Authored By: Aliya Khan
Newcastle University
A Case Summary of Halliburton Company v Chubb Bermuda Insurance Ltd
Halliburton Company v Chubb Bermuda Insurance Ltd (formerly known as Ace Bermuda Insurance Ltd) [2020] UKSC 48, [2021] AC 1083
Case citation and basic information
The appeal was decided by the Supreme Court of the United Kingdom on 27 November 2020. Lord Reed PSC, Lord Hodge, Lady Black, Lord Lloyd-Jones and Lady Arden sat. Lord Hodge gave the principal judgment, with which Lord Reed, Lady Black and Lord Lloyd-Jones agreed; Lady Arden gave a separate concurring judgment.[1]
Introduction
The case asked how English law should police apparent bias where one arbitrator accepts appointments in overlapping arbitrations but only one party is common to them. Its importance lies less in the result than in the architecture of the Supreme Court’s answer. The Court confirmed that every arbitrator owes a legal duty to disclose circumstances which would or might reasonably cause the fair-minded and informed observer to apprehend bias, while holding that breach does not automatically justify removal.[2] That distinction preserves disclosure as an early safeguard without turning it into an admission of partiality. It also exposes the central tension in private adjudication: specialist repeat appointments may be commercially valuable, yet secrecy can give the repeat party an informational and strategic advantage.
Material facts and procedural history
The dispute arose from the 2010 Deepwater Horizon disaster. BP leased the rig; Transocean owned and operated it; Halliburton supplied cementing and well-monitoring services. US proceedings apportioned blame 67 per cent to BP, 30 per cent to Transocean and 3 per cent to Halliburton, which had already settled related claims for about US$1.1 billion.[3] Chubb refused Halliburton’s insurance claim, disputing the reasonableness of the settlement and alleging that consent had been withheld. Their Bermuda Form policy required ad hoc London arbitration before three arbitrators, applied New York substantive law and excluded appeal on questions of law.[4]
After the party nominees failed to agree a chair, Flaux J appointed Kenneth Rokison QC in June 2015; Chubb had proposed him, and he disclosed earlier Chubb appointments.[5] In December 2015 he accepted Chubb’s appointment in Transocean’s related claim against Chubb, and in August 2016 a joint nomination in Transocean’s claim against another insurer. He disclosed Halliburton’s reference to Transocean but neither later appointment to Halliburton.[6] When Halliburton discovered them in November 2016, Rokison explained that the omission was an oversight, accepted that disclosure would have been prudent, and offered to resign from the later references if preliminary policy-construction rulings did not end them.[7]
Halliburton sought removal under Arbitration Act 1996, s 24(1)(a). Popplewell J dismissed the application, treating repeat appointments as common and often desirable and finding neither apparent bias nor a duty to disclose.[8] The Court of Appeal also dismissed the challenge. It held prospectively that disclosure should have occurred, but required “something more” of substance beyond overlapping appointments and treated the accidental omission, limited overlap and Rokison’s response as insufficient for removal.[9] Halliburton appealed to the Supreme Court; meanwhile, the later references ended on preliminary issues and the Halliburton tribunal issued a partial award for Chubb.
Legal issues
Two linked issues governed the appeal. First, could accepting multiple appointments in related references with one common party create apparent bias, judged by whether the fair-minded and informed observer would see a real possibility of bias?[10] Secondly, did English law impose a duty to disclose those appointments and, if so, what was the effect of non-disclosure? Resolving them required the Court to connect an arbitrator’s s 33 duty of impartiality with s 24 removal, while accommodating privacy, confidentiality and divergent trade practices.
Parties’ arguments and interventions
Halliburton did not allege actual or deliberate bias. It argued that Chubb’s repeat appointment gave Chubb privileged access: it could test submissions before the same decision-maker, observe his reactions and know about both proceedings, while Halliburton could neither object nor manage the overlap. Non-disclosure deepened that inequality and showed insufficient regard for procedural fairness.[11] Chubb answered that disclosure was legally required only where the known facts already satisfied the objective apparent-bias standard; anything wider was good practice, and innocent non-disclosure was merely one contextual factor.[12]
The interventions demonstrated why a universal numerical rule would misfire. The ICC, LCIA and CIArb emphasised international expectations of proactive disclosure and warned that non-disclosure may itself create justifiable doubts. GAFTA, the LMAA and ARIAS (UK) described specialist markets in which repeat appointments are routine, sometimes necessary, and understood by users.[13] Those competing accounts directed the Court towards a context-sensitive rule rather than either automatic disqualification or blanket tolerance.
