gruder QC to the BILA on 21 July 2017 in which he described the problem as being one of “inside information” or “inside knowledge”. It is to be noted, however, that he said that his resulting “sense of unease” did not relate to justifiable doubts as to the arbitrators’ impartiality.

We accept that inside information and knowledge may be a legitimate concern for the parties to have in overlapping arbitrations involving a common arbitrator but only one common party. We agree, however, with Leggatt J and Jeffrey Gruder QC that, in itself, it does not justify an inference of apparent bias.

As the judge held, the starting point is that an arbitrator should be trusted to decide the case solely on the evidence or other material adduced in the proceedings in question. In this context the judge referred to the AMEC case at [20]-[21] where Dyson LJ stated as follows:

“20.

In my judgment, the mere fact that the tribunal has previously decided the issue is not of itself sufficient to justify a conclusion of apparent bias. Something more is required. Judges are assumed to be trustworthy and to understand that they should approach every case with an open mind. The same applies to adjudicators, who are almost always professional persons. That is not to say that, if it is asked to redetermine an issue and the evidence and arguments are merely a repeat of what went before, the tribunal will not be likely to reach the same conclusion as before. It would be unrealistic, indeed absurd, to expect the tribunal in such circumstances to ignore its earlier decision and not to be inclined to come to the same conclusion as before, particularly if the previous decision was carefully reasoned. The vice which the law must guard against is that the tribunal may approach the rehearing with a closed mind. If a judge has considered an issue carefully before reaching a decision on the first occasion, it cannot sensibly be said that he has a closed mind if, the evidence and arguments being the same as before, he does not give as careful a consideration on the second occasion as on the first. He will, however, be expected to give such reconsideration of the matter as is reasonably necessary for him to be satisfied that his first decision was correct. As I have said, it will be a most unusual case where the second hearing is for practical purposes an exact rerun of the first.

21.

The mere fact that the tribunal has decided the issue before is therefore not enough for apparent bias. There needs to be something of substance to lead the fair-minded and informed observer to conclude that there is a real possibility that the tribunal will not bring an open mind and objective judgment to bear.”

As the judge observed, these comments are equally applicable to arbitrators and to references involving a common party. Arbitrators are assumed to be trustworthy and to understand that they should approach every case with an open mind. The mere fact of appointment and decision making in overlapping references does not give rise to justifiable doubts as to the arbitrator’s impartiality. Objectively this is not affected by the fact that there is a common party. An arbitrator may be trusted to decide a case solely on the evidence or other material before him in the reference in question and that is equally so where there is a common party.

In the context of Bermuda form arbitrations, the position has been summarised as follows in Liability Insurance in International Arbitration, 2nd ed (2011), at para 14.32:

“The decision in Locabail (UK) Ltd v Bayfield Properties Ltd [2000] QB 451, and the foregoing discussion, is also relevant in the fairly common situation where a loss, whether from boom or batch, gives rise to a number of arbitrations against different insurers who have subscribed to the same programme. A number of arbitrations may be commenced at around the same time, and the same arbitrator may be appointed at the outset in respect of all these arbitrations. Another possibility is that there are successive arbitrations, for example because the policyholder wishes to see the outcome of an arbitration on the first layer before embarking on further proceedings. A policyholder, who has been successful before one tribunal, may then be tempted to appoint one of its members (not necessarily its original appointee, but possibly the chairman or even the insurer's original appointee) as arbitrator in a subsequent arbitration. Similarly, if insurer A has been successful in the first arbitration, insurer B may in practice learn of this success and the identity of the arbitrators who have upheld insurer A's arguments. It follows from Locabail and AMEC Capital Projects Ltd v Whitefriars City Estates Ltd [2005] 1 All ER 723 that an objection to the appointment of a member of a previous panel would not be sustained simply on the basis that the arbitrator had previously decided a particular issue in favour of one or other party. It equally follows that an arbitrator can properly be appointed at the outset in respect of a number of layers of coverage, even though he may then decide the dispute under one layer before hearing the case on another layer.”

We accordingly agree with the judge that the mere fact that an arbitrator accepts appointments in multiple references concerning the same or overlapping subject matter with only one common party does not of itself give rise to an appearance of bias. As Dyson LJ said, “[s]omething more is required” and that must be “something of substance”.

