Court of Appeal for Ontario

Date: October 17, 2017

Docket: C61094

Judges: Gillese, MacFarland and Pepall JJ.A.

Between

Airia Brands Inc., Startech.Com Ltd. and QCS-Quick Cargo Service GMBH

Plaintiffs (Appellants)

and

Air Canada, AC Cargo Limited Partnership, Societe Air France, Koninklijke Luchtvaart Maatschappij N.V. dba KLM, Royal Dutch Airlines, Asiana Airlines Inc., British Airways PLC, Cathay Pacific Airways Ltd., Deutsche Lufthansa AG, Lufthansa Cargo AG, Japan Airlines International Co., Ltd., Scandinavian Airlines System, Korean Air Lines Co., Ltd., Cargolux Airlines International, Lan Airlines S.A., Lan Cargo S.A., Atlas Air Worldwide Holdings, Inc., Polar Air Cargo Inc., Singapore Airlines Ltd., Singapore Airlines Cargo PTE Ltd., Swiss International Air Lines Ltd., Quantas Airways Limited, and Martinair Holland N.V.

Defendants (Respondents)

Proceeding under the Class Proceedings Act, 1992, S.O. 1992, c. 6

Counsel

Paul Bates, Linda Visser and Emilie McLachlan Maxwell, for the appellants

Katherine L. Kay, Danielle K. Royal and James S.F. Wilson, for the respondents Air Canada and AC Cargo Limited

Robert S. Russell and Denes A. Rothschild, for the respondent British Airways PLC

Lisa La Horey and Brent Kettles, for the intervener Attorney General of Ontario

Heard: December 13 and 14, 2016

On appeal from: the order of Justice Lynne Leitch of the Superior Court of Justice, dated August 26, 2015, with reasons reported at 2015 ONSC 5332.

Pepall J.A.:

A. Overview

[1] This appeal concerns the applicable test for determining jurisdiction over absent foreign claimants ("AFCs") in a class action involving a claim of conspiracy to fix prices of air freight shipping services for shipments from or to Canada. The appellants allege that the airlines Air Canada, AC Cargo Limited Partnership ("Air Cargo"), and British Airways PLC ("BA") (collectively the "respondents") and other airlines, operating in locations all over the world, participated in a conspiracy to increase the price of air freight shipping services between the years 2000 and 2006. Specifically, they allege that the respondents and others conspired to limit or unduly lessen the supply of air freight shipping services, or to enhance unreasonably the price of associated fuel and security surcharges.

[2] The appellants sought an order certifying the class action with a class that included AFCs. The proposed class was to include persons who directly purchased air freight shipping services for shipments from or to Canada and indirect purchasers who purchased such services through a freight forwarder. The case is limited to the Canadian parameters of the Canadian conspiracy; that is to say, the illegal price fixing of fuel and security surcharges on air freight shipping services for shipments from or to Canada.

[3] In response, the respondents and others brought a motion for a declaration that the Ontario court did not have jurisdiction over AFCs and that the class should be defined to exclude such parties (the "jurisdiction motion"). For the purposes of the motion, AFCs were defined as persons who reside outside of Canada, entered into contracts for Air Freight Shipping Services outside of Canada, suffered any alleged losses outside of Canada, except those who expressly consent to the jurisdiction of the Ontario court.

[4] The respondents also served a notice questioning the constitutional applicability of the real and substantial connection test and ss. 27(3), 28(1), and 29(3) of the Class Proceedings Act, 1992 ("CPA"), which codify Ontario's opt out class proceedings regime, to AFCs.

[5] The respondents sought an order staying the proposed action insofar as it related to AFCs on the basis of jurisdiction simpliciter and alternatively, on the basis of forum non conveniens.

[6] The motion judge granted the respondents' jurisdiction motion, concluding that the real and substantial connection test for jurisdiction simpliciter should not be applied. Rather, the motion judge held that the jurisdictional analysis was to be guided by principles of order, fairness, and comity. She concluded that, because AFCs were not present in Ontario, or had not consented in some way to the jurisdiction of the court, jurisdiction had not been established. In the alternative, the motion judge held that there was no real and substantial connection and, in any event, Ontario should decline to exercise jurisdiction on the basis of forum non conveniens. In a subsequent order, she certified the action as a class proceeding, but excluded AFCs from the class.

[7] The appellants appeal from the motion judge's jurisdiction order. The Attorney General for Ontario has intervened in opposition to the motion judge's conclusion that the real and substantial connection test was inapplicable to AFCs.

[8] For the reasons that follow, I would allow the appeal. The motion judge erred in law in concluding that the real and substantial connection test was inapplicable and that jurisdiction existed only if AFCs were present in Ontario or consented to the Ontario court's jurisdiction. Moreover, she erred both in concluding that if the real and substantial connection test were to be applied, the test was not satisfied, and in her consideration of the principles governing forum non conveniens.

B. Facts

(1) Nature of the Claim

[9] Air freight shipping services are shipments of goods by air. Airlines typically sell their air freight shipping services to freight forwarders. Shipping customers in turn purchase these services from the freight forwarders. The freight forwarders in essence serve as intermediaries between the airlines and the shipping customers. The airlines may not know the identity of the shipper if a freight forwarder is used. Airlines also sell air freight shipping services directly to shipping customers. Air Canada, for example, has thousands of direct purchase shipper customers located throughout the world, but these direct purchasers represent only a small portion of Air Canada's customer base for air freight services.

[10] The appellants, the representative plaintiffs in this class action, include Airia Brands Inc., a Canadian manufacturer located in London, Ontario, and Startech.Com Ltd., a Canadian supplier also located in London, Ontario, both of whom purchased air freight shipping services from freight forwarders for the shipment of their products by air from and to Canada. QCS-Quick Cargo Service GMBH is a German freight forwarder that purchased air freight shipping services for shipments from and to Canada. The appellants assert a damages claim for conspiracy under the Competition Act, R.S.C. 1985, c. C-34. They also plead that the respondents are liable for the tort of civil conspiracy.

[11] The respondent airlines entered into transactions with customers, including AFCs, for shipments from and to Canada. Air Canada is a company domiciled in Canada. Air Canada's principal hub location for all its services, including air freight shipments, is Toronto (Pearson) International Airport. Air Cargo is a limited partnership that is controlled by Air Canada. Air Cargo provides air freight services. It has over 1,000 employees, about half of whom are located in Ontario. Approximately 50% of AC Cargo's Canadian cargo transportation revenue is generated from Toronto-based shipments. At the times material to this action, BA had cargo shipping operations located at Toronto (Pearson) International Airport. BA operated 16 direct flights per week from the U.K. to Toronto. Approximately 75% of its Canadian air cargo customers were located in the Greater Toronto Area.

[12] At the core of this class action is the issue of whether the respondents are, as the appellants allege, liable for conspiracy to fix prices for air freight shipping services, by imposing certain surcharges or extra fees on base prices for air freight shipping services. As mentioned, the appellants' claim is limited to the Canadian parameters of the global conspiracy; that is to say, the illegal price fixing of fuel and security surcharges on air freight shipping services for shipments from or to Canada. In furtherance of the conspiracy, the appellants allege that at least three in-person meetings occurred in Canada, of which at least one was in Ontario.

(2) Guilty Pleas and Settlements

[13] The conspiracy has been the subject of guilty pleas and judicial findings of guilt around the world. In Canada, the pleas generally related to outbound shipments only, but Cathay Pacific Airways Ltd. pled guilty to fixing surcharges on air freight shipping services to and from Canada.

[14] The action that is the subject matter of this appeal originally involved claims against other defendants, but those claims have been settled, leaving only three respondents remaining: Air Canada, Air Cargo, and BA. The first settlement by parties to the action was entered into in 2006. The class action as against Lufthansa AG, Lufthansa Cargo AG, and Swiss International Airlines Ltd. was certified on consent as a class proceeding in 2008 for settlement purposes. A global class was certified. An extensive notice program was undertaken worldwide, at a cost in excess of US$5 million. The notice program was to provide fair and adequate notice to class members by informing them about the litigation, the settlements, and the right to opt out. The notice program employed a number of notification methods, including mailed notice to approximately 310,000 persons, as well as notice by publication. The respondents approved the formal order made by the court. That said, the terms of the settlement did not bind them on any issue, including whether the action should be certified as a class proceeding.

