Over the last couple of years, arbitration practitioners across the country have had their eyes on the Ontario courts, which have considered several cases on the issue of reasonable apprehension of bias. For this reason, I’ve chosen to focus on Vento Motorcycles, Inc. v Mexico, 2025 ONCA 82, as a year-end reflection for 2025. Vento is an important reminder of the centrality of arbitrator independence to the institution of arbitration, and how its absence (whether real or perceived) can crumble the foundation of the arbitration process.
The facts underlying the Vento decisions by the Ontario Superior Court and Ontario Court of Appeal have been detailed in the following previous Arbitration Matters case notes: Ontario – No unfairness despite Browne v Dunn violation and arbitrator reasonable apprehension of bias – #796 – Arbitration Matters; Chris Reflects (2023): Arbitrator Bias and the Unanimous Award – #807 – Arbitration Matters; Jim Reflects (2023): Browne v Dunn is just a rule of fairness: a comment on the Vento case – #810 – Arbitration Matters; and Ontario – Award set aside where one of three arbitrators biased – #898 – Arbitration Matters.
Vento involved an application to set aside a final award in an arbitration arising out of a dispute under Chapter 11 of the North American Free Trade Agreement. Vento Motorcycles Inc. commenced an arbitration against the United Mexican States (“Mexico”) after Mexico denied preferential import tariffs to motorcycles that Vento assembled in the United States, which ultimately destroyed Vento’s business under a joint venture agreement with a Mexican company to sell and market motorcycles in Mexico.
The arbitral tribunal consisted of three arbitrators, with each party appointing one arbitrator and the tribunal president selected through an agreed process. While the arbitration was ongoing, Mexico’s tribunal appointee had a series of communications with members of the Mexican government, including Mexico’s lead counsel in the arbitration. The communications related to potential future arbitral appointments to a prestigious tribunal for Mexico’s tribunal appointee. These communications were not disclosed to Vento.
The tribunal rendered its final award, which was unanimously in Mexico’s favour. Vento subsequently learned of the undisclosed communications involving Mexico’s tribunal appointee, and applied to set aside the award on two grounds:
(1) that it was denied procedural fairness – this aspect of the application is not addressed further in this piece; for commentary on it, see Jim Reflects (2023): Browne v Dunn is just a rule of fairness: a comment on the Vento case – #810 – Arbitration Matters and Ontario – No unfairness despite Browne v Dunn violation and arbitrator reasonable apprehension of bias – #796 – Arbitration Matters; and
(2) that there were justifiable doubts as to the impartiality and independence of Mexico’s tribunal appointee.
On the latter issue, the Ontario Superior Court found that a reasonable apprehension of bias arose from Mexico offering its tribunal appointee valuable opportunities for future appointments, and that the appointee had a duty to disclose those ex parte communications. In fact, the arbitrator was notified, prior to issuing the final award, that he had been presented as one of Mexico’s candidates for appointment to rosters of arbitrators under the Comprehensive and Progressive Agreement on Trans-Pacific Partnership and the Canada-United States-Mexico Agreement, and that he had been appointed to those rosters. However, despite making a finding of reasonable apprehension of bias, the court exercised its discretion under article 34 of the Model Law to decline to set aside the tribunal’s award. In the court’s view, the reasonable apprehension of bias did not undermine the reliability of the award as the award was unanimous, and there was potential prejudice in requiring the arbitration to be reheard. This outcome was driven by the permissive language in article 34, which states that an arbitral award “may” be set aside by the court if certain listed criteria are met.
Vento appealed to the Ontario Court of Appeal. On the appeal, Mexico did not challenge the finding that there was a reasonable apprehension of bias; the only issue was whether it was appropriate for the application judge to have declined to set aside the award. The Court of Appeal overturned the application judge’s decision and set aside the award, on the basis that a finding of reasonable apprehension of bias necessarily results in the disqualification of the affected arbitrator and voids the award (if already rendered) because there is no way to remedy the damage caused by a reasonable apprehension of bias.
Mexico sought leave to appeal to the Supreme Court of Canada. The leave application was dismissed in September 2025.
It was a relief to see the Court of Appeal correct the application judge’s surprising analysis in this case. It is difficult to understand how a reasonable apprehension of bias in respect of one tribunal member could not taint the tribunal’s award. The award does not tell the whole story, and it cannot be assumed that the result would be the same because two of the three decision-makers were independent and impartial. Bias on the part of a single tribunal member could infect many aspects of the arbitration process, from procedural decisions, to questions during the hearing, to tribunal deliberations.