Whether and to what extent an arbitrator may accept such appointments without disclosure will be addressed below. For reasons there set out, we do not consider that the fact that such appointments may be accepted is determinative of whether disclosure should be given before accepting such appointments.

The Act sets out no requirements in relation to disclosure, but many institutional rules governing arbitration include provisions requiring disclosure to be made of facts or circumstances which may give rise to justifiable doubts as to an arbitrator’s impartiality.

Under the common law, judges should disclose facts or circumstances which would or might provide the basis for a reasonable apprehension of lack of impartiality.

In Davidson v Scottish Ministers (No 2) [2005] 1 SC 7 the House of Lords held that there was a risk of apparent bias where a judge was called upon to rule judicially on the effect of legislation which he or she had drafted or promoted. In addressing the issue of disclosure Lord Bingham described the position at common law in the following terms at [19]:

“Where a judge is subject to a disqualifying interest of any kind (‘actual bias’), this is almost always recognised when the judge first appreciates the substance of the case which has been assigned. The procedure is then quite clear: the judge should, without more, stand down from the case. It is rare in practice for difficulties to arise. Apparent bias may raise more difficult problems. It is not unusual for a judge, at the outset of a hearing, to mention a previous activity or association which could not, properly understood, form the basis of any reasonable apprehension of lack of impartiality. Provided it is not carried to excess, this practice is not to be discouraged, since it may obviate the risk of misunderstanding, misrepresentation or misreporting after the hearing. It is also routine for judges, before or at the outset of a hearing, to disclose a previous activity or association which would or might provide the basis for a reasonable apprehension of lack of impartiality. It is very important that proper disclosure should be made in such cases, first, because it gives the parties an opportunity to object and, secondly, because the judge shows, by disclosure, that he or she has nothing to hide and is fully conscious of the factors which might be apprehended to influence his or her judgment. When such disclosure is made, it is unusual for an objection to be taken ... (emphasis added)

There are of course a number of entirely honourable reasons why a judge may not make disclosure in a case which appears to call for it, among them forgetfulness, failure to recognise the relevance of the previous involvement to the current issue or failure to appreciate how the matter might appear to a fair-minded and informed observer who has considered the facts but lacks the detailed knowledge and self-knowledge of the judge. However understandable the reasons for it, the fact of non-disclosure in a case which calls for it must inevitably colour the thinking of the observer.”

The judgments of the Court of Appeal in Locabail (UK) Ltd v Bayfield Properties Ltd [2000] QB 451 and Taylor v Lawrence [2003] QB 528 are to similar effect.

In Locabail in the judgment of the court (Lord Bingham CJ, Lord Woolf MR, and Sir Richard Scott V-C) it was stated as follows at [21]:

“. . . If, in any case not giving rise to automatic disqualification and not causing personal embarrassment to the judge, he or she is or becomes aware of any matter which could arguably be said to give rise to a real danger of bias, it is generally desirable that disclosure should be made to the parties in advance of the hearing. If objection is then made, it will be the duty of the judge to consider the objection and exercise his judgment upon it. He would be as wrong to yield to a tenuous or frivolous objection as he would to ignore an objection of substance.” (emphasis added)

In Taylor Lord Woolf stated as follows in giving the judgment of the court at [64]:

“A further general comment which we would make, is that judges should be circumspect about declaring the existence of a relationship where there is no real possibility of it being regarded by a fair-minded and informed observer as raising a possibility of bias. If such a relationship is disclosed, it unnecessarily raises an implication that it could affect the judgment and approach of the judge. If this is not the position no purpose is served by mentioning the relationship. On the other hand, if the situation is one where a fair-minded and informed person might regard the judge as biased, it is important that disclosure should be made. If the position is borderline, disclosure should be made because then the judge can consider, having heard the submissions of the parties, whether or not he should withdraw. In other situations disclosure can unnecessarily undermine the litigant's confidence in the judge.” (emphasis added)

A recent example of a case in which it was held that disclosure should have been given in circumstances falling short of apparent bias is provided by the decision of the Cayman Islands’ Court of Appeal in Wael Almazeedi v Michael Penner and Stuart Sybermsa. That case concerned institutional links between a Qatari party involving Qatari government interests and the judge by reason of his appointment as a judge of the Qatar International Court and Dispute Resolution Centre (QICDRC). In giving the judgment of the Court on this issue Sir Bernard Rix stated at [50] as follows:

“50.