[15] After that settlement, the action was certified on consent for settlement purposes with a number of other defendants who entered into settlement agreements with the appellants. The respondents did not oppose the certification orders related to the settlement agreements, which again certified a global class. These subsequent settlement agreements similarly contained express provisions that the terms of the settlement did not bind the respondents in any way.

(3) U.S. Class Proceeding

[16] Over a decade ago, a parallel class proceeding was commenced in the United States (the "U.S. class action"). The U.S. class action was certified on behalf of persons who purchased air freight shipping services for shipments directly from or to the United States: In re Air Cargo Shipping Services Antitrust Litigation, MDL No. 06-1775, 2014 WL 7882100 (E.D.N.Y. Oct. 15, 2014), report and recommendation adopted in full by 2015 WL 5093503 (E.D.N.Y. Jul. 10, 2015), at p. 1. The certified class is global in nature, including both domestic and foreign purchasers of air freight services.

[17] Prior to the U.S. certification motion, the defendants in the American proceeding brought a motion to dismiss the claims of the representative plaintiffs on a number of grounds. In an order resulting from that motion, the U.S. court maintained the claims for inbound and outbound shipments, but dismissed claims for shipments to, from, or within the European Union (excluding from or to the United States), because they were based on alleged violations of European law. The court declined to exercise jurisdiction over the claims brought under European law on the grounds of forum non conveniens and international comity. However it assumed jurisdiction of claims brought under U.S. law that involved shipments between the European Union and the United States. The court specifically considered whether foreign plaintiffs had standing to bring claims under the applicable U.S. law, holding that they did. See: In re Air Cargo Shipping Services Antitrust Litigation, MDL No. 06-1775, 2008 WL 5958061 (E.D.N.Y. Sept. 26, 2008) report and recommendation on these issues adopted by 2009 WL 3443405 (E.D.N.Y. Aug. 21, 2009), affirmed 697 F.3d 154 (2nd Circ 2012).

[18] In making its decision, the U.S. court rejected the defendants' argument that the conduct at issue – charging fixed prices for air cargo services – occurred outside the United States. The U.S. court stated it would take a broader view of the conduct relevant to the plaintiffs' claims, and considered not just where communications about the price fixing occurred, but also where the air cargo services themselves were rendered. It concluded that it was significant that air freight services are not rendered in one location, but are "performed along entire transportation routes, touching both the country of origin and the country of destination": In re Air Cargo (Sept. 26, 2008), at p. 13.

[19] If the appellants succeed on this appeal, the Canadian and U.S. class proceedings would be substantially aligned.

(4) The Motion Judge's Jurisdiction Reasons

[20] Before considering the appellants' certification motion, the motion judge addressed the jurisdiction motion brought by the respondents and three other airlines who have since settled the proceeding against them. She commenced her analysis by noting that the representative plaintiffs sought to certify the following class pursuant to their certification motion:

All persons (excluding Defendants, their respective parents, employees, subsidiaries, affiliates, officers, directors, persons currently resident in Australia who paid more than AUD$20,000 for the carriage of goods to and from Australia by air during the period January 1, 2000 to January 11, 2007, and persons who commenced litigation in a jurisdiction other than Canada prior to the conclusion of the trial of the common issues) who purchased Airfreight Shipping Services* during the period January 1, 2000 to September 11, 2006, including those persons who purchased Airfreight Shipping Services through freight forwarders, from any air cargo carrier, including, without limitation, the Defendants, but not including Integrated Air Cargo Shippers.**

*Airfreight Shipping Services means airfreight cargo shipping services for shipments to or from Canada (excluding shipments to and from the United States).

**Integrated Air Cargo Shipper is defined as an air cargo shipper that manages an integrated system of people, airplanes, trucks and all other resources necessary to move airfreight cargo from a customer's point of origin to the delivery destination, and for greater certainty, includes but is not limited to FEDEX, UPS, DHL and TNT.

[21] The class that the appellants sought to certify was subject to two exceptions: persons who commenced litigation outside of Canada prior to the conclusion of the trial of the common issues, and persons and shipments that fell within the scope of the proposed class actions that had been commenced in the United States and Australia.

[22] The motion judge identified two issues to be addressed: did the court have jurisdiction over AFCs and if so, should it be declined based on forum non conveniens.

[23] After describing the parties' positions, she first turned to the expert evidence proffered by them. At para. 112 she concluded, based on the affidavits presented, that the prevailing law outside of Canada is that jurisdiction is based on presence, consent, or submission; that is, parties may only become plaintiffs if they bring the claim themselves or join in an existing claim. At para. 113, she agreed with the original respondents "that the real and substantial connection test is a radical departure from the norms adhered to by other countries and the Ontario opt-out regime set out in s. 27(3) of the [CPA] cannot be applied outside of Ontario." At paras. 114-115 she accepted, based on the record before her, the respondents' argument that an Ontario class action judgment would not be recognized and enforced by a foreign court, exposing the respondents to the potential of double recovery by AFCs.

[24] Having made these findings, the motion judge then turned to the issue of jurisdiction, noting that the CPA, a procedural statute, could not create jurisdiction where there is none. She considered some of the governing jurisprudence. She distinguished Currie v. McDonald's Restaurants of Canada Ltd. (2005), 74 O.R. (3d) 321 (C.A.), which adopted the real and substantial connection test to address whether an Illinois court had jurisdiction over non-resident Ontario class members in the context of enforcement of a foreign judgment. In contrast, the case before her raised the question of whether the court should assume jurisdiction where foreign conflict of law principles precluded other countries from recognizing an Ontario judgment.

[25] The motion judge also considered Meeking v. Cash Store Inc., 2013 MBCA 81, 367 D.L.R. (4th) 684, where, in circumstances in which an Ontario court in a class action had territorial jurisdiction over both the defendant and the representative plaintiff, common issues between the claims of the representative plaintiff and that of non-resident plaintiffs were held to be a presumptive connecting factor sufficient to give the Ontario court jurisdiction over the non-resident plaintiffs. The Manitoba Court of Appeal had concluded that this did not constitute an unconstitutional expansion of the real and substantial connection test. In considering Meeking, the motion judge reasoned at para. 173 that Meeking had limited significance because it related to a national class action and, as such, the real and substantial connection test was consistent with the principles of order and fairness. Moreover, it was unnecessary to decide if Meeking was wrongly decided and contrary to Van Breda v. Village Resorts Ltd., 2012 SCC 17, [2012] 1 S.C.R. 572 "because the assertion of jurisdiction based on the sharing of a common issue is equivalent to asserting jurisdiction based on a party being a necessary or proper party, a notion rejected in Van Breda."

[26] The motion judge noted the distinction made by the Supreme Court in Van Breda between the real and substantial connection test applied as a conflict of laws rule and as a constitutional principle. In addition, while the Court affirmed that the appropriate test to determine whether a court has jurisdiction is the real and substantial connection test, it did not oust the traditional private international law basis for court jurisdiction.

[27] The motion judge concluded that Van Breda invited courts to develop approaches to jurisdiction to meet constitutional requirements and that recognize order and fairness. Having regard to the unique circumstances presented by the case – parties who are not present in Ontario, have not consented to Ontario's jurisdiction, and have not attorned to Ontario's jurisdiction – she concluded that the principles of fairness, order, and comity should guide the jurisdictional analysis, not the real and substantial connection test.

[28] In particular, the motion judge agreed with the respondents that the potential for re-litigation and the multiplicity of further actions by AFCs was inconsistent with the objectives of class proceedings and contrary to the principles of order and fairness. Further, asserting jurisdiction over AFCs where the court could not reasonably expect that its judgment would be recognized in foreign countries would offend comity.

[29] At paras. 139-140, she considered Ramdath v. George Brown College of Applied Arts and Technology, 2010 ONSC 2019, 93 C.P.C. (6th) 106, in which Strathy J. (as he then was) rejected the defendant's argument that "jurisdiction should follow recognition" – that is, the notion that an Ontario court, in analyzing whether it has jurisdiction simpliciter over a claim, must consider whether its judgment would likely be recognized in foreign jurisdictions. Strathy J. noted that holding otherwise would mean that Ontario courts would be deprived of jurisdiction even in the presence of an obvious real and substantial connection to Ontario. This, in turn, would allow defendants to oust the Ontario court's jurisdiction by "simply point[ing]" to another country that would not recognize a potential judgment. The motion judge distinguished Ramdath on the basis that, in Ramdath, there was no issue of order and fairness and it was conceded that there was a real and substantial connection between Ontario and the claims of the international claimants. Both the international students and the defendant would expect that their relationship claims would be litigated in Ontario and there was a finite number of international students from identified countries unlike this case. Moreover, before Strathy J., there was limited evidence from foreign law experts. She also observed that in McKenna, he reached a different conclusion on different facts.

