Independence and Apparent Bias: Judicial and Arbitral Confidence: Halliburton v Chubb

The Goff Lecture 2022
Sydney and Hong Kong 

 Chief Justice Allsop AO^[1]  3 November 2022

1 Commerce is built on trust. That trust is not incautious wishful thinking. It is the requisite degree of trust in the circumstances, usually circumscribed by practical reality, including the mutual advantage of a continuing relationship, a body of information, and a degree of familiarity with the context.

2 Just as commerce is built on the requisite degree of trust, so it is important that the agency of adjudication, which is chosen to resolve disputes that cannot be resolved by mutually agreeable compromise be able to engender the requisite degree of trust in the parties, including as to honesty and fairness, and most importantly, impartiality.

3 By using the phrase “requisite degree of” I am not intending to water down the important underpinning place of honesty, fairness, impartiality, independence and trust. What I intend is to recognise that definitions, rules and moral absolutes have their limitations, along with their importance, in the real world. Dispute resolution, including adjudication, takes place in the real world with imperfect institutions and humans, acting in good faith.

4 The subject of this evening’s discussion is the decision of the United Kingdom Supreme Court in late 2020 in Halliburton Co v Chubb Bermuda Insurance Ltd (Halliburton).^[2] It contains important development and clarification of the common law of England on arbitration, recognising and incorporating real world realities in practice, but holding firm to a demand for the highest level of impartiality in arbitrations governed by English law.

5 It is first necessary to say something of the differences and similarities of judicial adjudication and arbitration to appreciate the existence and content of the legal duty of disclosure of the arbitrator or prospective arbitrator, and the duties of independence and impartiality of the judge and arbitrator.

6 The fundamental underpinning feature of judicial adjudication is the imposition of, or subjection of the disputant parties to, a form of state power: judicial power. Judicial resolution of a controversy is the state using its compulsory powers to resolve disputes brought to the jurisdiction^[3] of its courts by one party. Usually, this involves full public disclosure of all issues, and an appellate structure of one, possibly two, layers. The moving or claiming party in curial adjudication may have a choice of courts, but only within recognised limits of conflict of laws. For the responding party the venue is compulsory. Once in the hands of the court, the structure and application of procedure is by reference to the law of the relevant polity. Part of that imposed structure and law will be the institutional choice and imposition of the judge and an accompanying expectation of complete independence and impartiality.

7 The arbitral process, of its nature, contains within it a greater degree of choice: Arbitrators of chosen identity and known skill can be provided for; and trusted institutions, with known and understood rules, and seat courts of known skill can be chosen to supervise the resolution of the dispute.

8 The two concepts of independence and impartiality are not easily disentangled by verbal formula. It is doubtful whether the attempt to define them separately is worthwhile,^[4] other than by a recognition of their core meanings: impartiality is an absence of bias or of partiality to one side or its case; independence is a lack of control or of influence from relevant external sources.^[5] An exploration of circumstances that may lead to one being undermined or compromised generally assists in the consideration of whether the other is as well. The interconnection or relationship between the two concepts is not assisted by definition or seeking exhaustively to set out what may lead to, or be seen to lead to, an undermining of independence or a compromising of impartiality, although guidelines help.^[6] But guidelines often, over time, become seen as rules, with defined content, something unhelpful in a context-specific, fact-rich evaluation.^[7] It can be said that, ultimately, it is impartiality that matters. That was the view of the English Departmental Advisory Committee on Arbitration Law leading to the drafting of the Arbitration Act 1996^[8] and remains the, at least provisional, view of the Law Commission in its current review of the Act.^[9]

9 Definition of ethical values, principles and states of affairs underpinning trust and confidence, and honesty and fairness, risks arid debate and unhelpful distinctions, which likely pay insufficient regard to the importance of context and the necessary indeterminacy of meaning. This is not to promote uncertainty; rather it is to recognise the central place of the implicit and the contextual in the requisite degree of certainty in the application of principles of a behavioural or ethical kind.

10 That is why the conjoined expression “independence and impartiality” is helpful.^[10] The expression may perhaps be seen as, or as similar to, a hendiadys: a figure of speech in which a single complex idea is expressed by two words connected with “and” without subordination of one to the other or without parts of separate and equal standing.^[11]

11 The importance of the trust and confidence of the members of the polity served by the court, and of litigants who come before the court, is basal. This leads to the need to express the legal rule by reference to which a judge must decide that he or she cannot be independent or impartial, both by reference to the fact of independence and impartiality, and of their appearance externally. The fact of impartiality involves an examination of conscience. That may involve the contemplation of personal circumstance and attitude and an exploration of the possibility of subconscious bias. Of greater subtlety and practical importance is the rule or test for “apparent bias”. This requires the expression in words of an evaluative assessment: necessarily by reference to the practical human context in which the question is being posed. Hence, the necessary anthropomorphous (of human form) nature of relevant tests: Trust and confidence are human and emotional conceptions, linked to rational thought. Thus the test is best couched in human terms.