On the matter of disclosure, we consider that the judge ought to have disclosed his appointment to the QICDRC in a case which involved Qatari government interests in the form of the petitioning preference shareholders. That is independent of our decision on this appeal. The purpose of such disclosures, which often go beyond legitimate concerns as to independence and impartiality which would, subject to waiver, require a judge to recuse himself, is to enable the parties to consider the disclosures made and either to assure themselves in advance that there is no legitimate problem or to make submissions to the judge, or to finesse any potential problem by means of waiver. Such matters can only be efficiently and safely handled in advance. Once judgment has been entered, and a winner and a loser emerge, the matter becomes much more difficult. Losers feel aggrieved whatever the rights and wrongs of the situation are, specious claims to bias may be raised, and all the difficulties of retrospective consideration fall for debate and decision.”

At the time of the hearing of the present appeal the Privy Council decision on the appeal in the Almazeedi case was pending and it was agreed that this should be awaited and the parties should have the opportunity to make further submissions in the light of the Privy Council’s judgment. Judgment was handed down on 26 February 2018 – [2018] UKPC 3. Both parties have provided further written submissions in the light of that judgment, which we have taken into account. We also received written submissions in relation to the recent Court of Appeal decision concerning apparent bias in Bubbles & Wine Limited v Reshat Lusha [2018] EWCA Civ 468. That was a case on its own and different facts and we did not find it to be of any real assistance.

In Almazeedi the Privy Council agreed with what the Cayman Islands’ Court of Appeal had said about disclosure, observing at [60] that:

“21.

Against this background, the Court of Appeal repeated an observation made earlier in its judgment, namely that, although this was not the test of apparent bias, the judge ought to have disclosed his appointment in Qatar to enable the position to be clarified and considered and avoid possible later challenges such as the present. Mr Francis Tregear QC, representing the JOLs, realistically, did not take issue with this, and the Board need say no more than that it also agrees with it.”

By its decision the Privy Council (Lord Sumption dissenting) agreed with the Cayman Islands’ Court of Appeal that it had been inappropriate for the judge to sit without disclosure of his position in Qatar, but from an earlier date than that found by the Court of Appeal. The decision reached is encapsulated in [34] of the judgment:

“34.

In the result, the Board, with some reluctance, has come to the conclusion that the Court of Appeal was right to regard it as inappropriate for the judge to sit without disclosure of his position in Qatar as regards the period after 26 June 2013 and that this represented a flaw in his apparent independence, but has also come to the conclusion that that the Court of Appeal was wrong to treat the prior period differently. The judge not only ought to have disclosed his involvement with Qatar before determining the winding-up petition. In the Board's view, and at least in the absence of any such disclosure, a fair-minded and informed observer would regard him as unsuitable to hear the proceedings from at least 25 January 2012 on. The fact of disclosure can itself serve as the sign of transparency which dispels concern, and may mean that no objection is even raised. An alternative to disclosure might have been to ask the Chief Justice to deploy another member of the Grand Court, to which there would, so far as appears, have been no obstacle.”

The decision in Almazeedi supports the importance of disclosure, as borne out by the other authorities referred to. These authorities explain the important practical advantages of giving disclosure and addressing any issues which may arise at the outset. They also show that in borderline cases disclosure should be given - disclosure should be given of circumstances which would or might lead the fair-minded and informed observer, having considered the facts, to conclude that there was a real possibility that the tribunal was biased. An important practical reason for this approach is that it may be difficult at the time of considering whether to make disclosure to draw any firm conclusion as to whether or not the fair-minded and informed observer would so conclude. As the authorities make clear, the test is an objective one, to be judged by reference to what the fair-minded and informed observer would or might conclude.

We consider that the same approach applies to arbitral tribunals for the same reasons as have been given in the authorities we have mentioned. The test for apparent bias is the same and the practical advantages of early disclosure are just as important.

Many arbitration institutional rules impose a stricter test of disclosure, importing a subjective test. The IBA Guidelines, for example, require disclosure of facts or circumstances “that may, in the eyes of the parties, give rise to doubts as to the arbitrator’s impartiality or independence” (emphasis added) (General Principle (3)). The ICC Rules require disclosure of facts or circumstances which “might be of such a nature as to call into question the arbitrator’s independence in the eyes of the parties as well as any circumstances that could give rise to reasonable doubts as to the arbitrator’s impartiality” (emphasis added) (Article 11). The LCIA Rules require disclosure of “circumstances currently known to the candidate which are likely to give rise in the mind of any party to any justifiable doubts as to his or her impartiality or independence” (emphasis added) (Article 5.4).