[30] The motion judge addressed the ability of the appellants to rely on the Lufthansa notice program but held that it was irrelevant to the reasonable expectations of AFCs and she stated at para. 196 that "the fact of adequate representation of rights and procedural fairness becomes relevant only after a court concludes that it has jurisdiction."

[31] Finally, she concluded that the court was bound by the constitutional limits on its jurisdiction. At paras. 200-202, she stated:

[T]he territorial limits in s. 92 of the Constitution Act, 1867, prohibit the court from assuming jurisdiction over any class members who do not meet the traditional test of presence or consent recognized abroad, i.e. [AFCs].

The constitutional limits on the court's jurisdiction lead me to the conclusion that the real and substantial connection test ought not to be applied to establish jurisdiction over [AFCs].

Rather, I am satisfied that jurisdiction over class members can only be established if they are present in Ontario or have consented in some way to the jurisdiction of this court.

[32] As such, the court did not have jurisdiction over AFCs and the class action in relation to them was stayed. Having reached this conclusion, she declined to address the respondents' alternative argument that ss. 27(3), 28(1), and 29(3) of the CPA were constitutionally inapplicable to AFCs.

[33] The motion judge then addressed whether, if the real and substantial connection test did apply, it was satisfied in this case. In a brief discussion, she described the parties' arguments on the issue, rejected the new presumptive connecting factor described in Meeking and determined that the real and substantial connection test was not met. She did not discuss the issue of rebuttal of presumptive connecting factors.

[34] The motion judge also decided that if the court did have jurisdiction over the AFCs, such jurisdiction should be declined on the basis of forum non conveniens. She accepted the respondents' argument that in a multi-jurisdictional class action, the court should pay particular attention to whether its assumption of jurisdiction would be consistent with comity, prevailing international legal norms, and the reasonable expectations of the parties. The AFCs would have had no expectation that their rights would be adjudicated in Ontario. If they were included in the class, the Ontario court would have to apply the laws of at least 30 different countries in relation to matters that involve non-Canadians who entered into transactions outside of Canada. The evidence was that an Ontario judgment would not be recognized in other jurisdictions, reaffirming the risk of double recovery. The appellants' contention that the AFCs' claims would not be advanced anywhere at any time did not justify including those claims in the action. Accordingly, she stated that she would stay the proposed class action as it related to the AFCs on the basis that Ontario is forum non conveniens. She did not address whether an alternative forum was more appropriate.

C. Issues

[35] The issues for this court to consider on this appeal are:

(i) What is the applicable standard of review?

(ii) Did the motion judge err in her rejection of the real and substantial connection test for global class actions?

(iii) Did she err in her consideration of forum non conveniens?

[36] Before engaging in an analysis of these issues, I will first briefly address this court's jurisdiction to hear the present appeal. Section 30 of the CPA addresses the appropriate appeal routes for a number of types of orders and judgments made in class proceedings. The CPA however makes no reference to the appropriate avenue for appeal of jurisdiction orders. Where the CPA does not specifically address an avenue of appeal, s. 6(1)(b) of the Courts of Justice Act, R.S.O. 1990, c. C.43 governs whether an appeal in a class proceeding lies to this court: Locking v. Armtec Infrastructure Inc., 2012 ONCA 774, 30 C.P.C. (7th) 1, at para. 8; Cavanaugh v. Grenville Christian College, 2013 ONCA 139, 32 C.P.C. (7th) 1, at paras. 29-30; and Waldman v. Thomson Reuters Canada Ltd., 2015 ONCA 53, 71 C.P.C. (7th) 33, at paras. 5-7.

[37] The question therefore is whether the jurisdiction order under appeal is final and appealable to this court pursuant to s. 6(1)(b) or whether it is interlocutory. The jurisdiction order finally disposes of the issue of an Ontario court's jurisdiction over the AFCs, and therefore the present appeal is properly before this court.

D. Analysis

(1) Standard of Review

[38] This appeal addresses whether an Ontario court can take jurisdiction over class action proceedings involving AFCs and the appropriate test to be applied. This part of the analysis involves a question of law and therefore a correctness standard is applicable: Housen v. Nikolaisen, 2002 SCC 33, [2002] S.C.R. 235, at paras. 8-9, 27.

[39] In contrast to the identification of the correct legal test, the application of the correct legal test generally involves questions of mixed fact and law, which attract a standard of palpable and overriding error. However, where an error in the application of the correct legal test can be attributed to an extricable legal error – for example, the failure to consider a required element of a legal test, a mischaracterization of a legal test, or a similar error in principle – a standard of correctness applies: Housen, at paras. 33-36.

[40] This appeal also addresses the issue of forum non conveniens, which requires the application of the relevant legal test to the facts, and involves questions of mixed fact and law. Again, this issue is reviewable on a standard of palpable and overriding error, except where the motion judge's analysis involves an error that is traceable to an extricable question of law.

(2) Did the motion judge err in her rejection of the real and substantial connection test for global class actions?

(a) Positions of the Parties

[41] The appellants submit that the motion judge erred in numerous respects in rejecting the real and substantial connection test. They first argue that the operative test for determining jurisdiction simpliciter is that described in Currie, which incorporates the real and substantial connection test.

[42] Moreover, by holding that jurisdiction over the AFCs must be based on presence in the jurisdiction or consent to the jurisdiction, the motion judge replaced the real and substantial connection test with one based on imported foreign law; foreign law improperly became the determinant for jurisdiction. The appellants argue that factors of order and fairness described in Van Breda are subsumed by the real and substantial connection test and do not amount to a stand-alone test.

[43] The appellants additionally argue that, contrary to the motion judge's analysis, jurisdiction is not a function of enforcement. Her decision therefore conflicts with the decisions of Excalibur Special Opportunities LP v. Schwartz Levitsky Feldman LLP, 2016 ONCA 916, 135 O.R. (3d) 743, leave to appeal refused, [2017] S.C.C.A. No. 54; Ramdath, Nantais v. Telectronics Proprietary (Canada) Ltd. (1995), 25 O.R. (3d) 331 (Gen. Div.), leave to appeal refused (1995), 25 O.R. (3d) 331 at 347 (Gen. Div.); and Robertson v. Thomson Corp, (1999), 43 O.R. (3d) 161 (Gen. Div.), all of which provide that enforcement issues should not preclude certification of classes that include non-residents. In holding that the real and substantial connection test should be rejected based on concerns relating to recognition and enforcement, the motion judge allowed fairness to the respondents to overtake the analysis and become a higher priority than fairness to the AFCs. It was also an error to conclude that the prospect of re-litigation created order, fairness, and comity concerns, all the while ignoring that the risk of re-litigation was virtually non-existent due to the expiry of limitation periods.

[44] The respondents' position is that cases involving non-resident foreign class members in proposed class actions raise unique jurisdictional considerations. The motion judge did not err in concluding that it would be inappropriate to assert jurisdiction over the AFCs in light of: the unique and specific factual circumstances of the AFCs, the expert evidence that a decision of the Ontario court assuming jurisdiction over the AFCs on the basis of the real and substantial connection test would not be recognized, and the applicable principles of fairness and order. Contrary to the appellants' assertion, there is no rigidly defined Currie test. Currie supports the proposition that the court may examine both the real and substantial connection test and principles of order and fairness in determining whether jurisdiction should be assumed in an international class action. Subsequent decisions that the motion judge relied on in her reasons, such as McKenna v. Gammon Gold Inc., 2010 ONSC 1591, 88 C.P.C. (6th) 27, reversed on other grounds, 2011 ONSC 3782, 13 C.P.C. (7th) 232 (Div. Ct.); and Perell J.'s decision in Excalibur Special Opportunities LP v. Schwartz Levitsky Feldman LLP, 2014 ONSC 4118, 31 B.L.R. (5th) 46, have applied similar considerations.

