Arbitration - Jurisdiction. Uber Technologies Inc. v. Heller
In Uber Technologies Inc. v. Heller (SCC, 2020) the Supreme Court of Canada - in one of the messiest bits of law I recall for a long time - considers the exceptions to the rule that the presence of an arbitration agreement prevails, the further (unhelpful and circular) issue of who decides this, the arbitrator or the court and - apparently - another exception to those exceptions:
 The AA directs courts, on motion of a party, to stay judicial proceedings when there is an applicable arbitration agreement:. Alectra Utilities Corporation v. Solar Power Network Inc.
Stay But a court has discretion to retain jurisdiction and decline to stay proceedings in five circumstances enumerated in s. 7(2):
7 (1) If a party to an arbitration agreement commences a proceeding in respect of a matter to be submitted to arbitration under the agreement, the court in which the proceeding is commenced shall, on the motion of another party to the arbitration agreement, stay the proceeding.
ExceptionsThe only relevant exception here is para. 2 of s. 7(2), which gives a court discretion to refuse to grant a stay if the court determines that the arbitration agreement is invalid.
(2) However, the court may refuse to stay the proceeding in any of the following cases:
1. A party entered into the arbitration agreement while under a legal incapacity.
2. The arbitration agreement is invalid.
3. The subject-matter of the dispute is not capable of being the subject of arbitration under Ontario law.
4. The motion was brought with undue delay.
5. The matter is a proper one for default or summary judgment.
 The AA is silent on what principles courts should consider in exercising their discretion to determine the validity of an arbitration agreement under s. 7(2). But some criteria were set out in Dell Computer Corp. v. Union des consommateurs, 2007 SCC 34 (CanLII),  2 S.C.R. 801 and Seidel v. TELUS Communications Inc., 2011 SCC 15 (CanLII),  1 S.C.R. 531, which interpreted similar arbitration regimes in Quebec and British Columbia. In those decisions, this Court set out a framework for when a court should decide if an arbitrator has jurisdiction, instead of referring that question to the arbitrator out of respect for the competence-competence principle.
 Under the Dell framework, the degree to which courts are permitted to analyse the evidentiary record depends on the nature of the jurisdictional challenge. Where pure questions of law are in dispute, the court is free to resolve the issue of jurisdiction (para. 84). Where questions of fact alone are in dispute, the court must “normally” refer the case to arbitration (para. 85). Where questions of mixed fact and law are in dispute, the court must refer the case to arbitration unless the relevant factual questions require “only superficial consideration of the documentary evidence in the record” (para. 85).
 In setting out this framework, Dell adopted an approach to the exercise of discretion that was designed to be faithful to what the international arbitration literature calls the “prima facie” analysis test as regards questions of fact and questions of mixed fact and law (para. 83). Under this test, the court must “refer the parties to arbitration unless the arbitration agreement is manifestly tainted by a defect rendering it invalid or inapplicable” (para. 75). To be so manifestly tainted, the invalidity must be “incontestable”, such that no serious debate can arise about the validity (para. 76, quoting Éric Loquin, “Compétence arbitrale”, in Juris-classeur Procédure civile (loose-leaf), fasc. 1034, at No. 105). Rather than adopting these standards literally, Dell gave practical effect to what was set out in the arbitration literature by creating a test whereby a court refers all challenges of an arbitrator’s jurisdiction to the arbitrator unless they raise pure questions of law, or questions of mixed fact and law that require only superficial consideration of the evidence in the record (paras. 84-85).
 The doctrine established in Dell is neatly summarized in its companion case, Rogers Wireless Inc. v. Muroff, 2007 SCC 35 (CanLII),  2 S.C.R. 921, at para. 11:
The majority of the Court held that, when an arbitration clause exists, any challenges to the jurisdiction of the arbitrator must first be referred to the arbitrator. Courts should derogate from this general rule and decide the question first only where the challenge to the arbitrator’s jurisdiction concerns a question of law alone. Where a question concerning jurisdiction of an arbitrator requires the admission and examination of factual proof, normally courts must refer such questions to arbitration. For questions of mixed law and fact, courts must also favour referral to arbitration, and the only exception occurs where answering questions of fact entails a superficial examination of the documentary proof in the record and where the court is convinced that the challenge is not a delaying tactic or will not prejudice the recourse to arbitration. The parties agree that the framework from Dell and Seidel applies to Ontario’s Arbitration Act. We agree, based on the similarities between the arbitration regimes in Ontario, British Columbia and Quebec. The two exceptions to arbitral referral in Dell and Seidel therefore apply in Ontario. This case, according to Mr. Heller, engages one of those exceptions because it requires at most only a superficial review of the record.