The Court’s reasoning
Lord Hodge began from Porter v Magill: the question is whether the fair-minded and informed observer, having considered the facts, would conclude that there was a real possibility of bias.[14] The substantive standard is the same for judges and arbitrators, and for chairs and party-appointed arbitrators. Application nevertheless differs because arbitration is private, awards receive limited appellate scrutiny, arbitrators are remunerated and may depend on future appointments, and participants may bring different legal traditions to the role. These features place a “premium” on disclosure, while the observer remains neither complacent nor unduly suspicious and must recognise the risk of tactical challenges.[15]
From s 33 and the arbitrator’s contract, the Court derived a legal duty to disclose facts or circumstances which would or might reasonably cause the objective observer to apprehend bias.[16] Disclosure is therefore wider than disqualification: a matter may warrant transparency before its significance can be tested without ultimately proving a real possibility of bias. Institutional rules and the 2014 IBA Guidelines were influential evidence of practice, not sources of English law; importantly, their party-centred disclosure standard is not identical to the English objective test.[17] Non-disclosure is likewise evidence, not a self-executing sanction. An innocent omission may be explained; a deliberate or unexplained one may colour the observer’s assessment.[18]
Multiple appointments consequently admit no abstract answer. They may be objectionable because the common party alone can compare evidence, submissions and the arbitrator’s responses across references. Yet whether disclosure is required depends on the arbitration’s customs and the parties’ expectations at the relevant time.[19] The Court declined to make the Court of Appeal’s “something more” formulation a universal precondition. In Bermuda Form arbitration there was no established practice dispensing with disclosure, and the structural imbalance between a repeat insurer and a one-off insured made transparency necessary. Rokison therefore breached his duty when he failed to disclose the Chubb–Transocean appointment.[20]
Confidentiality did not defeat that conclusion. Consent to limited disclosure may be express or inferred from the arbitration agreement and relevant custom. Ordinarily, it is enough to identify the common party, the fact of the appointment and the shared incident or subject matter; more detail requires consent. If necessary disclosure cannot lawfully be made, the arbitrator should decline the later appointment.[21] Lady Arden preferred to characterise disclosure as a secondary obligation supporting the primary duty of impartiality, and stressed that breach remains a contractual wrong even where removal or damages do not follow.[22]
The decisive step was temporal. Whether disclosure was required is assessed when the duty arose; whether circumstances “exist” for removal under s 24 is assessed at the later court hearing.[23] By January 2017 the observer would know that Rokison’s omission was genuine, the law had been uncertain, the references’ overlap was limited, preliminary rulings were likely to end the later cases, and he had offered to resign if factual overlap emerged. There was no secret financial benefit or hostile response. Those facts neutralised the inference that non-disclosure might otherwise support. The duty had been breached, but no real possibility of bias existed at the removal hearing.[24]
Judgment, ratio and outcome
The Supreme Court unanimously dismissed the appeal. The narrow ratio is that English law requires an arbitrator to disclose matters which would or might reasonably lead the fair-minded and informed observer to apprehend a real possibility of bias; related repeat appointments may trigger that duty depending on custom and context; and breach is relevant but neither necessary nor sufficient for removal under s 24(1)(a).[25] Rokison should have disclosed the later Chubb appointment, yet the full circumstances at the hearing did not justify doubts about his impartiality. The extensive observations on particular specialist sectors are best treated as fact-sensitive guidance, not categorical exemptions.
Critical analysis and significance
The judgment’s strongest feature is its separation of prevention from remedy. A lower disclosure threshold lets parties investigate and waive concerns early, while the stricter removal test protects finality and discourages weaponised challenges. As Redfern and Hunter recognises, disclosure is not formalism: late challenges create replacement costs and may imperil enforcement.[26] The price is remedial fragility. Because later facts can cure the appearance that existed when disclosure was due, an arbitrator may breach a real legal duty yet face no practical consequence. Lady Arden’s insistence that contractual breach still matters partly answers that objection, but the judgment leaves sanctions underdeveloped.
Its reliance on sectoral custom is both commercially intelligent and normatively risky. Small maritime or commodities pools cannot operate as if expertise were unlimited. But custom is often generated by repeat players; a one-off insured may neither know nor meaningfully accept it. Courts should therefore demand a well-established, mutually intelligible practice before inferring consent to non-disclosure. The 2024 IBA Guidelines move in that direction: they preserve special treatment for fields with small pools, but say disclosure may still be desirable, while maintaining that non-disclosure alone does not establish a conflict.[27]
The later legal development confirms Halliburton’s durability. The Arbitration Act 2025 inserted s 23A into the 1996 Act, in force from 1 August 2025. It requires a prospective arbitrator, before accepting appointment, and an arbitrator throughout proceedings, to disclose relevant circumstances of which they are aware or ought reasonably to be aware. The reform codifies Halliburton’s central insight and strengthens it by expressly extending the duty to constructive awareness and the pre-appointment stage. It nevertheless leaves the content of relevance, sectoral practice and the consequences of breach to be worked out contextually. The 2024 IBA Guidelines similarly retain a flexible, fact-sensitive disclosure framework; they are persuasive international guidance, not an override of English law.[28][29][30]
Conclusion
Halliburton establishes transparency as part of, not a substitute for, impartial adjudication. Its calibrated rule respects specialist arbitration while confronting the informational asymmetry created by undisclosed overlapping appointments. The Court was right to find a breach yet reject removal on the later facts: otherwise disclosure would collapse into disqualification. The continuing challenge is to prevent “custom” from becoming repeat-player privilege. Mandatory s 23A now supplies a clearer statutory foundation, but Halliburton’s objective observer, temporal distinction and insistence on contextual judgment remain the controlling intellectual framework.