Whilst this may reflect good practice in international commercial arbitration, the authorities make it clear that the more certain standards of an objective observer apply to the issue of disclosure under English law.

We note that the LCIA Rules make it clear that disclosure is only required of facts or circumstances known to the arbitrator (as do the IBA Guidelines in the Explanation to General Standard (3) at (d)). We agree with that approach. You can only disclose what you know and there is no duty of inquiry.

The disclosure required depends on what the arbitrator knows. The fact that disclosure is required of circumstances that might lead to a conclusion of apparent bias, emphasises that the question of what is to be disclosed is to be considered prospectively. The question of whether or not disclosure should be made, or should have been made, depends on the prevailing circumstances at that time. A decision as to disclosure based on a conclusion which might be drawn can only be made on the basis of the circumstances as they were then known to be and, in principle, a determination of whether or not such disclosure should have been made should similarly be so judged. In this regard, we disagree with the approach of the judge who considered that the issue of whether disclosure ought to have been made was to be determined retrospectively by reference to whether, having regard to all matters known at the later stage to the fair-minded and informed observer, the circumstance would lead to the conclusion that there was a real possibility of bias.

In summary, we consider the present position under English law to be that disclosure should be given of facts and circumstances known to the arbitrator which, in the language of section 24 of the Act, would or might give rise to justifiable doubts as to his impartiality. Under English law this means facts or circumstances which would or might lead the fair-minded and informed observer, having considered the facts, to conclude that there was a real possibility that the arbitrator was biased.

The implications of this conclusion in relation to disclosure of appointments concerning overlapping subject matter but only one common party will be considered when addressing the facts of the present case.

There are, as it seems to us, two distinct questions for the court considering an allegation of non-disclosure after the event. First, the court needs to consider whether disclosure ought to have been made in accordance with the principles we have just enunciated. Secondly, the court needs to consider the significance of that non-disclosure in the context of the application with which the court is dealing. In the case of an application for removal of the arbitrator in question, the court will consider on the basis of all the factual information available when that application is heard (including the fact that there has been non-disclosure), whether the fair-minded and informed observer would conclude that there was a real possibility that the arbitrator was biased.

If a disclosure that ought to have been made has not been made, that will mean that the arbitrator will not have displayed the “badge of impartiality” which he should have done. As Lord Bingham observed in the Davidson case: the fact of non-disclosure “must inevitably colour the thinking of the observer”.

Non-disclosure is therefore a factor to be taken into account in considering the issue of apparent bias. An inappropriate response to the suggestion that there should be or should have been disclosure may further colour the thinking of the observer and may fortify or even lead to an overall conclusion of apparent bias – see, for example, Paice v Harding [2015] EWHC 661, [2015] BLR 345 and Cofely Ltd v Bingham [2016] EWHC 240.

Non-disclosure of a fact or circumstance which should have been disclosed, but does not in fact, on examination, give rise to justifiable doubts as to the arbitrator’s impartiality, cannot, however, in and of itself justify an inference of apparent bias. Something more is required – see, for example, the comments of Lord Mance in Helow v Home Secretary at [58].

For reasons already given, viewed objectively, we do not consider that the mere fact of an appointment in a related reference with only one common party would in and of itself justify an inference of apparent bias. We accept that M’s acceptance of a closely related appointment also involving Chubb may giverise to legitimate concerns in the eyes of Halliburton, and that these might have been alleviated by disclosure. However, as has already been explained, in order to lead to the objective conclusion of apparent bias something more would be required, and that must be “something of substance”.

In this case, Halliburton relies on a number of aspects of M’s behaviour above and beyond the mere fact of his acceptance of a related appointment to which Chubb was the only common party. The main points are as follows:

More generally, Halliburton relied on the views expressed by the third arbitrator, N, “an international arbitrator of great experience and eminence”, as a proxy for the views of the fair-minded and informed observer.

M was appointed by the High Court in accordance with the agreed contractual procedure set out in the arbitration agreement. Having been so appointed, the appropriateness of his appointment is not open to question. Moreover, Halliburton’s main objection to Chubb’s candidates, namely that they were English lawyers, was generic and did not relate to M personally or to issues of a possible lack of impartiality.

