Chris Reflects (2023): Arbitrator Bias and the Unanimous Award – #807

When will a court confirm a unanimous arbitral award issued by a three-person panel where one of those arbitrators was biased? This case note reviews three cases that try to answer that question. In each, the Court applied the Model Law. In one recent casethe Ontario Superior Court of Justice upheld the award, finding that the bias did not cause actual prejudice. The other two cases, one from India, the other from Germany, reached the opposite conclusion, highlighting the pernicious, and often unseen, effect that bias can have on the deliberative process.  

In Vento Motorcycles, Inc. v. United Mexican States, 2023 ONSC 5964 (“Vento”), an investor applied to set aside a NAFTA award alleging, among other things, that one of the three arbitrators on the tribunal was biased. That arbitrator failed to disclose communications with the government of Mexico, a party to the arbitration, by its lead counsel during the arbitration. The communications concerned the potentially lucrative opportunity of being named to Mexico’s panel of tribunal chairpersons under the Comprehensive and Progressive Agreement on Trans-Pacific Partnership and another trade agreement. 

When Vento discovered these communications, it sought to set aside the arbitral award. The Court had no hesitation in finding that the arbitrator’s conduct gave rise to a reasonable appreciation of bias. The Court, for example, wrote: 

“From a reputational perspective, these appointments can only be seen as being advantageous, as demonstrated by [the impugned arbitrator’s] statement in his e-mail… that he was “honored and grateful”.  Further, as Vento put it, these appointments to rosters were the “gateway” to future work and remunerative appointments to panels.  Thus, the appointments of [the arbitrator] to the rosters of panelists conferred upon him a professional benefit and the potential for future financial benefits.”

The Court, likewise, emphasized that the arbitrator, “had an incentive to please Mexico” and that, ‘whether consciously or unconsciously’, would have ‘a leaning, inclination bent or predisposition towards’ Mexico, or that he could be influenced by factors other than the merits of the case as presented by the parties in reaching his decision.”

The Court, moreover, accepted that the impugned arbitrator spent “significantly” more time on the case and, therefore, “may have done a significant part of drafting the Award.”

The Court, nonetheless, declined to set aside the award on the grounds that it could not find that the integrity of the award rendered by the three-member tribunal was tainted by the bias of the one member. In reaching this conclusion, the Court centred its analysis on the fact that the three-person tribunal rendered a unanimous decision. The Court, for example, noted that:

“[I]t appears unlikely that the Tribunal would have adopted a process where one arbitrator would have been assigned the task to go through the case to brief the other arbitrators.  Ultimately, all arbitrators signed the Award, and the reasonable conclusion is that all three arbitrators shared the same view as to the disposition of the arbitration and the reasons set out in the Award.

Based on this, and a presumption of impartiality that the other two arbitrators enjoyed, the Court determined that any bias of the one arbitrator did not produce real unfairness or real practical injustice.  This finding is surprising because the test is objective – what would a fair-minded and informed observer think?  

The Court’s conclusion about the tribunal’s process were speculative and do not sit easily with the actual evidence that the impugned arbitrator spent “significantly” more time on the case. It is also unknowable, as these processes are protected by deliberative secrecy which safeguards the confidentiality of deliberations among arbitrators during the proceedings. This confidentiality will only be lifted in the rarest of cases where a court determines that it is in the interests of justice to do so. 

The Court’s decision, moreover, runs against at least two decisions applying Section 34(2) of the Model Law that set aside arbitral awards in the face of bias by only one arbitrator on a three-person tribunal. 

Firstly, in M/S. Lanco-Rani (JV) vs. National Highways Authority of India Limited, 2016 SCC OnLine SC1616, a 2016 decision of the High Court of New Delhi, the petitioner challenged an arbitration under Section 34 of the Arbitration and Conciliation Act, 1996, (the “ACA”), which mirrors the Model Law.

The underlying dispute stemmed from the delay in completing a raised roadway project known as the Kishanganj Flyover. The petitioner sought compensation for additional costs and losses resulting from the delay in construction, following which the matter was referred to the Dispute Review Expert (“DRE”) and later to the arbitration tribunal as per the parties’ agreement.