12 In English law the test for apparent bias in decision-making was settled in 2002 and 2008, in Porter v Magill and Helow v Secretary of State for the Home Department: whether a “real possibility” exists that a decision-maker is biased. A “real possibility” exists where the circumstances would lead a fair-minded and informed observer to conclude that there was a real possibility that the decision-maker was biased.^[12] This replaced (or modified) Lord Goff’s expressions “real danger” and “real likelihood” in the sense explained by Lord Goff in R v Gough^[13] as in terms of possibility not probability, which had been criticised by the High Court of Australia as being too stringent, over-emphasising the court’s view and under-emphasising the public perception of the circumstances, and not followed in Canada and New Zealand.^[14]

13 In Porter v Magill, Lord Hope described as a “modest adjustment”^[15] the introduction of the fair-minded and informed observer concluding that there was a real possibility of bias.^[16] Given Lord Goff’s explanation^[17] that the court’s perspective was the reasonable man observing and that the perception was one in terms of possibility not probability, and given the view of Lord Phillips of Worth Matravers MR in In re Medicaments^[18] that “real danger” was the same as “real possibility” (thereby removing the connotation of immediacy as in “clear and present danger”), that was a fair comment by Lord Hope, even though his Lordship discarded the phrase “real danger” as not useful.

14 In Australia, the reasonable apprehension test, at least for non-arbitral bias, is whether “a fair-minded lay observer might reasonably apprehend that the judge might not bring an impartial mind to the resolution of the question the judge is required to decide.”^[19]

15 A different formulation applies, however, to arbitration under both Commonwealth and State and Territory arbitration Acts. The relevant provisions give content to Article 12 of the Model Law (“justifiable doubts as to [the arbitrator’s] impartiality or independence”) by such being satisfied “only if there is a real danger of bias”.^[20]

16 This statutory formulation, taken from R v Gough, was advocated strongly by those who saw the need for a higher threshold in adjudicators who were not public officials.^[21]

17 Neither the English nor the Australian non-arbitral test requires certainty, only a “real possibility” or “reasonable apprehension” of bias. These standards support the fundamental principle that justice must be both done and be seen to be done. The qualifiers “real” and “reasonable” import the notion that the human construct must consider that the potential for bias is material, that it logically or experientially arises from the surrounding circumstances, for a judge to be disqualified from hearing a matter. As to the nature of the human subject, in the English test, the fair-minded observer is an independent person, not a person in the position of a litigant before the judge.^[22] Further, the fair-minded English observer appreciates the context of the material before them, and reserves judgment until they have fully understood both sides of the argument.^[23] In the Australian non-arbitral test, the fair-minded observer is a lay person, who is not unduly sensitive or suspicious, and who has an understanding of the relevant circumstances of the case and ordinary judicial practice.^[24]

18 The need for this degree of detachment from the special knowledge and expertise of judges and lawyers reflects the constitutional importance of the judicial process as the open and public exercise of state power. As will be seen in Lord Hodge’s judgment in Halliburton the context of the arbitral process perhaps lessens the need for emphasis on public confidence, but heightens the need for emphasis on participant confidence and general confidence in the system of arbitration, as part of an international system of justice.^[25]

19 The institutional integrity of a court should not be able to be impugned too lightly. At the same time, the challenge should not be so onerous as to permit judges to adjudicate in circumstances where a reasonable human perception would call that into question.

20 The textual expressions of the test in England and Australia reflect the difficulty of this balance. There is a lack of equivocation in the English test: there is a real possibility of bias, because the circumstances would lead to the fair-minded and informed observer to that conclusion. The non-arbitral test in Australia may perhaps be seen as more attenuated: the observer might reasonably apprehend that the judge might not bring an impartial or unprejudiced mind to bear. Though not to be viewed in a mathematical way, two “mights”, even if “firmly established” as they have to be^[26] seem less stringent and more open to equivocation than one “would” leading to a “real possibility”.

21 All this said, one needs to be careful of over-precise linguistic expression when the proposition involves the implicit, the less than certain, the intuitive and human feeling, albeit derived from rational thought. The underlying notion is elegantly expressed in the Model Law: Are there justifiable doubts about independence or impartiality? That is a factual question that is to be examined contextually. The two contexts of a public judicial institution, and of a private panel of people partly chosen by the parties are very different. That difference does not necessarily entail the need for a differently framed question; but, rather, the recognition of a different factual context in which to answer it.

22 So, is the arbitral test in Australia higher than that for Australian judges and other public decision-makers? Those advocating for, and making, the change certainly thought so, as a way of stopping what was seen as an increasingly common procedural tactic in high value international commercial arbitrations.^[27] The Explanatory Memorandum to the Act amending the Commonwealth Act expressly referred to the decision of Lord Goff in Gough, as a different (and more stringent) standard to that derived from cases such as Ebner.^[28] Yet, as Lord Goff himself made clear, the test of danger or likelihood was one by reference to possibility, not probability. Australia’s arbitral test is clearly linked to R v Gough. The real question is whether it will be construed as modestly adjusted and explicated in and by Porter v Magill, Helow v Secretary and In re Medicaments.