For reasons already given, the mere fact of overlap does not give rise to justifiable doubts as to impartiality. In any event, the degree of overlap in this case was in fact very limited. References 2 and 3 were decided on the basis of a preliminary issue of law that did not arise in reference 1. The facts never had to be investigated. Further, although there was a similarity in the further issues which would otherwise have arisen (reasonableness of settlement and withholding of consent), they would have fallen for consideration in a very different factual context. In particular, in Halliburton’s case the context was before and in anticipation of the court judgment by the federal court of the Eastern District of Louisiana, whilst in Transocean’s case it was after and in knowledge of the judgment.

We do not consider this to be a matter of significance, essentially for the reasons given by the judge at [20]-[21] of his judgment. In essence, the argument goes too far and would mean that a remuneration benefit which an arbitrator receives from his appointing party (even indirectly) is a disqualifying benefit. If that were so it would equally apply to party-appointed arbitrators in a single arbitration and would be wholly inconsistent with the manner in which commercial arbitration is routinely conducted. The alleged “secrecy” of the benefit adds nothing. Either the benefit is disqualifying or it is not. If it is, then objection could be made to every party-appointed arbitrator, which would be absurd.

The judge approached the issue of disclosure with the benefit of hindsight. Having decided that the appointment of M did not give rise to justifiable doubts as to his impartiality, he considered that it followed that there were no circumstances to disclose. For reasons already given, we agree with Halliburton that the issue of whether disclosure ought to have been made has to be addressed at the time that disclosure potentially falls to be made and not with the benefit of hindsight. In this case, it should be considered at the point M took the decision to accept the Transocean appointment.

The question to be asked is whether there were facts or circumstances known to M which would or might lead the fair-minded and informed observer, having considered the facts, to conclude that there was a real possibility that M was biased. If so, M was under an obligation to disclose them.

The relevant facts and circumstances are M’s proposed appointment in related arbitrations and in particular reference 2, which involved Chubb alone as a common party. Halliburton submits that this should have been disclosed because of the legitimate concerns to which such appointments give rise, as addressed above.

As explained above, the starting point is as stated by the judge, namely that an arbitrator should be trusted to decide the case solely on the evidence or other material adduced in the proceedings in question. For reasons already given, viewed objectively, we do not consider that the fact of such an appointment would or might, in and of itself, justify an inference of apparent bias such as to give rise to an obligation to make disclosure.

We recognise, however, as Leggatt J did in the Guidant case, that in the eyes of the parties there might be cause for legitimate concern, at least in a case of substantial material overlap. From Halliburton’s perspective, such disclosure would enable it to identify potential problems and inform the arbitrator of the significance of those problems, to play a role in policing the arbitrator’s duty of impartiality, and to propose practical measures so as to prevent any potential disadvantage arising.

In the context of an international commercial arbitration, we accept that it is good practice for disclosure to be given in the light of party concerns such as this. Indeed, M himself recognised that disclosure would have been “prudent”. He also stressed the fact that, as matters had turned out, he had learned no factual information that would be unknown to his co-arbitrators in the Haliburton reference, thereby implicitly recognising the relevance of unknown information. Further, in so far as the IBA Guidelines reflect international commercial arbitration practice, it is to be noted that the present case may be said to fall within the IBA Guideline Orange List 3.1.5., which calls for disclosure where an arbitrator serves in an arbitration on a related issue involving one of the parties.

We would therefore approach the issue of whether disclosure ought to have been made on the basis that in the context of international commercial arbitration, as a matter of good practice, such disclosure should have been made. It is then necessary to consider whether, considered at the time of the appointment, disclosure ought, as a matter of law to have been made. We bear in mind that we consider this question from an objective standpoint, even though the objective observer would realise that a party to whom disclosure was not made might reasonably have the concerns we have alluded to.

It was telling that counsel for Chubb accepted that 10 appointments for one party might objectively give rise to justifiable doubts as to the impartiality of the arbitrator. This was not such a case, but it was a case where M had accepted a number of appointments by Chubb or in cases involving Chubb. Chubb stressed that this had not been the basis of Halliburton’s objection to M, but the test to be applied is an objective one. Moreover, given that M had at the outset disclosed his appointments in other Chubb arbitrations, the natural expectation of the fair-minded and informed observer would be that he would do likewise in relation to any further proposed Chubb appointment. Since the question at the time that disclosure ought to have been contemplated is not whether the fact would have provided the basis for a reasonable apprehension of lack of impartiality, but whether it might have provided such a basis, this concession is, we think, important.