Bibliography
Cases
Halliburton Co v Chubb Bermuda Insurance Ltd [2017] EWHC 137 (Comm), [2017] 1 WLR 2280
Halliburton Co v Chubb Bermuda Insurance Ltd [2018] EWCA Civ 817, [2018] 1 WLR 3361
Halliburton Company v Chubb Bermuda Insurance Ltd (formerly known as Ace Bermuda Insurance Ltd) [2020] UKSC 48, [2021] AC 1083
Helow v Secretary of State for the Home Department [2008] UKHL 62, [2008] 1 WLR 2416
Porter v Magill [2001] UKHL 67, [2002] 2 AC 357
Legislation
Arbitration Act 1996
Arbitration Act 2025
Arbitration Act 2025 (Commencement) Regulations 2025, SI 2025/905
Books
Blackaby N, Partasides C and Redfern A, Redfern and Hunter on International Arbitration (7th edn, OUP 2022)
Official materials and institutional guidance
International Bar Association, IBA Guidelines on Conflicts of Interest in International Arbitration (2014)
International Bar Association, IBA Guidelines on Conflicts of Interest in International Arbitration (2024)
Law Commission, Review of the Arbitration Act 1996: Final Report and Bill (Law Com No 413, 2023)
Word count (main text): 1,659 — excluding the title, case citation, basic-information component, headings, footnotes and bibliography.
[1]Halliburton Company v Chubb Bermuda Insurance Ltd (formerly known as Ace Bermuda Insurance Ltd) [2020] UKSC 48, [2021] AC 1083, title page and [1] (Lord Hodge).
[2] Halliburton Company v Chubb Bermuda Insurance Ltd (formerly known as Ace Bermuda Insurance Ltd) [2020] UKSC 48, [2021] AC 1083 [3]–[5], [151]–[157] (Lord Hodge) (‘Halliburton (UKSC)’).
[3] Halliburton (UKSC) [7]–[9].
[4] ibid [10]–[11].
[5] ibid [12]–[13].
[6] ibid [15]–[18].
[7] ibid [19]–[21].
[8] Halliburton Co v Chubb Bermuda Insurance Ltd [2017] EWHC 137 (Comm), [2017] 1 WLR 2280 [16], [23], [55]–[56] (Popplewell J).
[9] Halliburton Co v Chubb Bermuda Insurance Ltd [2018] EWCA Civ 817, [2018] 1 WLR 3361 [71], [77]–[79]; Halliburton (UKSC) [33]–[39].
[10] Halliburton (UKSC) [2]–[3], [52]; Porter v Magill [2001] UKHL 67, [2002] 2 AC 357 [103].
[11] Halliburton (UKSC) [41], [61].
[12] ibid [46].
[13] ibid [42]–[45].
[14] Porter (n 9) [103]; Halliburton (UKSC) [52].
[15] Halliburton (UKSC) [55]–[69]; Helow v Secretary of State for the Home Department [2008] UKHL 62, [2008] 1 WLR 2416 [1]–[3].
[16] Arbitration Act 1996, ss 24(1)(a), 33; Halliburton (UKSC) [74]–[81], [107]–[116].
[17] Halliburton (UKSC) [70]–[72]; International Bar Association, IBA Guidelines on Conflicts of Interest in International Arbitration (2014) General Standards 2–3.
[18] Halliburton (UKSC) [73], [117]–[118].
[19] ibid [61], [125]–[136].
[20] ibid [130], [137]–[138], [145]–[147].
[21] ibid [82]–[105], [146].
[22] ibid [159]–[169], [173]–[188] (Lady Arden).
[23] ibid [119]–[123].
[24] ibid [148]–[150].
[25] ibid [151]–[158].
[26] Nigel Blackaby KC, Constantine Partasides KC and Alan Redfern, Redfern and Hunter on International Arbitration (7th edn, OUP 2022) paras 4.69–4.72, 4.88–4.101.
[27] International Bar Association, IBA Guidelines on Conflicts of Interest in International Arbitration (2024) General Standard 3(g), Orange List 3.1.3 and 3.1.5, n 3 <https://www.ibanet.org/document?id=guidelines-on-conflicts-of-interest-in-international-arbitration-2024> accessed 10 August 2026.
[28] Arbitration Act 1996, s 23A, inserted by Arbitration Act 2025, s 2; Arbitration Act 2025 (Commencement) Regulations 2025, SI 2025/905, reg 2.
[29]Law Commission, Review of the Arbitration Act 1996: Final Report and Bill (Law Com No 413, 2023) paras 3.64–3.72, 3.95–3.99; Arbitration Act 2025, s 2.