23 Debates about refinements of degrees of possibility are likely to be arid and unhelpful. This is why Lord Hodge’s judgment, with its emphasis on the arbitration context, rather than textual comparisons, is very valuable.

24 It is important to recall that the apprehension is one of bias or partiality or pre-judgment^[29]: not of loss of the adjudication. Litigants often fail to grasp this. The fact that a judge has decided a legal point before is not, of itself, a ground of objection. The apprehension must be of not being able to bring an open mind to bear on the arguments, not that the judge is more likely to follow precedent perceived to be unfavourable.

25 But some kinds of prior decisions can reasonably be seen by the hypothesised human construct to be more difficult to eradicate, consciously or subconsciously, from even the trained judicial mind. An earlier finding of fraud or dishonesty may be impossible to remove from the mind.^[30]

26 Such prejudicial earlier findings raise the issue of unconscious bias.^[31] There is a growing body of literature which, drawing on modern psychology, addresses the impact of subconscious forces upon the judicial (and arbitral) mind.^[32] This includes dealing with the operation of confirmation bias and the role of intuitive first perceptions and ordinary human prejudices.^[33] The risk of unconscious bias may rise to the level of apparent bias.^[34]

27 Related to any formulation of apparent bias is the question of disclosure of matters to parties. Many judges feel more comfortable in approaching the question of disclosure broadly: disclosing matters of connection that may not justify recusal; but they do so to ensure confidence and trust in the parties of the process, given that one party will always lose. It is part of bolstering the institution, rather than complying with any legal duty expressed as an incident of judicial power. In Australia, the High Court has said^[35] that where a judge is unclear whether he or she should recuse him or herself, disclosure of the circumstances should be made as a matter of prudence and professional practice, saying that it was neither useful nor necessary to describe this practice in terms of rights and duties. The place of disclosure and, for arbitrators, the legal duty attending it, was at the centre of Halliburton.

28 Legislation governing arbitration, model laws, and arbitral rules generally express independence and impartiality as core tenets of the arbitral process,^[36] variously, but similarly, dealing with the test for them and with questions of disclosure.^[37]

29 It is to be noted that some of the tests relating to disclosure are explicitly directed to what would be thought by the parties.^[38] The reference to the perspective of a participating party is a distinct feature of these tests, as compared with tests for judicial bias and may be appropriate in the context of arbitrations where considerations of the exercise of public power and the interest of the public at large are absent, and the confidence of the parties is paramount.

30 The reality of the place of party autonomy in choice and the role of the party-appointed arbitrator should be confronted. The different institutional contexts of curial and arbitral adjudication are, again, important. In the case of a litigant before a court, a judge is allocated to a case because the state has identified a particular judicial resource is suitable. The parties have no choice in the matter. Therefore, a challenge to a judge involves an argument that the state’s process of allocation of a judge to be imposed on the parties should be overturned as a matter of law: This is a legal right to demand of the state that a particular person should not remain allocated. In contrast, in arbitration, the question is whether the parties will consent to submit their dispute before a particular person or persons. Disclosure is thus not merely a question of confidence; it goes to the very foundation of whether the parties will choose this person or persons to adjudicate their dispute. The first point of disclosure arises at the point of choice of arbitrator. Especially because of the confidentiality of arbitration, matters of importance may not be public knowledge and may therefore suggest disclosure. Thus, there must be some scope for parties to communicate with a prospective appointee in order that the party ascertain the suitability of the person for nomination by the party. Institutional rules reflect this freedom.^[39]

31 Although it is generally accepted that party-appointed arbitrators must be independent and impartial, as required by national laws and institutional rules,^[40] the requisite degree of impartiality required of a party-appointed arbitrator remains a vexed question in international arbitration. Indeed, the contested role of the party-appointed arbitrator has led some, most notably Jan Paulsson, to call for the abandonment of party-appointed arbitrators altogether.^[41] The epicentre of the contemporary debate appears to be whether party-appointed arbitrators do or should play a particular role, sometimes dubbed a ‘special role’ in an arbitration as compared with a neutral, presiding arbitrator. For Raymond Bishop and Lucy Reed, a ‘special role’ arises because of the “natural distinction between [parties’] expectations for the party-appointed arbitrators and the presiding arbitrator.”^[42] By virtue of being selected by a party, and consequently being likely to share a similar national, linguistic, cultural and legal background to that party, a party-nominated arbitrator may be seen to have a particular obligation to understand their appointing party’s case, ensure that the presiding arbitrator is not biased against their appointing party, and ‘translate’ their appointing party’s case when language, cultural or legal background may be distorting the other tribunal members’ understanding of their claims.^[43] Redfern and Hunter describe such a function as “useful” in ensuring “due process for the party that nominated him or her, without stepping outside the bounds of independence and impartiality.”^[44] Bishop and Reed contend that a distinctive role for the party-appointed arbitrator is the prevailing practice in international arbitration.^[45] From one perspective, these views are not a denial or qualification of the duty of impartiality;^[46] rather they reflect how practically that duty should be seen to operate in the real world. It is not partiality for one party; it is ensuring that that party’s case is well understood by the whole tribunal. The distinction is subtle, but perhaps real, although the danger of sliding towards partiality can be recognised.^[47] 