In our judgment, the fact that best practice in international commercial arbitration would have required disclosure of the other appointments, taken together with the clear possibility that other factors, such as the actual degree of overlap about which M knew little at the time and the nature of other connections, might have been argued to combine together to give the fair-minded and informed observer a basis for a reasonable apprehension of lack of impartiality. In these factual circumstances, we consider that disclosure ought, as a matter of law, to have been made. In so far as that impacts on the arbitrator’s duty of confidentiality in relation to the other arbitration, it must be regarded as being an exception to that duty, a duty which is recognised not to be absolute. The enquiry is obviously fact-intensive, but we take the view that M’s own instincts were the right ones. He recognised that it would have been “prudent” to make disclosure. His innocent oversight in failing to do so is not something for which he can be blamed, but it cannot also excuse non-disclosure of a fact that ought properly to have been disclosed on the basis of the principles we have explained.

M dealt with Halliburton’s concerns appropriately for all the detailed reasons given by the judge. In particular, there is no justification for criticising M for making his offer to resign conditional on the consent of Chubb. Having reached the conclusion in good faith that Halliburton’s complaints did not require him to resign, in circumstances where Chubb wished him to remain in office, it was appropriate for him to conclude that he should continue to discharge his responsibilities until the parties agreed or the court ordered otherwise. In doing so, he did not align his interests with Chubb; even if he had, this would have been limited to the preliminary issue in references 2 and 3 and would have had no impact on reference 1.

The court has to form its own judgment on the conclusion which the fair minded and informed observer would draw, on the basis of the evidence and argument presented to it, applying English law. Moreover, it is to be noted that although N has throughout been a party to these proceedings, he never expressed any concerns about the procedural fairness of the arbitration until the Award was issued, and did not dissent from the substantive decision in the Award.

We conclude that M ought as a matter of good practice and, in the circumstances of this case, as a matter of law to have made disclosure to Halliburton at the time of his appointments in references 2 and 3.

The question then becomes whether, at the time of the hearing to remove, the non-disclosure taken together with any other relevant factors would have led the fair-minded and informed observer, having considered the facts, to conclude that there was in fact a real possibility that M was biased.

In answering this question we would in particular take the following factors into account from the perspective of the fair-minded and informed observer: (1) the non-disclosed circumstance does not in itself justify an inference of apparent bias; (2) disclosure ought to have been made, but the omission was accidental rather than deliberate; (3) the very limited degree of overlap means that this is not a case where overlapping issues should give rise to any significant concerns; (4) the fair-minded and informed observer would not consider that mere oversight in such circumstances would give rise to justifiable doubts as to impartiality; and (5) there is no substance in Halliburton’s criticisms of M’s conduct after the non-disclosure was challenged or in the other heads of complaint raised by them.

Halliburton suggested that M’s oversight in relation to disclosure might be indicative of unconscious bias. In the Almazeedi case the Privy Council cited with approval at [19] a passage from Lord Bingham’s judgment in the case of Prince Jefri Bolkiah v State of Brunei Darussalam (No 3) [2007] UKPC 62, [2008] 2 LRC 196 at [18] where he said (quoting from the Court of Appeal of Brunei):

Although the Privy Council stated that they did not “take this passage to mean that the fair-minded and informed observer would discount the risk of unconscious bias in all situations”, the Prince Jefri case does illustrate that relevant experience is material to the risk of such bias. In this case, as the judge observed, M is a “well known and highly respected international arbitrator” with very extensive experience as an arbitrator.

In addition, the fact that Halliburton’s concerns had been fully explained and expressed before any substantive consideration of the case by M makes such bias unlikely. M would thereby have been made conscious of the matters of which it is suggested he might otherwise have been unconscious. In the light of the objections raised, M would be likely to have done all he could to ensure that nothing that transpired in references 2 and 3 influenced in any way his approach in reference 1. In any event, due to the determination of the preliminary issue in references 2 and 3 in favour of Chubb, there was in fact no substantive overlap between these references and reference 1.

For all these reasons, having carefully considered both individually and collectively all the arguments advanced by Halliburton (orally and in writing), we agree with the judge’s overall conclusion that the fair-minded and informed observer, having considered the facts, would not conclude that there was a real possibility that M was biased.

For the reasons outlined above, we dismiss the appeal.