Indeed, in Vento, the fees of Mexico’s appointee were significantly higher than those of the other arbitrators. It can easily be speculated that the absence of that member from the various stages of the arbitration could have led the other tribunal members to arrive at a different conclusion.
By overturning the application judge’s decision, the Court of Appeal aligned the result in Vento with the law in other arbitration cases (e.g. Szilard v Szasz, 1954 CanLII 4 (SCC)), as well as administrative contexts (e.g., Sparvier v Cowessess Indian Band (T.D.), 1993 CanLII 2958 (FC)), to the effect that a reasonable apprehension of bias on the part of one member of an adjudicative tribunal is sufficient to disqualify the whole tribunal. The court in Vento stated that “[t]he parties to an arbitration are entitled to an independent and impartial tribunal, not simply the decision of a quorum of panel members who are unbiased” (para 46).
The Court of Appeal also made clear that there’s a hard-line rule in relation to reasonable apprehension of bias: where found, it necessarily results in disqualification of the arbitrator and the voiding of any final award rendered. There are relatively few hard-line rules, with no residual judicial discretion or contextual analysis, in law. However, such a strict approach to this issue makes sense in the arbitration context because arbitrations are fundamentally grounded in contract, and independence and impartiality are key components of that contract. In their absence – even in respect of only one of several tribunal members – the parties have not gotten what they bargained for and the structure collapses. A strict rule protecting the impartiality of the arbitral tribunal is necessary to preserve the integrity of arbitration as a dispute resolution mechanism.
The flip side is that, in order for the parties to get what they bargained for, it may be necessary for the tribunal to be reconstituted with a new member who will need to get up to speed, or for the arbitration to be restarted altogether with a partly or wholly new tribunal. As previously noted by my colleague, Lisa Munro (in Ontario – Award set aside where one of three arbitrators biased – #898 – Arbitration Matters), this is a costly consequence to the parties – who may not have any fault for the circumstances giving rise to the reasonable apprehension of bias, or who may attempt to game the system and withhold knowledge of a risk to impartiality until a later date so as to subject an innocent party to elevated legal fees or extended timelines for resolution.
Where a reasonable apprehension of bias results in a set aside, the potentially drastic increase in the cost of the arbitration process may lead some parties to seek or feel pressured to consider alternatives to repeating the arbitration, such as settling on relatively unfavourable terms or even pursuing their disputes through the court system. In the latter case, a party attempting to move the dispute to the court system will need to be prepared to face a stay motion, which could be a new source of cost and delay – but perhaps a worthwhile risk for exceptionally complex arbitrations where a repeat would involve even more cost and delay.
I query whether a party facing a stay motion in this context may be able to rely on the unconscionability analysis set out in Uber Technologies Inc. v Heller, 2020 SCC 16, to argue that the increased costs associated with repeating an arbitration render the arbitration agreement improvident and unenforceable at that point. This seems like an unlikely outcome, but where arbitration costs skyrocket, parties may see it as worthwhile to roll the dice and try to pursue a relatively less costly process.
As a final point of note, the court in Vento provided interesting commentary on the discretion afforded to judges under article 34 of the Model Law, attempting to thread the needle to balance the seriousness of a reasonable apprehension of bias against the courts’ inherent authority to control their own processes and prevent abuse. The court explained that, unlike other grounds on which a court may set aside an award under article 34, there is no discretion to refuse to set aside an award where a reasonable apprehension of bias is found. This is because “[a] finding of a reasonable apprehension of bias is necessarily serious and must be made in the knowledge of the result that follows: the adjudicator is disqualified and the substantive decision is void” (para 33). However, the court also recognized that judges’ inherent discretion to control the court process remains, such that judges may refuse to consider allegations of bias in appropriate circumstances (e.g. where an applicant has not raised a bias allegation at the earliest opportunity, and therefore may be considered to have waived protection from bias).
These comments are difficult to reconcile. If a finding of reasonable apprehension of bias is so serious as to remove judicial discretion and demand an onerous outcome, then how can a party’s protection from that serious circumstance be waived by mere inference? And how can a judge reasonably avoid entertaining an argument that a reasonable apprehension of bias exists, and potentially allow that circumstance to continue?
Consequences aside, Vento serves as a reminder that independence and impartiality are bedrock principles of arbitration. One or both principles typically are guaranteed to the parties through the statutes and/or rules that are incorporated into arbitration agreements (unless the parties specifically opt out, depending on what is permitted by the statute and/or rules applicable in a given case). And when they are absent (without party agreement to that state of affairs), a court will be compelled to remove the affected arbitrator and unwind the process to the extent necessary to remove their impact, including by setting aside an award in which they participated.