32 Others are opposed to any recognition of a ‘special role’ for party-appointed arbitrators on the basis that so doing inevitably introduces confusion as to the expected standards of impartiality and independence.^[48] The Chief Justice of Singapore wrote in 2017 that he was “fundamentally uncomfortable” with the suggestion that party-appointed arbitrators owe their appointing party any special duty.^[49] Chief Justice Menon stressed that all arbitrators are obliged to consider carefully the cases presented by both sides and cautioned that “the line between driving a lively discussion to ensure that a party’s case is fully understood and urging that party’s position upon the arbitral tribunal may be so fine that it is impossible to work with in practice.”^[50] Chief Justice Menon’s driving concern was the maintenance of confidence in the institution of arbitration, and the perception that it is a procedurally fair process of dispute resolution.^[51]

33 It is difficult, however, to avoid the recognition that, as Gary Born says, party-appointed arbitrators are an important pillar supporting the trust and confidence of commercial parties in international arbitration.^[52] As an incident of the centrality of party autonomy in international arbitration, and as a matter of practical reality,^[53] the ability for parties to name a person of their choice to the arbitral tribunal engenders an arbitration with legitimacy for them.^[54] Such legitimacy may be seen as crucial to the parties’ investment in the fairness of the process of the arbitration and their respect for its outcome, whether favourable to them or not.^[55] It may be better to say that the characteristics of party-appointed arbitrators may differ from those of presiding arbitrators without offending the basal requirement that all arbitrators be independent and impartial. As Born notes, co-arbitrators may have qualities such as a shared nationality with an appointing party, particular experience, philosophical outlook and professional ties that may be desirable for a party-appointed arbitrator but inappropriate for a presiding arbitrator. Such characteristics can be seen to bring the necessary cultural and conceptual diversity to the resolution of a complex case which might involve cultural, or legal or other diversity.^[56]

34 I now turn to Halliburton. The case concerned a large insurance claim arising from the explosion and fire on the Deepwater Horizon drilling rig in the Gulf of Mexico in 2010. Various parties were liable for damages and penalties, including a BP company which leased the rig, Transocean which owned the rig and provided crew and drilling teams, and Halliburton which provided cementing and other services to BP in relation to relevant activities. Halliburton had taken out liability insurance with Chubb. It made a claim for the amount of money it had paid in the settlement of private claims of US$1.1bn. Transocean settled the private claims for US$212m. Later, in 2014, after the settlements, the Federal Court apportioned blame between these three parties in the proportions of BP 67%, Transocean 30% and Halliburton 3%. Chubb declined Halliburton’s insurance claim on the basis, inter alia, that it was not a reasonable settlement. Transocean also made insurance claims against its liability insurers, which included Chubb. Chubb contested its liability to Transocean on substantially the same grounds as its defence to Halliburton’s claims. Both Transocean and Halliburton had arranged their cover under the Bermuda Form with an ad hoc arbitration provision. Both their policies were layered, with Chubb the top layer of each policy, and the policies were, relevantly in identical terms. The policies were governed by New York law and the arbitration clause in each provided for arbitration in London by a tribunal of three arbitrators, two party-appointed, the third chosen by agreement of the two party-appointed arbitrators, or failing agreement, by the High Court in London.

35 Halliburton nominated Professor Park of Boston University, an experienced arbitrator; Chubb nominated Mr Cole, an accomplished insurance executive, counsel and arbitrator. They could not agree upon a chair. After a contested hearing, on 12 June 2015, the High Court appointed Mr Rokison QC whom Chubb had proposed (amongst others) to the Court. Halliburton’s objections to all Chubb’s proposed arbitrators were that they were English lawyers or former judges known to Chubb.

36 Mr Rokison had disclosed to the Court and to Halliburton that he had previously acted as an arbitrator in arbitrations where Chubb was a party, including some as a party-appointed arbitrator nominated by Chubb and that he had two pending references in which Chubb was a party. There was no suggestion that this disclosure was in any way deficient.

37 The matter before the High Court in which Mr Rokison was appointed can be referred to as it was in the decision as reference 1.

38 Six months later, in December 2015, Mr Rokison accepted appointment by Chubb in an arbitration concerning an excess liability claim by Transocean arising out of the same incident. This second arbitration was referred to as reference 2. Before accepting appointment in reference 2, Mr Rokison disclosed his appointment in reference 1 and the other Chubb arbitrations which he had disclosed to Halliburton in reference 1. There was no objection to Mr Rokison by Transocean. Mr Rokison did not, however, then disclose to Halliburton in reference 1 of his proposed appointment in reference 2.