 Neither Dell nor Seidel fully defined what is meant by a “superficial” review. The essential question, in our view, is whether the necessary legal conclusions can be drawn from facts that are either evident on the face of the record or undisputed by the parties (see Trainor v. Fundstream Inc., 2019 ABQB 800, at para. 23 (CanLII); see also Alberta Medical Association v. Alberta, 2012 ABQB 113, 537 A.R. 75, at para. 26).
 Although it is possible to resolve the validity of Uber’s arbitration agreement through a superficial review of the record, we are of the view that this case also raises an issue of accessibility that was not raised on the facts in Dell and justifies departing from the general rule of arbitral referral. As Dell itself acknowledged, the rule of systematic referral of challenges to jurisdiction requiring a review of factual evidence applies “normally” (para. 85; see also Muroff, at para. 11). This is one of those abnormal times.
 The underlying assumption made in Dell is that if the court does not decide an issue, then the arbitrator will. As Dell says, the matter “must be resolved first by the arbitrator” (para. 84). Dell did not contemplate a scenario wherein the matter would never be resolved if the stay were granted. This raises obvious practical problems of access to justice that the Ontario legislature could not have intended when giving courts the power to refuse a stay.
 One way (among others) in which the validity of an arbitration agreement may not be determined is when an arbitration is fundamentally too costly or otherwise inaccessible. This could occur because the fees to begin arbitration are significant relative to the plaintiff’s claim or because the plaintiff cannot reasonably reach the physical location of the arbitration. Another example might be a foreign choice of law clause that circumvents mandatory local policy, such as a clause that would prevent an arbitrator from giving effect to the protections in Ontario employment law. In such situations, staying the action in favour of arbitration would be tantamount to denying relief for the claim. The arbitration agreement would, in effect, be insulated from meaningful challenge (see Jonnette Watson Hamilton, “Pre‑Dispute Consumer Arbitration Clauses: Denying Access to Justice?” (2006), 51 McGill L.J. 693; Catherine Walsh, “The Uses and Abuses of Party Autonomy in International Contracts” (2010), 60 U.N.B.L.J. 12; Cynthia Estlund, “The Black Hole of Mandatory Arbitration” (2018), 96 N.C. L. Rev. 679).
 These situations were not contemplated in Dell. The core of Dell depends on the assumption that if a court does not decide an issue, the arbitrator will.  Against these real risks of staying an action in favour of an invalid arbitration, one could pit the risk of a plaintiff seeking to obstruct an arbitration by advancing spurious arguments against the validity of the arbitration. This concern animated Dell (see paras. 84, 86).
 In our view, there are ways to mitigate this concern that make the overall calculus favour departing from the general rule of referring the matter to the arbitrator in these situations. Courts have many ways of preventing the misuse of court processes for improper ends. Proceedings that appear vexatious can be handled by requiring security for costs and by suitable awards of costs. In England, courts have awarded full indemnity costs where a party improperly ignored arbitral jurisdiction (Hugh Beale, ed., Chitty on Contracts (33rd ed. 2018), vol. II, Specific Contracts, at para. 32-065; A. v. B. (No.2),  EWHC 54 (Comm.),  1 All E.R. (Comm.) 633, at para. 15; Kyrgyz Mobil Tel Limited v. Fellowes International Holdings Limited  EWHC 1329, 2005 WL 6514129 (Q.B.), at paras. 43-44). Further, if the party who successfully enforced an arbitration agreement were to bring an action, depending on the circumstances they might be able to recover damages for breach of contract, that contract being the agreement to arbitrate (Beale, at para. 32-052; West Tankers Inc. v. Allianz SpA,  EWHC 854 (Comm.),  2 All E.R. (Comm.) 395, at para. 77).
 Moreover, Dell itself makes clear that courts may refer a challenge to arbitral jurisdiction to the arbitrator if it is “a delaying tactic”, or would unduly impair the conduct of the arbitration proceeding (para. 86). This provides an additional safeguard against validity challenges that are not bona fide.
 How is a court to determine whether there is a bona fide challenge to arbitral jurisdiction that only a court can resolve? First, the court must determine whether, assuming the facts pleaded to be true, there is a genuine challenge to arbitral jurisdiction. Second, the court must determine from the supporting evidence whether there is a real prospect that, if the stay is granted, the challenge may never be resolved by the arbitrator.