[45] Furthermore, the respondents submit that in Van Breda, the Supreme Court recognized the distinction between the real and substantial connection test as a conflicts rule and as a constitutional principle. The Supreme Court elaborated on the test as a conflicts rule, including discussion of presumptive connecting factors and their potential for rebuttal, but declined to elaborate on the content of the constitutional principle, deferring that question to an appropriate case in the future. The respondents submit that the motion judge properly concluded that asserting jurisdiction over the AFCs would offend the constitutional requirements of order and fairness. This is because an Ontario judgment would not be recognized by foreign courts, creating a potential for double-recovery. Furthermore, AFCs would reasonably expect to bring their actions in their home jurisdiction.

[46] The Attorney General, in its capacity as intervener, submits that the motion judge erred in holding that jurisdiction existed only if the AFCs were present in Ontario or consented to the Ontario court's jurisdiction and in holding that the real and substantial connection test was constitutionally inapplicable.

[47] The Attorney General points to over 25 years of jurisprudence from the Supreme Court establishing that the real and substantial connection test governs whether a provincial superior court may assume jurisdiction over a dispute for both common law and constitutional purposes. Van Breda did not invite judges to devise their own test for jurisdiction. It held the opposite, expressly cautioning judges against resolving jurisdiction issues on a case-by-case, ad hoc basis. The common law conflicts rules must recognize Canadian constitutional, statutory, and common law requirements. The real and substantial connection test strikes an appropriate balance between these objectives.

[48] The Attorney General further argues that neither the real and substantial connection test, nor ss. 27(3), 28(1), and 29(3) of the CPA (the opt out provisions), are constitutionally inapplicable. The real and substantial connection test promotes the purposes of the CPA. In contrast, the motion judge's holding defeats those purposes.

[49] The Attorney General also challenges the factors considered by the motion judge in her analysis of jurisdiction. Consideration of foreign enforceability conflates the analysis of jurisdiction simpliciter with forum non conveniens. Requiring a plaintiff to prove that it can enforce a judgment before a court has jurisdiction reverses the onus and upends the analysis set out in Van Breda. The risk of duplicative litigation and double recovery are inherent in the nature of conflicts of laws problems and are not fatal to the assumption of jurisdiction. Risk is one of many factors a court may consider.

[50] Lastly, the Attorney General submits that the real and substantial connection test is a constitutional imperative. The presence or consent test unduly narrows the jurisdiction of provincial superior courts to deal with matters that may have a real and substantial connection to a province and may leave prospective class members who have suffered harm, which is real and connected to Ontario, without the ability to seek redress in Ontario.

(b) Discussion

(i) The Evolution of the Real and Substantial Connection Test

[51] As Lebel J. observed at para. 66 of Van Breda, striking a proper balance between flexibility and predictability, or between fairness and order, has been a constant theme in Canadian jurisprudence on jurisdiction.

[52] Starting in 1990 with the Supreme Court's decision in Morguard Investments Ltd. v. De Savoye, [1990] 3 S.C.R. 1077, and continuing with this court's decision in Excalibur, overturning Perell J.'s decision in 2014 ONSC 4118, and for which leave to appeal was refused, the real and substantial connection test has been a dominant, although not exclusive, test governing the issue of jurisdiction.

[53] In Morguard, the Supreme Court addressed jurisdiction in the context of extra-provincial enforcement of judgments. Justice La Forest noted at pp. 1108-1109 that the real and substantial connection approach to jurisdiction provides a reasonable balance between the rights of the parties. In Beals v. Saldanha, 2003 SCC 72, [2003] 3 S.C.R. 416, the Supreme Court expanded the application of the real and substantial connection test to the recognition of foreign judgments. Justice Major held at para. 23 that a substantial connection with the subject matter of the action will satisfy the test even in the absence of such a connection with the defendant in the action. Accordingly, a Canadian court could assume jurisdiction if the foreign court had a real and substantial connection with the subject matter of the action or the parties. Justice Major further noted at para. 37 that the presence of traditional indicia of jurisdiction such as attornment, consent, residence, and presence serve to bolster the real and substantial connection.

[54] In Van Breda, a case dealing with the assumption of jurisdiction in tort actions, Lebel J. observed that, in developing the real and substantial connection test, the Supreme Court had crafted a constitutional principle rather than a simple conflicts rule. Justice Lebel sought to draw a clearer distinction between the constitutional and private international law dimensions of the test. The constitutional dimension of the rule concerns the territorial limits of jurisdiction. The real and substantial connection test establishes boundaries within which conflicts rules could be applied. The purpose of the constitutionally-imposed territorial limit is to ensure the existence of the relationship or connection needed to establish the legitimate exercise of state power. A weak or hypothetical connection casts doubt on the legitimacy of the exercise of state power over the affected persons. In contrast, conflict rules decide when jurisdiction can be assumed over a dispute, what law will govern a dispute, or how an adjudicative decision from another jurisdiction will be recognized and enforced. As Lebel J. stated, at para. 33, "The constitutional territorial limits … are concerned with setting the outer boundaries within which a variety of appropriate conflicts rules can be elaborated and applied."

[55] In recognizing the need to develop an appropriate framework for the assumption of jurisdiction based on the real and substantial connection test, Lebel J. specified, at para. 73, that the framework for the assumption of jurisdiction cannot be "an unstable, ad hoc system made up on the fly on a case-by-case basis — however laudable the objective of individual fairness may be." He went on to say, at paras. 78-79:

In my view, identifying a set of relevant presumptive connecting factors and determining their legal nature and effect will bring greater clarity and predictability to the analysis of the problems of assumption of jurisdiction, while at the same time ensuring consistency with the objectives of fairness and efficiency that underlie this branch of the law.

From this perspective, a clear distinction must be maintained between, on the one hand, the factors or factual situations that link the subject matter of the litigation and the defendant to the forum and, on the other hand, the principles and analytical tools, such as the values of fairness and efficiency or the principle of comity. These principles and analytical tools will inform their assessment in order to determine whether the real and substantial connection test is met. However, jurisdiction may also be based on traditional grounds, like the defendant's presence in the jurisdiction or consent to submit to the court's jurisdiction, if they are established. The real and substantial connection test does not oust the traditional private international law bases for court jurisdiction.

[56] Fairness, efficiency, and comity inform the test but are not stand alone connecting factors. At para. 84, Lebel J. specifically excluded "general principles or objectives of the conflicts system, such as fairness, efficiency or comity" from the list of presumptive connecting factors, although he did note that these principles may influence the appropriateness of new factors.

[57] Justice Lebel outlined a list of presumptive connecting factors for torts. The onus is on a plaintiff to establish that one or more of the presumptive factors exists. If one or more of the presumptive connecting factors applies, the court should assume that it has jurisdiction over the claim. However, the presumption of jurisdiction is rebuttable. A defendant may rebut the presumptive effect of any of the factors to demonstrate that the court lacks jurisdiction. To do so, the defendant must establish, as outlined at para. 95, that the factor "does not point to any real relationship between the subject matter of the litigation and the forum", or that the relationship between the subject matter and the forum is weak.

[58] The Supreme Court set out the following four, non-exhaustive presumptive connecting factors for tort cases that, prima facie, entitle a court to assume jurisdiction over a dispute:

(a) the defendant is domiciled or resident in the province;

(b) the defendant carries on business in the province;

(c) the tort was committed in the province; and

(d) a contract connected with the dispute was made in the province.

[59] At para. 91, Lebel J. described the following considerations to assist in identifying new presumptive factors:

(a) similarity of the connecting factor with the recognized presumptive connecting factors;

(b) treatment of the connecting factor in the case law;

(c) treatment of the connecting factor in the statute law; and

(d) treatment of the connecting factor in the private international law of other legal systems with a shared commitment for order, fairness and comity.

[60] Justice Lebel also discussed whether a court would be limited in hearing only that part of a multi-jurisdictional claim that could be directly connected with the jurisdiction. In holding that a court would not be limited in such a manner, he explained at para. 99 that "[t]he purpose of the conflicts rules is to establish whether a real and substantial connection exists between the forum, the subject matter of the litigation and the defendant." Where a real and substantial connection exists "in respect of a factual or legal situation, the court must assume jurisdiction over all aspects of the case."

