chevron-down Created with Sketch Beta.


Enforcing Standard-Form Arbitration Agreements in Class Actions: Lessons from the Uber Saga in Canada

Gannon G Beaulne, Maya Bretgoltz, and Ranjan Agarwal

Enforcing Standard-Form Arbitration Agreements in Class Actions: Lessons from the Uber Saga in Canada
fotostorm via Getty Images

In June 2020, the Supreme Court of Canada released its decision in Uber Technologies Inc. v. Heller, 2020 SCC 16, a proposed class action alleging violations of Ontario’s employment standards legislation. The decision arose from a preliminary request by Uber to stay the proposed class action in favor of arbitration under an arbitration clause in a standard-form agreement governing drivers. The plaintiff argued that the arbitration clause was unfair and should not be given effect because it required drivers to start arbitration proceedings in Amsterdam, which involves incurring US$14,500 in filing fees. Uber argued that any issue concerning the arbitration agreement’s validity should be referred to arbitration under the general rule of systematic referral of challenges to jurisdiction, as laid down by the Supreme Court in Dell Computer Corp. v. Union des consommateurs, 2007 SCC 34, another proposed class action.

Uber’s stay request underscores a tension between two precepts of private law when they interact in the context of class actions and arbitration: on the one hand, party autonomy and freedom of contract—including that the arbitrator, not the courts, should typically decide questions of arbitral jurisdiction—and, on the other hand, access to justice.

In Uber, the Supreme Court decided in favor of access to justice. It found that it had jurisdiction to determine the validity of Uber’s arbitration agreement because there was a “real prospect” that the issue would, for practical reasons, never be resolved if referred to the arbitrator. Then seven of the nine Supreme Court judges found Uber’s arbitration agreement invalid under the contract law doctrine of unconscionability (a novel extension of that doctrine in the arbitration context).

For U.S. companies operating in Canada, and their lawyers, Uber may raise questions about whether mandatory arbitration clauses remain an effective tool for managing class action risk. Because arbitration cannot always match class actions in accessibility for lower-income parties, an access to justice exception to systematic referral might reasonably be viewed with distrust.

This article unpacks Uber and assesses its implications for enforcing arbitration agreements in class actions. It considers how companies can seek to stay ahead of the curve and how Uber itself has responded to the decision by adjusting its standard-form agreement. It concludes that Uber likely does not change the basic posture of Canadian courts in favor of arbitration but merits careful consideration by companies interested in managing class action risk. Ultimately, Uber is contextually particular. The default remains deference, with systematic referral still the operating principle in normal circumstances. But arbitration agreements that do not safeguard some realistically attainable dispute mechanism will be subject to potential challenge after Uber.


David Heller—a Toronto-based food services and delivery driver—started a proposed class action in Ontario against Uber, alleging violations of Ontario’s Employment Standards Act (ESA). To become a driver, Mr. Heller had to sign a standard-form services agreement. Under the agreement, drivers were required to resolve any disputes with Uber by mediation or mandatory arbitration in the Netherlands, under the International Chamber of Commerce (ICC) rules. Starting an ICC arbitration means paying US$14,500 in up-front administrative and filing fees, amounting to most of Mr. Heller’s annual income as an Uber driver. In practice, the arbitration clause thus preemptively shut down most claims that drivers were likely to bring.

Uber moved to stay Mr. Heller’s action in favor of arbitration. The motion judge granted the stay, holding that the arbitration clause was valid. A unanimous panel of the Court of Appeal for Ontario overturned that decision, accepting that it could consider the validity of the arbitration agreement, and found that Uber’s standard-form clause was unconscionable and unenforceable.

Uber sought and obtained leave to appeal to the Supreme Court of Canada.

The Majority Decision

A New Exception to the Systematic Referral Rule

In its 2007 decision in Dell and its 2011 decision in Seidel v. TELUS Communications Inc., 2011 SCC 15, the Supreme Court established a framework to help courts determine when they can rule on the validity of an arbitration agreement rather than refer that question to arbitration. Under that framework, courts can depart from the general rule of systematic referral if the challenge to the agreement’s validity is a question of pure law or mixed law and fact requiring only a “superficial review” of the evidentiary record.

Unlike in Dell and TELUS, the arbitration agreement in Uber created practical obstacles to drivers’ access to justice, which, the majority found, the Ontario legislature could not have intended when granting courts the power to grant or refuse a stay. The majority acknowledged that previous Supreme Court decisions “did not contemplate a scenario wherein the matter would never be resolved if the stay were granted” and that the up-front ICC fees “impose[d] a brick wall between Mr. Heller and the resolution of any of the claims he has levelled against Uber.”