39 Later, in August 2016, Mr Rokison accepted appointment in another Deepwater Horizon arbitration on the joint nomination of parties, Transocean and another insurer on the same layer of insurance as Chubb’s layer in reference 2. This was referred to as reference 3. The proposed appointment to reference 3 was also not disclosed to Halliburton.

40 In both references 2 and 3 there was a preliminary issue of policy interpretation on uncontested facts which was potentially dispositive of both claims. This was heard in November 2016.

41 In November 2016, Halliburton learned of Mr Rokison’s appointments in references 2 and 3. Halliburton’s attorney wrote to Mr Rokison, referring to the IBA Guidelines on Conflicts which contemplated a continuing duty of disclosure of matters on the “Orange List”. In explanation (accepted by Halliburton as truthful) Mr Rokison said that it had not occurred to him that he was under a duty to disclose the later appointments and that he did not believe that the IBA Guidelines had been breached; but he said that with the benefit of hindsight it would have been prudent to inform Halliburton; and he apologised. He offered to resign from references 2 and 3 if the preliminary issue did not bring the arbitrations to an end. Halliburton’s lawyer repeated his concern and called for Mr Rokison’s resignation. (Mr Justice Popplewell described some of the terms of the attorney’s correspondence as “grossly offensive”.) Chubb would not, however, agree to Mr Rokison’s resignation. Mr Rokison again wrote, repeating his views that the IBA Guidelines had not been breached, but again accepting that, with hindsight, it would have been prudent to have made disclosure. He said that he had not learnt of any facts about the incident in references 2 and 3 that were not public knowledge. He said that he would resign if it were only up to him and suggested that the parties agree upon a substitute chair. On 21 December 2016, Halliburton sought Mr Rokison’s removal as arbitrator in the High Court under s 24(1)(a) of the 1996 Act. The application was heard by Mr Justice Popplewell who dismissed it on 3 February 2017.

42 The hearing in reference 1 meanwhile took place between 27 January and 6 February 2017.

43 On 1 March 2017, awards were issued in references 2 and 3 deciding the preliminary issues in favour of Chubb and the other insurer. This brought an end to the claims, without the need to consider any relevant evidence or submissions.

44 On 5 December 2017, the Tribunal in reference 1 issued its final award on the merits, deciding in Chubb’s favour. The award was signed by all three arbitrators, though Professor Park qualified his signature with “separate observations” to confirm his participation; but he expressed an inability to join in the award because of his “profound disquiet about the arbitration’s fairness”. The separate observations did not, however, dissent from any finding of fact, statement of applicable law, process of reasoning or final conclusions.

45 Interestingly, for the discussion of party-appointed arbitrators, the evidence in the High Court challenge revealed that Professor Park had been Halliburton’s party-appointed arbitrator in three references against different insurers in Deepwater Horizon claims, without formal disclosure. The Halliburton attorney who had confronted Mr Rokison in the correspondence suggested in evidence that Professor Park was in a different position as a party-appointed arbitrator: to be distinguished from Mr Rokison’s position as a chair or umpire. The use of the word “umpire” implied an advocacy role for a party-appointed arbitrator beyond any notion of cultural translation.

46 From these facts the Supreme Court dealt with the questions of disclosure, confidentiality and apparent bias. In (if I may be permitted to say) a thoughtful and sophisticated judgment, Lord Hodge JSC for the Court (with additional insightful comments from Lady Arden JSC) clarified a number of important aspects of the English common law in relation to apparent bias and disclosure in the context of arbitration under English law. In summary, the conclusions were:

(1) The obligation of impartiality is a core principle and the duty applies equally to party-appointed arbitrators and the presiding arbitrator or chair. This resolves, for English law, the debate about the role of the party-appointed arbitrator, though in the operation of the duty to disclose and of the test of apparent bias, the existence of the debate and differences of views were recognised to be important.

(2) The relevant human construct to assess both whether a duty of disclosure arises and whether there is a real possibility of bias is the fair-minded and informed observer and the test of bias is whether the observer would consider there was a real possibility of bias, assessed objectively. This assessment is made having regard to the customs and practices of international arbitration in any particular field. This equates, as Lord Goff had urged in R v Gough, the expression of the test for apparent bias in arbitration with that of curial adjudication.

(3) There is a legal duty of disclosure which is a component of, or drawn from, the statutory and contractual duty to act fairly and impartially. The duty of disclosure is wider than matters that would give rise to an apprehension of bias and extends to matters that would or might give rise to an apprehension of bias. The duty of disclosure is a continuing one. This can be seen to distinguish the position of arbitrator from that of judge if, as in Australia, the similar disclosure is a matter of prudence and professional practice.

(4) The duty of disclosure does not override the arbitrator’s duty of privacy and confidentiality, but absent a contract restricting disclosure, certain matters can be disclosed without breach of the duty of privacy and confidentiality and without obtaining specific or express consent: the identity of the common party in multiple references; whether a proposed appointment to another reference was to be a party-appointment or as nomination as the third arbitrator; a statement of the fact that the other reference arises from the same incident or had the same or overlapping subject matter; and a high-level statement about whether similar issues were likely to arise in the other reference. Consent to such disclosures can be inferred.