The petitioner’s challenge to the award centred on allegations of bias concerning the arbitrator appointed by the respondent, National Highways Authority of India Limited (“NHAI”). After the tribunal rendered the award, the petitioner discovered that NHAI had retained the arbitrator as an advisor/technical expert in a different arbitration involving NHAI, during the Kishanganj Flyover arbitration, a fact neither he nor NHAI disclosed. 

The petitioner argued that this lack of disclosure created justifiable doubts about the arbitrator’s independence and impartiality. The Court agreed and set aside the arbitral award, concluding that the failure to disclose the arbitrator’s employment with NHAI compromised the fairness of the arbitration proceedings and rendered the award unsustainable in law.

The Court, moreover, explicitly addressed and rejected arguments that it should uphold the award because only one of the three tribunal members to the unanimous award was found to be biased, writing: 

“27. The fact that the Award may have been unanimous and that [the arbitrator] was only one of the [arbitrators on the tribunal] does not make even one bit of a difference to the above conclusion. This aspect of the matter has already been dealt with by the Supreme Court in A.K. Kraipak v. Union of India (supra). There again, it was urged that the mere fact that one of the Members of the Board was biased against some of the Petitioners cannot vitiate the entire proceedings. The Supreme Court negated this plea since the Court was essentially concerned with the question whether the decision taken by the Board can be considered as having been taken fairly and justly. This was because of the conflict between duty and interest. In other words, even if one of the [arbitrators] has compromised the essential requirement of fairness by failing to disclose the circumstances which may give rise to justifiable doubts as to independence and impartiality, the [award] would get vitiated.

Secondly, in a decision indexed at docket number Beschluss des I. Zivilsenats vom 11.12.2014 – I ZB 23/14, the German Supreme Court, or Bundesgerichtshof, came to the same conclusion in another case involving allegations of bias against one of three arbitrators, and in the face of a unanimous award.

As background, the case arose from a lease dispute for a thermal bath. The agreement required that disputes relating to the lease would be heard by a three-person arbitration tribunal in Germany. Each party appointed a nominee arbitrator, who in turn appointment a third arbitrator as the tribunal’s chairperson. One party challenged the appointment of the tribunal’s chairperson, citing bias, which the reasons for judgment do not discuss. 

The tribunal itself dismissed the challenge. That party then brought a challenge in the German courts while the arbitration continued, relying on legislation that incorporated the Model Law. In April 2013, the tribunal rendered a unanimous award before the German court released a decision. In January 2014, however, the Higher Regional Court of Munich agreed with the applicant and concluded that the tribunal was not properly constituted. 

The German Supreme Court upheld the decision, and in doing so addressed the effect of a unanimous award should weigh against setting aside the award. It dismissed arguments that unanimity could undermine the effect of the bias, reasoning that in multi-arbitrator panels, deliberation precedes voting, leaving room for an arbitrator’s influence. Therefore, establishing a direct causal link between flawed composition and the award was unnecessary, given the potential for an arbitrator’s influence during deliberations: “[i]t is always possible that the behavior of a judge at the deliberation and voting influences the formation of opinions and the voting behavior of the other judges.” Given this possibility, it was unnecessary for the applicant to establish a causal link between the potential bias of the arbitrator and the outcome of the award. 

These cases paint a vivid picture of the critical role the reasonable apprehension of bias or justifiable doubts tests plays in the sanctity of the arbitration process. While the recent Ontario Superior Court of Justice in Vento declined to set aside the award despite identified bias, the decisions from India and Germany are more consistent with the harms that reasonable apprehension of bias seeks to protect against, and more closely adhere to idea is reflected in the famous quote that it “… is of fundamental importance that justice should not only be done, but should manifestly and undoubtedly be seen to be done (The King v. Sussex Justices, [1924] 1 K.B. 256 at 259).

These decisions underscore the paramount importance of not just actual bias but also the perception or appearance of bias, recognizing that this perception alone can stain the integrity of the decision-making process. Deliberative secrecy, as protected in these instances – and rightly so – must be given immense weight, but it reveals the possibility of potential vulnerabilities in tribunal dynamics, especially when one arbitrator significantly shapes the award. These cases collectively illuminate the delicate balance courts must strike between considering the gravity of a breach in procedure (which can never be proven) within the larger context of the fairness of the arbitration process to ascertain its impact on the final award.