[61] Jurisdiction may therefore be based on traditional grounds such as presence or consent, or on the existence of a real and substantial connection. Moreover, as stated in Beals v. Saldanha, the former may bolster the latter.

[62] That traditional grounds are an appropriate basis for jurisdiction was reiterated in Chevron Corp. v. Yaiguaje, 2015 SCC 42, [2015] 3 S.C.R. 69, a case involving an action for recognition and enforcement of a foreign judgment in Ontario. In such an action, there is no need to establish a real and substantial connection between the enforcing forum and either the judgment debtor or the dispute.

[63] Of course, despite having jurisdiction, a court has discretion to decline to exercise its jurisdiction and to stay the proceedings based on the doctrine of forum non conveniens.

[64] For the purposes of this appeal, however, the first issue to be addressed is the thorny issue of the intersection of jurisdiction and class action proceedings, the topic I will next discuss.

(ii) Class Action Proceedings and the Unresolved Dilemma of Jurisdiction

Class Action Regime

[65] Having discussed the applicable principles for jurisdiction, it is helpful to return to first principles governing class actions.

[66] A class action is brought by a named plaintiff who serves as the representative plaintiff for a class of named or unnamed persons. The AFCs are an example of an unnamed class.

[67] In an article entitled "Are National Class Actions Constitutional?" (2010) 26 Nat'l J. Const. L. 279, Peter W. Hogg and S. Gordon McKee provide a concise outline, at p. 281, of the gatekeeping function of the court in a class action proceeding:

Before a class action can proceed to the trial of the common issues, it must be certified by the court. The court needs to be satisfied that the proceedings disclose a cause of action, that there is an identifiable class, that the class will be adequately represented by the representative plaintiff, that the claims of the class raise common issues, and that the class proceeding is the preferable procedure to adjudicate the common issues of the claims asserted in the action. If certification is denied, the class action will not proceed to trial, although the plaintiffs will then be free to bring individual actions against the defendant. Only if certification is granted, does the class action proceed to trial.

[68] Class actions were available at common law. As long ago as 1938, class action legislation was introduced in the United States. In Canada, the Province of Quebec was in the vanguard introducing class action legislation in 1978. Ontario followed in 1993. Now, the federal government and all provinces except Prince Edward Island have enacted class action legislation. As has frequently been described, there are three public policy purposes that have fueled the development of class actions: access to justice, behaviour modification, and judicial economy.

Expansive and Restrictive Approaches to Jurisdiction over AFCs in Class Actions

[69] Cases involving class actions raise unique jurisdictional challenges. Unlike traditional litigation, which consists of an identifiable plaintiff actively selecting a jurisdiction and hence consenting to the jurisdiction, class actions involve at least one identifiable representative plaintiff and frequently unidentified members of a larger class. Typically in litigation, the question is whether the Ontario court has jurisdiction over a defendant, not whether it has jurisdiction over AFCs. Indeed, the purpose of class actions is, in part, to confer a benefit on absent class members. It is difficult to reconcile class actions that include unidentified claimants with traditional approaches to jurisdiction. In those circumstances, a real and substantial connection on an individualized basis may be elusive. To allow for jurisdiction, either the members must be identified and present or consent to jurisdiction or there must be another doctrinal mechanism available to anchor jurisdiction.

[70] Two very different approaches to this jurisdictional dilemma have emerged, one expansive and the other more restrictive.

[71] One arm of the expansive approach focuses on the commonality of interest between the claims of resident and non-resident class members, as Tanya J. Monestier notes in her article "Personal Jurisdiction Over Non-Resident Class Members: Have I Gone Down the Wrong Road?" (2010) 45 Tex. Int'l L.J. 537, at pp. 546-548.

[72] Meeking provides an example of this more expansive approach. In Meeking, the Manitoba Court of Appeal focused on commonality between resident and non-resident class members to provide a basis for jurisdiction. In that analysis, the identity or confluence of an interest shared between non-resident plaintiffs and the representative plaintiff served to anchor jurisdiction in circumstances where the court had territorial jurisdiction over both the representative plaintiff and the defendant. McCutcheon v. The Cash Store (2006), 80 O.R. (3d) 644 (S.C.J.) (Meeking's Ontario counter-part); and Harrington v. Dow Corning Corp., 2000 BCCA 605, 82 B.C.L.R. (3d) 1, leave to appeal refused [2001] S.C.C.A. No. 21, both adopted the same approach.

[73] The expansive approach has attracted criticism. In particular, Professor Monestier argues that the common issues approach is artificial and also conflates the test for jurisdiction with the test for certification.

[74] Similarly, but without citing Meeking or the other aforementioned cases, Hogg and McKee dismiss an approach where a real and substantial connection for the entire class is established through a real and substantial connection between the representative plaintiff and the forum, coupled with commonality between the issues of the representative plaintiff and the rest of the class. They too argue that this conflates certification (which requires common issues) and the test for jurisdiction.

[75] Another arm of the expansive view of class action jurisdiction is reflected in American jurisprudence. In Phillips Petroleum Co. v. Shutts, 472 U.S. 797 (1985), the U.S. Supreme Court held that in a class action, so long as the defendant had a sufficient connection to the forum state, the forum court had jurisdiction over a plaintiff class that included non-resident persons with no connection to the forum state provided they had been given adequate notice plus an opportunity to be heard, an opportunity to opt out of the action, and there was adequate representation by the representative plaintiff.

[76] Justice Rehnquist reasoned at pp. 808-810 that non-resident plaintiffs are not in the same degree of jeopardy as out-of-state defendants because their interests are protected by the representative plaintiffs. Moreover, a court's inquiry at the certification stage helps ensure that those interests will be adequately represented and are sufficiently common with the interests of the representative plaintiffs. See also Ortiz Fibreboard v. Corp., 527 U.S. 815 (1999), at pp. 847-848; Carlough v. Amchem Prod., Inc., 10 F.3d 189 (3rd Cir. 1993), at p. 199.

[77] At p. 291 of their article, Hogg and McKee dismiss the approach taken in Phillips for Canada because it is based on due process rights rather than the doctrine of extraterritoriality applicable in this country. It bears noting that their conclusion serves to negate class actions with unidentified class members, be they national or international.

[78] It also bears noting that, although not identified as such, at its heart, the approach adopted by Sharpe J.A. in Currie is akin to the formulation set out in Phillips. As I later discuss in greater detail, Currie similarly held that jurisdiction over non-resident plaintiffs need not turn on whether they have opted into a class proceeding, so long as their interests are adequately represented and they are accorded sufficient procedural fairness, including adequate notice.

[79] In contrast, a more restrictive approach to the real and substantial connection test requires that there be a substantive connection, beyond common interests, between the non-resident class members and the adjudicating forum: Monestier, at p. 548. For example, in HSBC v. Hocking, 2008 QCCA 800, [2008] R.J.Q. 1189, the Quebec Court of Appeal held that a shared interest in the common issues was insufficient to establish a real and substantial connection where there was otherwise no connection to the jurisdiction.

[80] The difficulty with the restrictive approach, as Professor Monestier observes, is that it is almost impossible for a court to assert jurisdiction over a non-resident class member owing to a lack of actual connection between such a class member and the adjudicating forum. She notes, at pp. 551-552:

The fundamental problem with a restrictive approach to the real and substantial connection test is that it essentially undercuts the ability of the class action to act as a vehicle for the resolution of issues that transcend provincial borders and are perhaps best suited to being addressed in class form.

[81] An obvious problem created by the lack of a uniform approach to class action jurisdiction, as Hogg and McKee discuss at pp. 286-287, is that a defendant needs to be assured that it will not be exposed to further litigation on the same issue by non-resident members of the plaintiff class. If the enforcing court does not regard the adjudicating forum as having jurisdiction over the non-resident class members, the judgment will not be enforceable and nothing prevents the non-resident class member from proceeding with a claim in the enforcing forum. This creates concern about fairness to the defendant, who remains exposed to the risk of re-litigation. See also Monestier, at pp. 552-554.

Opt In and Opt Out Regimes

[82] One option that avoids the jurisdictional issues associated with the real and substantial connection test when it comes to non-resident class members is the implementation of opt in and opt out regimes.