 While this second question requires some limited assessment of evidence, this assessment must not devolve into a mini-trial. The only question at this stage is whether there is a real prospect, in the circumstances, that the arbitrator may never decide the merits of the jurisdictional challenge. Generally, a single affidavit will suffice. Both counsel and judges are responsible for ensuring the hearing remains narrowly focused (Hryniak v. Mauldin, 2014 SCC 7 (CanLII),  1 S.C.R. 87, at paras. 31-32). In considering any attempt to expand the record, judges must remain alert to “the danger that a party will obstruct the process by manipulating procedural rules” and the possibility of delaying tactics (Dell, at para. 84; see also para. 86).
 As a result, therefore, a court should not refer a bona fide challenge to an arbitrator’s jurisdiction to the arbitrator if there is a real prospect that doing so would result in the challenge never being resolved. In these circumstances, a court may resolve whether the arbitrator has jurisdiction over the dispute and, in so doing, may thoroughly analyze the issues and record.
 Turning to the appeal before us, we would first observe that Mr. Heller has made a genuine challenge to the validity of the arbitration agreement. The clause is said to be void because it imposes prohibitive fees for initiating arbitration and these fees are embedded by reference in the fine print of a contract of adhesion. Second, there is a real prospect that if a stay is granted and the question of the validity of the Uber arbitration agreement is left to arbitration, then Mr. Heller’s genuine challenge may never be resolved. The fees impose a brick wall between Mr. Heller and the resolution of any of the claims he has levelled against Uber. An arbitrator cannot decide the merits of Mr. Heller’s contention without those — possibly unconscionable — fees first being paid. Ultimately, this would mean that the question of whether Mr. Heller is an employee may never be decided. The way to cut this Gordian Knot is for the court to decide the question of unconscionability.
 We would therefore resolve the arguments Mr. Heller has raised against the validity of Uber’s arbitration agreement rather than refer those arguments to arbitration in the Netherlands.
 We observe, incidentally, that departing from the general rule of arbitral referral in these circumstances has beneficial consequences. It will prevent contractual drafters from evading the result of this case through a choice of law clause. A choice of law clause could convert a jurisdictional question that would be one of law (and which therefore could be decided by the court) into a question as to the content of foreign law, which would require hearing evidence in order to make findings as to the content of foreign law, something that one would not ordinarily contemplate in a superficial review of the record.
 This is a significant loophole for contractual drafters to exploit. Indeed, Uber’s contract here includes a foreign choice of law clause. As the intervener Canadian Federation of Independent Business (“CFIB”) submitted, this Court should presume that Dutch law governs the question of whether the arbitration agreement is unconscionable because the contracts have a choice of law clause indicating Dutch law (I.F., at para. 34; see also Vita Food Products, Inc. v. Unus Shipping Co., 1939 CanLII 269 (UK JCPC),  A.C. 277 (P.C.), at pp. 289-91). Neither party, however, chose to lead evidence of Dutch unconscionability law. Since the parties chose not to lead evidence of Dutch law, this Court must address the issue of unconscionability according to Ontario law (see Pettkus v. Becker, 1980 CanLII 22 (SCC),  2 S.C.R. 834, at pp. 853-54; Das v. George Weston Limited, 2017 ONSC 4129, at para. 215 (CanLII)). If Uber had adduced evidence of Dutch law, then under the two exceptions to arbitral referral recognized in Dell, this Court would have had to grant the stay in favour of an arbitrator determining the unconscionability argument.
 As well, even though this case could have been resolved based on undisputed facts, such an approach may not be sustainable in future cases. An approach to arbitral referral that depends on undisputed facts would invite parties to dispute facts. Were that standard to apply, unreasonably disputing facts would allow a party to evade any review of the merits, by use of an arbitration clause. There would be no negative consequence, in this context, to a party unreasonably disputing facts if it meant the stay in favour of arbitration would be granted. This differs significantly from the standard civil litigation context, wherein unreasonable disputes as to facts can be deterred by costs awards.
 We turn finally to the validity of the arbitration agreement. As mentioned, Mr. Heller raised two independent arguments as to why the arbitration agreement with Uber is invalid: first, the clause is void for unconscionability; and, second, the clause is void because it contracts out of the ESA.
In Alectra Utilities Corporation v. Solar Power Network Inc. (Ont CA, 2019) the Court of Appeal discusses the integral element of jurisdiction in an arbitration court appeal:
The role of the court under the Arbitration Act, 1991
 The starting point in exercising the court’s role under the Arbitration Act, 1991 is the recognition that appeals from private arbitration decisions are neither required nor routine. Section 45 of the Act makes clear that the parties are free to establish or to preclude an appeal to the court on a question of law, fact, or a mixed question of law and fact. If the arbitration agreement is silent on this point, an appeal to the court on a question of law lies only with leave, which may be granted only if the court is satisfied that the matter is sufficiently important to justify an appeal and that determination of the question of law at issue will significantly affect the rights of the parties.