(5) A failure by the arbitrator to make disclosure which should have been made is a factor to be taken into account by the fair-minded and informed observer construct as to whether there is a real possibility of bias.

(6) The relevant dates to evaluate whether the duty of disclosure has been complied with, or whether there is a real possibility of bias are different. For the former, the test is prospective: by reference to facts and circumstances at the time the duty arose. For the latter, the test is by reference to facts and matters known at the date of the relevant application to the Court, such as to remove the arbitrator.

47 Time does not permit detailed examination of the reasons of Popplewell J or the Court of Appeal. Neither found a basis for apprehended bias, but the Court of Appeal said that Mr Rokison had failed to comply with a legal duty of disclosure. 

48 The appeal to the Supreme Court by Halliburton attracted interventions by the LCIA, the ICC and CIArb that the Court of Appeal’s judgment was out of step with international standards in their discussion of multiple common appointments. Other interventions were made by LMAA,^[57] GAFTA^[58] and ARIAS (UK)^[59] which were anxious to see unimpaired the ability to appoint common arbitrators to chain or string, supply, transport, shipping and reinsurance arbitrations.

49 Halliburton’s case was one of unconscious apparent bias: in Mr Rokison taking the benefit of paid appointments in references 2 and 3, Mr Rokison gave Chubb an unfair advantage in being a common party in related arbitrations where (to the ignorance of Halliburton) he could be influenced by evidence and submissions in references 2 and 3 which could affect his approach in reference 1; Chubb would know of his responses to evidence and submissions in references 2 and 3, but Halliburton would not; the failure to disclose deprived Halliburton of forming its own view about the position; and Mr Rokison was said not to have paid proper regard to Halliburton’s interests in the fairness of procedure.

The duty of impartiality – discussion of principle^[60]

50 Lord Hodge applied in the arbitral context the relevant test for apparent bias in judicial proceedings taken from Porter v Magill^[61] and Helow v Secretary of State for the Home Department.^[62] Without going into detailed textual comparison (and, if I may say, rightly so), he said that the English common law test was “similar” to the test of “justifiable doubts” in the UNCITRAL Model Law.^[63]

51 This test was held to apply to judges and all arbitrators. However, the importance of the reasons of Lodge Hodge can be found in his Lordship’s emphasis on context and his helpful discussion of the contextual differences between judicial and arbitral adjudication.^[64]

52 First, frank disclosure in arbitration is crucial because of the privacy of consensual arbitration. There may be no public record to which to have regard in evaluating an arbitrator, which makes it difficult or impossible for an arbitrator’s participation in multiple references and the submissions and evidence led in them to be known by those not participating in those references.

53 Secondly, there is usually no appeal from an arbitral award, and there are very limited powers of review. Therefore, the choice of the arbitrator is of critical importance.

54 Thirdly, the judge holds public office and is paid by the state; the arbitrator is chosen by one or both parties or by an institution and is remunerated by the parties, and ultimately, by the losing party. An arbitrator in practice has an interest in future appointments. This may give rise to an interest in avoiding action that may alienate the parties, such as assertive case management.

55 Fourthly, there is a diverse range of persons who may be appointed as arbitrators, coming from many jurisdictions and different legal traditions, with potentially diverging views on ethically acceptable conduct.

56 Fifthly, and importantly, Lord Hodge recognised the differing understandings of the obligations of the party-appointed arbitrator. His Lordship referred to the “lively debate” as to the role of party-appointed arbitrators, referring to the articles and writings of Paulsson, van den Berg, Brower and Rosenberg, Born, and Redfern and Hunter, discussed above. Lord Hodge rejected the nuances involved in the debate: No allegiance or duty is owed; no representative role is undertaken. Once appointed, independence and impartiality are unqualified. ^[65]

57 Lord Hodge was expressly influenced by the place of London in arbitral adjudication, and the English legal system’s reputation for guaranteeing neutrality and impartiality.^[66] His Lordship drew from Chief Justice Menon’s article that the requirement that all arbitrators had the same undifferentiated duty reflects an increasingly widespread legal norm. Whilst recognising that the view was not universal, and that some legal systems do accept that the party-appointed arbitrator has a special role in relation to his or her appointing party, this was not so in English law. So, in applying the relevant test, one does not have regard to the characteristics of the party including whether one or more was foreign.

58 The contrast in the approach in different legal systems was illuminated by the discussion of the 2018 Second Circuit decision in Certain Lloyd’s Underwriters v Florida^[67] which expressly recognised a distinction between party-appointed arbitrators and “neutral” arbitrators and that party-appointed arbitrators were expected to espouse the view of the appointing party and serve as de facto advocates.