[83] An opt in or opt out approach is reflected in certain statutes. In Quebec and Ontario, a class action judgment is binding on all members of the class unless they have opted out of the proceeding: Code of Civil Procedure, R.S.Q., c. C-25, art. 1027 and the CPA, s. 27(3). In British Columbia, in contrast, the governing statute provides for an opt out for class members who are residents of the province and an opt in provision for those who are resident outside of that province: Class Proceedings Act, R.S.B.C. 1996, c. 50, ss. 16(1), 16(2). Clearly a non-resident class member who has opted into the proceeding has consented to the jurisdiction of the court and is bound by the result.

[84] While Professor Monestier notes, at p. 558, that the opt in approach is doctrinally sound, it also "thwart[s] the policy objectives of class actions, such that they are no longer able to achieve the very goals for which they were designed."

[85] In a similar vein, Professor Janet Walker, in "Cross Border Class Actions: A View from Across the Border" (2004) 3 Mich. St. L. Rev. 755 observes that the opt in alternative may provide for an under-inclusive class. She states at p. 770:

First, to the extent that class actions are intended to have a regulatory affect by requiring market actors to internalize the costs of wrongful conduct, under-inclusive plaintiff classes mean that the costs internalized are less than the costs generated by the wrongful conduct … Second, to the extent that class actions are intended to facilitate compensation for wrongs suffered, under-inclusive plaintiff classes result in the failure of members of the plaintiff class to receive compensation … Finally, to the extent the class actions are intended to also bring closure to matters for defendants, the under-inclusiveness of plaintiff classes means that defendants will be left with unresolved claims that might be brought in other actions or in other fora.

[86] Opt out regimes also present certain challenges, as Hogg and McKee comment at p. 287:

We have noticed that the courts will insist that "sufficient notice" be given to the members of the plaintiff class, but so far the courts have not insisted that actual notice be given to every member of the class. Therefore, a class action judgment in Ontario (for example) may apply to some members of the plaintiff class who in fact know nothing about the proceedings brought on their behalf.

[87] Professor Monestier posits that it is necessary to rethink whether a real and substantial connection is needed to ground jurisdiction over a non-resident plaintiff class, noting at p. 539 that "[t]he test cannot be readily transposed to the separate question of whether a court has jurisdiction over an amorphous class of unnamed plaintiffs." She suggests that courts should instead reorient the analysis towards insuring that procedural safeguards are afforded to non-resident plaintiffs.

(iii) Jurisprudence on Class Actions and Jurisdiction

[88] In Ontario, the real and substantial connection test has been consistently applied to the question of jurisdiction in class actions. The first case to deal with the issue of jurisdiction over non-resident class members in Ontario was Currie, released in 2005. The particular issue in Currie was whether an Illinois judgment approving the settlement of a global class action against McDonald's was enforceable against Currie, an Ontario resident, so as to preclude him from launching his own class action in Ontario against McDonald's Canada.

[89] Justice Sharpe, writing for this court, noted at paras. 14-15 that legislation in several provinces specifically contemplated the inclusion of non-resident class members, and that strong policy reasons favoured the fair and efficient resolution of interprovincial and international class action proceedings. In appropriate cases, conflict of law rules should recognize the importance of having claims finally resolved in one jurisdiction. He highlighted one of the concerns associated with such enforcement that had been identified by Henry Paul Monaghan in "Antisuit Injunctions and Preclusion against Absent Nonresident Class Members" (1998) 98 Colum. L. Rev. 1148. The article describes, at pp. 1155-1156, the need to guard against potential abuses arising from class action defendants using the settlement of class action proceedings as a means to limit liability at "bargain-basement" prices.

[90] Justice Sharpe provided guidance on addressing this and other concerns arising from class actions that included non-resident class members. In doing so, he explained at para. 20 that the position of a class action plaintiff is not the same as that of a typical defendant, and that rules for recognition and enforcement of class action judgments should reflect those differences:

Class action regimes typically impose upon the court a duty to ensure that the interests of the plaintiff class members are adequately represented and protected. This is a factor favouring recognition and enforcement against unnamed class members.

[91] To address the concern for fairness, Sharpe J.A. held, at para. 25, that the court had to be satisfied with the procedures adopted in the proceeding:

Respect for procedural rights, including the adequacy of representation, the adequacy of notice and the right to opt out, could fortify the connection with [the foreign] jurisdiction and alleviate concerns regarding unfairness.

[92] He rejected the need for an opt in regime, noting that this would effectively negate meaningful class action relief. While he acknowledged that it may be easier to justify the assumption of jurisdiction in interprovincial cases than in international cases, a failure to opt out could be seen as a form of passive attornment sufficient to support jurisdiction over a non-resident, unnamed plaintiff, provided that certain conditions were met: a real and substantial connection links the cause of action to the foreign jurisdiction, the rights of non-resident class members are adequately represented, and non-resident class members are accorded procedural fairness, including adequate notice.

[93] Justice Sharpe ultimately determined, on the case before him, that because the unnamed plaintiffs were not afforded adequate notice, an Ontario court ought not recognize and enforce the foreign judgment against the Canadian representative class member and the other Canadian class members he sought to represent.

[94] The principles described in Currie were followed by Strathy J. in Ramdath and McKenna, both decided in 2010. In Ramdath, Strathy J. concluded that the court should take jurisdiction over non-resident students of George Brown College in a class action claim alleging that the College misrepresented the benefits of a certain programme. He held that, provided the conditions in Currie were met, neither foreign law evidence that a Canadian class action judgment might not be given preclusive effect in a foreign jurisdiction, nor the absence of actual notice, should preclude an Ontario court from taking jurisdiction over the entire class. There was a real and substantial connection between Ontario and the subject matter of the action, and procedural fairness could be provided to the plaintiff class. In McKenna, Strathy J. again applied the real and substantial connection test and the framework set out in Currie in assessing the court's jurisdiction over non-resident class members. He noted at para. 108:

Although Currie involved the enforcement in Ontario of a judgment in a foreign class action, the mirror image of the principles stated by the Court of Appeal are applicable to the exercise of jurisdiction by this court in a class action that seeks to include class members outside the jurisdiction.

[95] As noted earlier, in 2006 in McCutcheon, Cullity J. certified a class proceeding involving non-residents. He determined that he could bind non-resident class members to an Ontario proceeding because there was a sufficient real and substantial connection between non-resident class members and Ontario, and the principles of order and fairness were met. Justice Cullity concluded that a common interest between the class members served to establish the real and substantial connection. Similarly, In Baxter v. Canada (Attorney General), [2005] O.T.C. 391 (S.C.J.), Winkler J. (as he then was) observed that certified common issues in a class action could serve as a basis for jurisdiction over extra-provincial parties.

[96] These cases all pre-date Van Breda, which was released in 2012, and must now be read through the lens of that decision. As discussed earlier, Van Breda confirmed the real and substantial connection test as a constitutional imperative and a conflict of law rule, and founded presumptive connecting factors which, if present, establish jurisdiction.

[97] After Van Breda, in its 2013 decision in Meeking, the Manitoba Court of Appeal grappled with issues similar to those in this appeal except that it dealt with recognition and enforcement of an Ontario judgment against non-resident, unnamed plaintiffs. There, the court held at paras. 93 and 97 that:

The Canadian jurisprudence to which I have referred that would allow for common issues to be considered as a presumptive connecting factor in the real and substantial connection test is persuasive. Fairness to non-resident plaintiffs is achieved through the notification process and opt-out provisions, while, at the same time, the policy considerations favouring class actions described by McLachlin C.J.C. in Western Canadian Shopping Centres Inc. v. Dutton, 2001 SCC 46, [2001] 2 S.C.R. 534 are fulfilled. Further, the constitutional principle of federalism is respected.

Therefore, I would conclude that, in circumstances where the court has territorial jurisdiction over both the defendant and the representative plaintiff in a class action proceeding, common issues between the claim of the representative plaintiff and that of non-resident plaintiffs is a presumptive connecting factor, sufficient to give the court jurisdiction over non-resident plaintiffs.

[98] Subsequently, Abdula v. Canadian Solar Inc., 2015 ONSC 53, 126 O.R. (3d) 459, adopted the same approach at paras. 58-59.