Faced with this “Gordian Knot,” the majority recognized a new exception to the systematic referral rule: If there is a genuine challenge to arbitral jurisdiction and a “real prospect” that referring it to the arbitrator would result in the challenge never being resolved, the court could decide the arbitration agreement’s validity.

The Majority’s Finding that the Arbitration Agreement Was Unconscionable

Having established jurisdiction to decide validity, the Supreme Court found Uber’s mandatory arbitration clause invalid as unconscionable. The unconscionability doctrine applies when there is proof of both inequality in the position of the contracting parties and an improvident bargain.

On the first criterion, the majority focused on the standard-form nature of the driver agreement. Practically, Mr. Heller could not negotiate the agreement’s terms. The details (including expense) of the ICC arbitration were not clearly or effectively described. There was also a disparity in the parties’ bargaining power and sophistication.

On the second criterion, the majority found the ICC fees prohibitive and disproportionate to any reasonably foreseeable arbitral award. The majority criticized designating Netherlands law as the governing law and Amsterdam as the place of the arbitration because Uber gave drivers the impression that they would have to travel overseas at their own expense to pursue claims.

The Concurrent Judgment and Dissent

In a concurring decision, Justice Brown reached the same conclusion as the majority but would have grounded the invalidity of Uber’s arbitration agreement in the rule of law because precluding meaningful access to justice offends public policy. The solution, in his view, was not to “vastly expan[d] the scope of the [unconscionability] doctrine’s application and remov[e] any meaningful constraint” on it.

In a dissent, Justice Côté stated that the majority’s approach clashes with party autonomy, freedom of contract, legislative intent, and commercial practicality. She noted that the gig economy depends on standard-form contracts and, if companies in that sphere are not assured of certainty in contractual arrangements, the gig economy may suffer. Justice Côté also observed that, if the legislature was concerned about arbitration agreements in the gig economy, it could have amended the ESA to restrict those clauses, as it did with consumer protection legislation. Even so, Justice Côté would have granted the stay only if Uber agreed to advance the ICC fees.

What Uber Means for Enforcing Arbitration Agreements

The pursuit of access to justice and the enforcement of arbitration agreements have historically been complementary objectives. But the growing gig economy has, at times, pitted the two aims against one another. Multinational companies often rely on arbitration agreements to attempt to bar class actions. For prospective class members, those agreements can decrease access to meaningful recourse, given that a class action may be the only way lower-income workers can overcome the prohibitive expense of bringing claims. For multinational companies, however, mandatory arbitration clauses may be a business necessity, without which they could face proliferating disputes in manifold jurisdictions.

Canadian courts have generally resolved this conflict in favor of arbitration. The Supreme Court in TELUS confirmed this inclination, holding that valid arbitration clauses—even in contracts of adhesion—are enforceable unless prohibited by statute. Similarly, in TELUS Communications Inc. v. Wellman, 2019 SCC 19, the Supreme Court refused to allow claims subject to an arbitration clause to be included in a class action (although a strong dissent in Wellman presaged the Uber majority’s concern for access to justice). But eight of nine Supreme Court judges in Uber found that access to justice, at least in a compelling factual context, trumps presumptive deference to arbitration when the two principles are squarely at odds. The majority found that Uber’s standard-form driver agreement provided “illusory” recourse, creating friction between the systematic referral rule and the meaningful access to justice on which the class action system is premised. The majority held: “Respect for arbitration is based on it being a cost-effective and efficient method of resolving disputes. When arbitration is realistically unattainable, it amounts to no dispute mechanism at all.”

Companies and their lawyers drafting arbitration agreements for potential use in Canada should consider whether the dispute mechanism they propose to put into place creates a realistically attainable method of resolving disputes.

Some critics of the decision suggest, in line with the dissent’s reasoning, that Uber creates commercial uncertainty about the enforceability of arbitration agreements. Based on early returns, that concern is likely overstated. The Supreme Court’s decision is consistent with deferring to arbitration agreements in the ordinary course, especially in commercial settings. The new access to justice exception that the majority recognized is grounded in a host of factual and legal particularities in the Uber case, although Mr. Heller is not likely to be the last plaintiff in a proposed class action to attempt to rely on that exception. The decision holds neither that valid arbitration agreements deny access to justice per se nor that arbitration agreements are invalid when class actions provide merely better recourse than arbitration. The requirement of a “real prospect” that the jurisdiction challenge realistically would never be resolved if referred to the arbitrator appears to set a reasonably high bar, subject to how courts interpret that language.