59 Lord Hodge then sought to reconcile this international divergence in the perception of the role of party-appointed arbitrator in a subtle and important discussion of the attribution of knowledge to the human construct: The fair-minded and informed observer would be credited with the knowledge that some, perhaps many, parties and arbitrators have the view reflected in the Second Circuit’s reasons, and that there is a debate within the arbitration community as to the role of the party-appointed arbitrator and the compatibility of that role with the requirement of impartiality.^[68] This was not to recognise a different obligation within the application of the test in England; rather, it is to understand the context in which the objective judgment (by the Court through the human construct) will be made.

The duty of disclosure: discussion of principle ^[69]

60 The Court of Appeal was, in Lord Hodge’s view, correct to hold that there is a legal duty to disclose in English law. In arbitration, it arises from the obligation under the 1996 Act to act fairly and impartially, which gives rise to an implied term for the need to disclose matters.

61 To align or reconcile the duty to disclose with the duty of privacy and confidentiality, it becomes necessary to obtain the consent of relevant parties to disclosure, failing which consent the offered appointment must be declined. However, the context of particular trades or institutions (such as dealt with by the LMAA and GAFTA) may found an implied consent to multiple appointments. Importantly also, there was the recognition of the entitlement to disclose, confidentially, the existence of a current or past arbitration involving the common party and the identity of the common party, but not the identity of other parties without obtaining consent from all such parties. Such entitlement comes from common practice. Jurisprudentially this can be seen, as Lord Hodge said^[70], as the law recognising the realities of accepted and acceptable commercial and arbitral practice as a guide to the formulation of legal rules. Further, there was a public interest in upholding the integrity of arbitration as a system of dispute resolution calling for proper disclosure and calling for the giving of certainty to the legal standing of established arbitral practice.

62 Thus, at least in Bermuda Form arbitrations, an arbitrator may (subject to contrary agreement) make disclosure on a confidential basis of the existence of the arbitration and the identity of the common party without the express consent of other parties. Consent of the common party is inferred.

63 As to the content of the duty, Lord Hodge considered that with one qualification the duty as expressed by the Court of Appeal was correct:

…Disclosure should be given of facts and circumstances known to the arbitrator which … would or might give rise to justifiable doubts as to his impartiality … [which] means facts or circumstances which would or might lead the fair-minded and informed observed, having considered the facts, to conclude that there was a real possibility that the arbitrator was biased.^[71]

64 The qualification of Lord Hodge was to what was “known”. Generally there is no duty to inquire; one must disclose what one knows. However, Lord Hodge would not rule out circumstances raising a duty to make reasonable inquiries, noting that the IBA Guidelines, Part 1, General Standard 7(d) provide for a duty to make reasonable inquiries.^[72] His Lordship thus left to future cases the development of the question by reference to arbitral practice and rules.

65 Lord Hodge then discussed the phrase “would or might”.^[73] Any obligation to disclose something that “might” give rise to justifiable doubts can only arise, his Lordship said, where the matter might reasonably give rise to such doubts. Lord Hodge discussed the human contextual reality: the existence of matters that, if unexplained, would give rise to justifiable doubts, so the necessity to disclose and explain; or matters, more than trivial, which could, by themselves or with other facts, give rise to justifiable doubts, if later discovered. This analysis, necessarily less than precise, begins to explain the wider notion of disclosure and its relationship to the test for apparent bias, by reference to appropriate human conduct to maintain trust in the particular tribunal members and in the institution and process of arbitration as a whole.^[74]

66 As the discussion of Lord Hodge makes clear, the good sense of this approach comes from the consensual character of the process and the central value of party autonomy. It is the parties’ process. That is the attraction of arbitration: The existence of matters that might reasonably found a justifiable doubt as to impartiality, then or later, calls forth a sense of entitlement in the parties to know of the matter and to consider their positions for themselves. Reasonable minds can, and often do, differ about these sorts of questions. To recognise that and the policy of supporting trust and confidence in a consensual relationship built on party choice and autonomy justifies this wider penumbra of the obligation to disclose.

67 The development of English law to impose this duty was seen to be consistent with developments elsewhere.^[75]

68 Whilst the expression of the obligation under English law is not directed, as many arbitral rules are,^[76] to the perceptions of the parties, the distinction is unlikely to be sharp in practice. As expressed in summary form by Lord Hodge^[77] it is sufficient if matters are such that they are relevant and material to an assessment of the arbitrator’s impartiality and could reasonably lead to such an adverse conclusion. In practice, and in the context of the reality of the parties before the arbitrator, it is difficult to see how the objectively perceived relevance to the parties is not central to the objective assessment by the fair-minded and informed observer.

69 The duty of disclosure is continuing and subject to changing circumstances.^[78] Its fulfilment is to be tested prospectively, and not retrospectively by reference to matters known later to the fair-minded and informed observer at the point of assessing bias.