[99] Finally and most recently, in 2016, this court addressed the application of the real and substantial connection test in Excalibur. At para. 34, MacFarland J.A. (Cronk J.A. concurring) affirmed that the test to determine whether to take jurisdiction over foreign class members "begins with an inquiry into jurisdiction simpliciter, on the principles set out in Club Resorts Ltd. v. Van Breda". She noted that Currie was not to be interpreted as standing for the proposition that an Ontario court should approach the issue of taking jurisdiction in a restrained manner. Furthermore, the reasonable expectations of the non-resident class members were not to be treated as an independent consideration in determining whether to take jurisdiction in a global class proceeding.

[100] In Excalibur, the representative plaintiff and the respondent accounting firm were both based in Toronto, but the remaining 56 plaintiffs were non-residents of Ontario. Justice MacFarland concluded that three of the four presumptive connecting factors identified in Van Breda were present. There were also no serious concerns of order and fairness as the identity of all but one class member was known and all class members could receive appropriate notice.

[101] In dissent, Blair J.A. reasoned that the motion judge had not disputed that the court had jurisdiction simpliciter and could take jurisdiction; rather the motion judge was addressing whether the Ontario court should take jurisdiction. Justice Blair held that the motion judge did not err in considering the reasonable expectations of the proposed non-resident class members or in taking an approach of "restraint" to the issue of whether to assume jurisdiction over the global class.

[102] The Supreme Court refused leave to appeal Excalibur in June of this year.

(iv) The motion judge erred in rejecting the real and substantial connection test

[103] In light of the foregoing, it is clear that the motion judge erred in simply anchoring her jurisdiction analysis in a negation of the traditional bases for jurisdiction, namely presence or consent, and in failing to apply the real and substantial connection test articulated in Van Breda. As explained in Van Breda, order, fairness, and comity are not independent roots of jurisdiction but are subsumed by the real and substantial connection test. As stated by Lebel J., the purpose of the constitutionally-imposed territorial limit is to ensure the existence of the relationship or connection needed to establish the legitimate exercise of state power.

[104] It is true, as noted by the motion judge and as stated by Lebel J. at para. 79 of Van Breda, that jurisdiction may be based on traditional grounds. This would include presence in the jurisdiction and consent to the jurisdiction. However, relying on those factors to ground jurisdiction is very different from rejecting jurisdiction based on their absence. The motion judge erred in doing the latter. Here the traditional grounds for jurisdiction were not available. As such, the real and substantial connection test had to be the start of the analysis.

[105] As recognized by Sharpe J.A. in Currie and the commentary discussed previously, the situation of unidentified non-resident class members in a proposed class action presents unique concerns with respect to ensuring that a sufficient connection exists to justify taking jurisdiction over them. In my view, this is what the Courts of Appeal in Ontario and Manitoba sought to address in Currie pre-Van Breda, and in Meeking, post-Van Breda, albeit in different ways. Moreover, it seems to me that a robust analysis may encompass the approach of both appellate courts.

[106] In setting out the contours of what the framework should be in these circumstances, I will address a preliminary issue raised during the appeal, which informs what is relevant to the assessment. I agree with the appellants that jurisdiction is not a function of foreign recognition and enforcement standing alone. Put differently, when addressing jurisdiction, issues of foreign recognition and enforcement are not preclusive of all other factors. Here, the motion judge permitted foreign law governing recognition to dominate her analysis to the exclusion of all other relevant factors.

[107] Turning to the appropriate framework, jurisdiction may be established over AFCs where:

1. there is a real and substantial connection between the subject matter of the action and Ontario, and jurisdiction exists over the representative plaintiff and the defendants;

2. there are common issues between the claims of the representative plaintiff and AFCs; and,

3. the procedural safeguards of adequacy of representation, adequacy of notice, and the right to opt out as described in Currie are provided, thereby serving to enhance the real and substantial connection between AFCs and Ontario.

[108] In my view, this framework provides the necessary safeguards to establish that jurisdiction properly exists and ensures the protection of the values of order and fairness. Quite apart from that consequence, a positive result of this framework is that the objectives at the heart of class actions are served.

[109] As the motion judge had already determined that the Ontario court did not have jurisdiction over the AFCs, her subsequent analysis of the issue of jurisdiction was very brief. She agreed at para. 210 with the original respondents that the new presumptive factor described in Meeking could not be recognized, as this would be inconsistent with Van Breda's guidance that, in identifying new presumptive factors, courts should consider the treatment of the connecting factor in the private international law of other legal systems having a shared commitment for order, fairness, and comity. Then, after describing both the appellants' and the respondents' positions, she concluded at paras. 213-214:

It cannot be said here as it was in Currie, Ramdath, McKenna and Excalibur for example, that it would come as no surprise to, or it would not be unreasonable from the perspective of, the [AFCs] that legal claims arising from their purchase of Airfreight Shipping Services outside Canada would be litigated in an Ontario court.

I am not satisfied that in these circumstances the real and substantial connection test is met.

[110] Apart from her discussion of the new presumptive factor in Meeking and the relevance of reasonable expectations, it is unclear whether the motion judge accepted all or some of the respondents' submissions and some or none of those of the appellants regarding the application of the presumptive connecting factors discussed in Van Breda. Given her rejection of the real and substantial connection test and in the absence of a consideration of the established presumptive connecting factors, it falls to this court to conduct the jurisdiction analysis anew applying the aforementioned analytical framework.

[111] Dealing with the first factor, unquestionably, there is a real and substantial connection between the subject matter of the action and Ontario. All of the three respondents carry on business in Ontario. As Gascon J. stated in Chevron Corp. at para. 89, "If a defendant maintains a business in Ontario, it is reasonable to say that the Ontario courts have an interest in the defendant and the disputes in which it becomes involved." Moreover, there is jurisdiction over the three respondents based on presence in the jurisdiction and over the three representative plaintiffs based on presence and consent to the jurisdiction. Airia, located in London, Ontario, purchased shipping services for transport from Ontario and Startech, also located in London, Ontario, purchased shipping services for transport from and to Ontario. QCS is a German freight forwarder that purchased shipping services to Canada from Air Canada and BA, and also purchased services for shipment from Canada using agents located in Ontario.

[112] The claim of the plaintiffs alleges the tort of conspiracy. There is also a claim for damages for breach of the Competition Act. Three specific meetings in furtherance of the conspiracy are alleged to have taken place in Canada, one of which was in Toronto on June 22, 1999. Although the appellants define the 'Relevant Period' for the conspiracy as beginning on or about January 1, 2000 and continuing until September 11, 2006, the Toronto meeting is pleaded as being in furtherance of the conspiracy. The respondents argue that the elements required to complete the tort of conspiracy did not occur in Canada. However, this is not determinative: see Central Sun Mining Inc. v. Vector Engineering Inc., 2013 ONCA 601, 117 O.R. (3d) 313. Moreover, the tortious conduct relates to air freight shipments linked to Canada, which served as either the shipments' country of origin or country of destination.

[113] Turning to the second factor, there are common issues between the representative plaintiffs over whom the court has jurisdiction and the AFCs. In her decision on the representative plaintiffs' related motion for certification, reported at 2015 ONSC 5352, the motion judge held that the representative plaintiffs' claim raised common issues among the members of the class she certified. The common issues certified by the motion judge relate to the core question of whether the respondents are liable to the class members for the tort of conspiracy and breaches of the Competition Act arising from the supply of airfreight shipping services to and from Canada in the relevant time period. These common issues clearly extend to the AFCs.

[114] Lastly, as a result of the settlements with some of the other defendants, the AFCs have been afforded the three procedural safeguards described in Currie. Notice has been effected in 140 countries and as mentioned, mailed notice has gone to 310,000 persons. The designed website had received more than 54,000 visitors from 135 countries as of November 2008 and the interactive telephone service had fielded 3,900 calls. As of January 2009, 270 class members from 33 different countries registered to receive further information about the Canadian litigation by mail. Each round of settlements involved a further notice campaign, including direct mail.

[115] The respondents argued before the motion judge that the notice program should not be considered as it amounted to an action to carry out the Lufthansa settlement that, based on the terms of the settlement, was not to be referred to or received in evidence in any proceeding. The motion judge chose not to resolve the issue as she was of the view that it could not have an effect on the determination of the jurisdiction motion and the notice program was irrelevant to the reasonable expectations of the AFCs vis-à-vis the respondents and could not prejudice them. There is no evidence that any motion to strike was brought by the respondents.