 In this case the parties – sophisticated commercial parties represented by counsel – chose not only to resolve their contractual dispute by arbitration rather than litigation, but also to preclude appeals to the court. Section 7.8(1) of the PAMA provides as follows:
There shall be no appeal from the determination of the arbitrator to any court. Judgment upon any award rendered by the Arbitrator may be entered in any court having jurisdiction thereof. There is no ambiguity here: there is no appeal to the court, period. The arbitrator’s determination is final and binding.
 Accordingly, the only basis for the respondent to challenge the award was under s. 46(1) of the Arbitration Act, 1991, which authorizes the court to set aside an arbitration award on the limited and specific grounds it enumerates. A court may set aside an arbitration award if:
1. A party entered into the arbitration agreement while under a legal incapacity. These grounds for setting aside an arbitration award are, in general, not concerned with the substance of the parties’ dispute. They concern issues such as the establishment and composition of the arbitration tribunal, compliance with Ontario law, and the requirements of procedural fairness.
2. The arbitration agreement is invalid or has ceased to exist.
3. The award deals with a dispute that the arbitration agreement does not cover or contains a decision on a matter that is beyond the scope of the agreement.
4. The composition of the arbitral tribunal was not in accordance with the arbitration agreement or, if the agreement did not deal with that matter, was not in accordance with this Act.
5. The subject-matter of the dispute is not capable of being the subject of arbitration under Ontario law.
6. The applicant was not treated equally and fairly, was not given an opportunity to present a case or to respond to another party’s case, or was not given proper notice of the arbitration or of the appointment of an arbitrator.
7. The procedures followed in the arbitration did not comply with this Act.
8. An arbitrator has committed a corrupt or fraudulent act or there is a reasonable apprehension of bias.
9. The award was obtained by fraud.
10. The award is a family arbitration award that is not enforceable under the Family Law Act.
 Although the court cannot apply s. 46(1)3 without having regard to an arbitrator’s decision, the court’s authority to set aside an arbitration award under that subsection depends on the mandate the arbitration agreement confers on the arbitrator to resolve a particular dispute. In order to succeed on an application to set aside an arbitration award, an applicant must establish either that the award deals with a dispute that the arbitration agreement does not cover or contains a decision on a matter that is beyond the scope of the arbitration agreement.
 For example, if an arbitration agreement provides that an arbitrator shall resolve a particular question and the arbitrator does so, the court has no authority to set aside the award on the basis that the arbitrator’s decision is unreasonable or incorrect. If, however, in the course of resolving the particular question remitted the arbitrator asks and answers an additional second question, the award may be set aside – not because the arbitrator’s answer to the second question is unreasonable or incorrect, but because the arbitrator had no authority to reach any conclusion on the second question at all.
 In short, s. 46(1)3 requires that arbitrators act within the bounds of the authority granted by the arbitration agreement pursuant to which they are appointed – no less, but no more. Section 46(1)3 is not an alternate appeal route and must not be treated as such.
The concept of jurisdiction
 The parties agree that s. 46(1)3 allows review for jurisdictional error – “true” jurisdictional error, as they describe it – but they disagree on what jurisdictional error is and how it is properly established. There should be no surprise in this. The difficulty in identifying jurisdictional error is well known.
 Jurisdictional error has a long and controversial history in Canadian administrative law that need not be recounted here. Suffice it to say that, historically, the courts often exercised their judicial review authority to interfere in matters that were committed by the legislature to determination by administrative tribunals. In C.U.P.E. v. N.B. Liquor Corporation, 1979 CanLII 23 (SCC),  2 S.C.R. 227 at p. 233, Dickson J. admonished courts not to do so:
The question of what is and is not jurisdictional is often very difficult to determine. The courts, in my view, should not be alert to brand as jurisdictional, and therefore subject to broader curial review, that which may be doubtfully so. But this proved easier said than done, and the concept of jurisdiction has continued to bedevil Canadian law despite Dickson J.’s admonition. That is so because, as Lord Denning explained in Pearlman v Keepers and Governors of Harrow School,  EWCA Civ 5, the distinction between an error that is jurisdictional in nature – which justifies judicial intervention – and an error made within jurisdiction – which does not – is so fine as to be manipulable. The same matter can be characterized as jurisdictional or non-jurisdictional depending on whether one seeks to intervene or defer.
 This is not the law in Ontario. The role of courts in addressing claims of jurisdictional error in the context of private arbitration is far more limited than the respondent would have it.