70 However, the question of whether justifiable doubts exist is to be assessed by reference to the fact as they exist at the time of the application.^[79]

71 Against that detailed discussion, Lord Hodge concluded that Mr Rokison should have disclosed to Halliburton the multiple references, but that in all the circumstances, by the time the matter came before the Court, the fair-minded and informed observer would not reach a conclusion of apparent bias.^[80] Particular relevant circumstances to that evaluation included that Mr Rokison gave an explanation of his failure to disclose his appointments in references 2 and 3 which was accepted by Halliburton’s lawyers as a genuine oversight, Mr Rokison was of the view that there was no material overlap between the references and had offered to resign if the separate issues in references 2 and 3 did not dispose of those references, and Mr Rokison at all times adopted courteous and measured responses.^[81]

72 Where does all this leave us? Halliburton highlights, with clarity, the intimate relationship between the duties of impartiality and of disclosure. Parties are only equipped to make a judgment upon whether they are satisfied that an arbitrator is sufficiently impartial notwithstanding a potential lack of complete independence if they are aware of all of the relevant circumstances, including multiple or overlapping appointments.^[82] An explicit duty of disclosure is a necessary incident of the private and confidential nature of most arbitrations, which may prevent the parties from ascertaining a complete picture of the circumstances from their own inquiries alone. In what may be seen as a welcome development supporting the all-important trust of parties in the arbitral process, the legal obligation to disclose a wider range of matters than those which would found a conclusion of apparent bias can now be taken to be recognised not just in many arbitral institutional rules, but also by the common law of England.

73 Commentary on the Supreme Court’s decision has canvassed areas for further development of the common law of England with respect to disclosure and remaining issues for debate.^[83] That does not reveal an omission in the judgment. Rather, it reflects that decision-making, in the context of concrete problems rather than exhaustive rule-making, is the foundation of the common law’s support for honest good sense in commerce.

74 The emphasis (with respect, correctly) placed by Lord Hodge on context may contain the key for those in Australia who may have to determine, as a matter of statutory construction, whether there really is any practical difference (at least in likely result) between the Gough test introduced into Australian arbitration legislation, with or without its “modest adjustment” in Porter v Magill of a fair-minded and informed observer (bearing in mind Lord Phillips’ view as to the identity of “real possibility” and “real danger”, removing any connotation of the immediacy of the threat from the word “danger”, and Lord Hodge’s view of the “similarity” of the Model Law “justifiable doubts” to the English test) and the Australian judicial test if applied in the context of arbitral adjudication in the relevant field of commerce in which the arbitration takes place, and if firmly established.

75 An issue that is perhaps under-explored in the commentary to date is the impact of the decision on the position of the party-appointed arbitrator. I make several final observations in this regard.

76 First, English law has set its face against recognition of any form of special duty or role of a party-appointed arbitrator, or of the chair.

77 That said, secondly, there is a recognition of the legitimacy of other views and other approaches about this in other jurisdictions.

78 Thirdly, it is not clear how taking these matters into account as part of context in the appreciation of the fair-minded and informed observer’s task will play out. That, perhaps, is only an aspect of the intensely factual and contextual nature of the relevant evaluation.

79 Fourthly, the use by Lord Hodge of the Second Circuit decision in the Florida case may not, with respect, fully deal with the subtleties involved. The American position, even after the change of the AAA/ABA Code of Ethics in 2004 towards recognition that all arbitrators should abide by the same formal ethical standards, is expressed, partly by reference to pre-2004 prevailing authority in the Supreme Court^[84] discussed in Florida, in language accepting the place of advocacy on the part of party-appointed arbitrator. The views of others as to the importance of the party having an impartial, but culturally and conceptually aware nominee, can be seen as more subtle than the American position. The confidence of parties in the arbitral process is crucial. That confidence is bolstered by an ultimate demand for honesty and impartiality; but it also may be seen to be bolstered by parties being happy with their informed choice of arbitrator to decide their dispute with the other side’s appointee and a chair chosen by them or appointed independently. There is already a qualification to independence: the arbitrator is chosen not imposed. But there need not, and should not, be a qualification to impartiality, unless the parties wish to choose a system and supervising court that finds impartiality to be legitimately qualified by a duty to advocate one side’s case.

80 Fifthly, whether one personally agrees with the views of Chief Justice Menon and Jan Paulsson, or with those of Gary Born, Raymond Bishop and Lucy Reed, may not matter. Different arbitral laws, statutory and general, may express themselves differently, as the Second Circuit and the Supreme Court did in Florida and Halliburton. This displays a choice that is given to parties and their advisors about the law of the arbitration and the seat. Parties can choose the approach they wish to have taken.

81 Sixthly, and finally, the debate about party-appointed arbitrators may underscore the distinction between institutional and ad hoc arbitration. A denial of the right of party-appointment would throw into high relief the need for integrity of the institution appointing the arbitration panel or the pool from which the panel is chosen. Reliance on the institution may simply push the relevant questions of independence, impartiality and disclosure further away from the knowledge of the parties and from the operation of party autonomy. Or, it may give institutions the opportunity to fashion bespoke solutions to any perceived problems of compromise of impartiality by party-appointment, such as by “blind” appointments, where the arbitrators do not know whether they are party-appointed or not.