Post-Uber, multinational companies can continue confidently using valid arbitration clauses that have the purpose or practical effect of managing class action risk. But companies should ensure that their arbitration clauses are not so barrier-laden as to insulate them from any practical method or form of recourse.

Multinational companies using or considering mandatory arbitration clauses in standard-form contracts should use Uber as an impetus to reevaluate and strengthen the practicality, and therefore enforceability, of their contractual arbitration language. How Uber itself has responded to this decision may be instructive, although each agreement should be considered in its own context and with the benefit of legal advice.

As has been publicly reported, Uber Canada has made significant modifications to the arbitration clause in its standard-form agreements. Drivers now have the right to opt out of the arbitration clause within 30 days, with no repercussions. To opt out, drivers need only email Uber making that request (though, in practice, many drivers are unlikely to appreciate the implications of the opt-out provision). The modified arbitration clause also provides that arbitrations are now local for each driver (not overseas) and filing fees are capped at the cost of filing a claim in the driver’s local Canadian court (with the balance of fees, if any, payable by Uber). In addition, the modified arbitration clause preserves the right of drivers to file “a claim with a government agency.” Thus, for instance, a driver would still be able to file a complaint with the Ontario Ministry of Labour, Training, and Skills Development for an ESA violation. Finally, Uber’s modified arbitration clause contains a “class action waiver” provision, barring drivers who do not opt out of arbitration from pursuing or participating in any class actions against Uber for violating the standard-form agreement.

Whether Uber’s modified arbitration clause will achieve its objective remains to be seen, but it shows the many ways in which companies can respond to the Supreme Court’s concerns about access to justice. For example, companies can, depending on the context and their legal advice,

  • tailor the place of arbitration to where the counterparty lives;
  • select a place of arbitration that allows virtual hearings;
  • select a governing law based on where the counterparty lives;
  • agree to pay or subsidize the counterparty’s filing costs and implement a rule for cost shifting if a claim is found invalid or dismissed;
  • rather than pay the counterparty’s filing costs, choose a place of arbitration with lower fees or special rules for modest claims;
  • agree to reimburse a counterparty for any filing fees if the claim succeeds;
  • make any arbitration agreement detailed to ensure that the counterparty has notice of unusual or onerous requirements; and
  • give the counterparty a window of opportunity to opt out of an arbitration clause.

The practical effect of Uber is that companies will be more likely to choose local seats for their mandatory arbitrations and make arbitration more cost-effective for their counterparties. As a result, U.S. companies operating in Canada should be prepared to participate in more Canadian arbitrations in the future.

After Uber, class arbitration will likely become a more prominent topic of discussion. Class arbitration remains uncharted territory in Canada, unlike in the U.S. Interestingly, Uber’s new “class action waiver” provides that drivers cannot commence class arbitrations. It will be worth watching whether plaintiffs try to get around class action waivers by bringing mass arbitrations and whether corporations may come to regret the waivers. This has been a phenomenon in the U.S. where, for instance, DoorDash (another gig economy company) is facing over 5,000 arbitration applications, and about US$9 million in filing fees, after specifying that disputes would be heard by the American Arbitration Association and that it would pay most associated filing fees. Because of the influx of applications, DoorDash sought to stay the arbitrations in favor of a class action settlement. In denying the motion, Judge William Alsup held in Abernathy v. DoorDash, Inc., 438 F. Supp. 3d 1062 (N.D. Cal. 2020): “[I]n irony upon irony, DoorDash now wishes to resort to a class-wide lawsuit, the very device it denied to workers, to avoid its duty to arbitrate. This hypocrisy will not be blessed, at least by this order.”


In Uber, the Supreme Court of Canada recognized a side constraint on the concepts of party autonomy and freedom of contract in the arbitration context to reflect the operation of the principle of access to justice. In arbitration-friendly Canada, after Uber, the default rule remains that valid arbitration agreements should be enforced, and questions of arbitral jurisdiction should typically be referred to the arbitrator, especially in commercial settings. But Uber clarifies that arbitration agreements, especially in contracts of adhesion, should give effect to contracting parties’ genuine intention to refer disputes to a realistically attainable arbitration process and may be scrutinized by the courts to ensure that they do not preemptively shut down disputes and thus render contractual rights illusory.