[116] I do not accept that the extent of and facts relating to notice would be captured by the terms of the Lufthansa settlement and that such evidence would be precluded from consideration in the proceeding against the respondents. In any event, even if they were, I would note that the CPA specifically provides for notice. Moreover, other significant protections are provided by the CPA. To ensure its fair and expeditious determination, a court may make any order under s. 12 that it considers appropriate regarding the conduct of the class action. The certification criteria in s. 5 provide that there be a representative plaintiff who would fairly and adequately represent the interests of the class among other matters. Pursuant to s. 29 of the CPA, a class action may only be discontinued or settled with court approval. Section 31 ensures that class members other than the representative plaintiffs are not liable for costs except in respect of the determination of their individual claims. Section 32 provides that the court must approve the fees of class counsel. These protections are all provided within the context of the gatekeeping function of the court in a class action proceeding as previously described in these reasons. In my view, there is a sufficient basis on which to ground jurisdiction over the AFCs.

[117] As in Abdula, this is not a case of a defendant being forced to defend a lawsuit in a jurisdiction in which it has no connection. I would also observe that assuming jurisdiction over the AFCs serves the objectives of class proceedings.

[118] Nor has the assumption of jurisdiction been rebutted. The respondents' business in Ontario includes the sale of services that are the subject matter of the alleged conspiracy. As mentioned, part of the conduct in furtherance of the alleged conspiracy took place in Ontario, and there are shared common issues between the representative plaintiffs and the AFCs.

[119] Accordingly, I would conclude that an Ontario court has jurisdiction over the AFCs in this class action proceeding.

(3) Did the motion judge err in her consideration of forum non conveniens?

[120] The motion judge also decided that jurisdiction should be declined on the basis of forum non conveniens. I will now turn to this issue.

[121] In Van Breda, at para. 110, Lebel J. set out a list of non-exhaustive factors to consider in the forum non conveniens analysis: the location of the parties and the witnesses; the cost of transferring the case to another jurisdiction or declining the stay; the impact of a transfer on the conduct of the litigation or on related or parallel proceedings; the possibility of conflicting judgments; problems related to recognition and enforcement of judgments; and the relative strength of the connection of the parties.

[122] The burden is on the respondents to show that another jurisdiction has a real and substantial connection to the claim and also the availability of a clearly more appropriate forum than Ontario: LaPointe Rosenstein Marchand Melançon LLP v. Cassels Brock & Blackwell LLP, 2016 SCC 30, [2016] 1 S.C.R. 851, at para. 52, and Van Breda at paras. 103-105. As stated in Van Breda at para. 104: "Forum non conveniens recognizes that there is a residual power to decline to exercise jurisdiction in appropriate, but limited, circumstances in order to assure fairness to the parties and the efficient resolution of the dispute." Consistent with the need to identify a clearly more appropriate forum than Ontario, the principle of comity informs the application of forum non conveniens.

[123] The motion judge accepted the respondent airlines' arguments that Ontario was forum non conveniens and suggested that AFCs pursue individual actions within their own jurisdictions. However, she failed to ask whether any jurisdiction was clearly more appropriate than Ontario. This was in error.

[124] I would note that, in concluding Ontario was not the appropriate forum for the AFCs' claims, the motion judge relied extensively on this court's decision in Kaynes v. BP, PLC, 2014 ONCA 580, 122 O.R. (3d) 162. In Kaynes, Sharpe J.A. held that the United States and United Kingdom were more appropriate forums than Ontario for a class proceeding involving a claim in misrepresentation arising out of share purchases, the bulk of which were made on foreign exchanges.

[125] An examination of the facts in Kaynes reveals why Kaynes is distinguishable from the present case. First, the appellant in Kaynes was able to point to specific alternative fora that were more appropriate than Ontario, which the respondents here have failed to do. Second, litigation had commenced in the U.S. with respect to the exact same claims that the representative plaintiff sought to litigate in Ontario, whereas the claims put forth in the present class proceeding, dealing with shipments from or to Canada, are not the subject of any other existing proceedings. Indeed, the representative plaintiffs seek to exclude those claims relating to shipments between Canada and countries with ongoing proceedings arising from the same alleged conspiracy. Third, the majority (over 99%) of the proposed class in Kaynes lacked a link to Canada, as they had purchased shares outside of Canada. In contrast, to be part of the class in the present proceeding, an AFC must have purchased air freight services from or to Canada. Lastly, the substantive claims in Kaynes rested in part on U.S. securities law, unlike the claims in the present case, which rest solely on Canadian law.

[126] I would also observe that in a subsequent decision of this court in Keynes v. BP, P.L.C., 2016 ONCA 601, 133 O.R. (3d) 29, the same panel decided that the stay that it had previously imposed on the Ontario action based on forum non conveniens was to be lifted. This was because it had become clear that the plaintiffs could not proceed with a claim in the proceedings in the United States as had previously been anticipated and the claim was now conceded to be governed by Ontario law. Another forum was not clearly more appropriate.

[127] In this case, there is no such forum. Furthermore, the evidence clearly demonstrates a robust connection between the parties and Ontario. All of the respondents carry on business in Ontario, as do at least two of the representative plaintiffs. To be part of the class, AFCs must have purchased air freight services from the respondents for shipments from or to Canada. Based on the respondents' own evidence, the majority of these services were rendered for shipments from or to Ontario. Some of the events in furtherance of the conspiracy claim unfolded in part in Ontario. Litigation relating to these same claims has not been brought in another jurisdiction and if it is in the future the plaintiffs are excluded from the class. The evidence of the respondents disclosed a minimum of 11,000 customers in Ontario for shipping services that arguably were impacted by the conspiracy.

[128] Moreover, the availability of class actions, contingency fees, and the ability to defer costs favour Ontario as the most efficient and cost-effective forum for resolution of the conspiracy claim. The motion judge's ruling has the effect of fracturing a single class of AFCs into countless possible actions brought by individual AFCs in their home jurisdictions. This outcome lacks the efficiency and cost-effectiveness provided by the present class proceedings, while creating the potential for a multiplicity of proceedings and increasing the likelihood for conflicting decisions.

[129] Section 28 of the CPA provides for the tolling of limitation periods. AFCs can only benefit from this particular provision if their claims are prosecuted in Ontario. Moreover, as the conspiracy in issue is alleged to have taken place between 2000 and 2006, if the AFCs are denied recourse in Ontario, their claims may very well be time-barred in other jurisdictions. This suggests that the risk of re-litigation is low thereby undermining the concerns of order and fairness raised by the respondents.

[130] The motion judge found that the court would have to apply the laws of at least 30 countries if AFCs were included in the class. In her decision on the related motion for certification, the motion judge held at para. 133 that the issue of whether foreign law would have to be applied to the litigation was "uncertain at this stage of the litigation". I agree that the extent to which foreign law will apply in these proceedings is unclear.

[131] First, the only pleading in which foreign law was pleaded was that of a defendant who has since settled the claims against it. Second, foreign law must be proven and in the absence of same, domestic law is applicable. Third, even if foreign law is considered to be applicable, it will have to be applied to those AFCs who consented or attorned to the jurisdiction in any event.

[132] Recognition and enforcement are two factors of many to consider. However, it may be dangerous to accord primacy to these factors, particularly in the face of the multitude of other factors that contribute to the inquiry of forum non conveniens. This danger is perhaps best illustrated by the evidence on the position taken by the respondent BA in related litigation in the U.K. There, it relies on orders made by the Canadian courts, including the court in Ontario, approving the Lufthansa settlement agreement, to preclude the litigation of claims in the U.K. Here, in essence, its position is to the contrary.

[133] In conclusion, there is no clearly more appropriate forum to resolve the AFCs' claims than Ontario, and the respondents have not met their onus in demonstrating otherwise.

(4) Other

[134] Lastly, in their notice of constitutional issue, the respondents posited that even if the court possessed adjudicative jurisdiction over AFCs based on the real and substantial connection test, the territorial limits of provincial legislative power in s. 92 of the Constitution Act would still prevent the court from applying ss. 27(3), 28(1), and 29(3) of the CPA as set out in Appendix A, attached hereto. Those subsections address opting out, suspension of limitation periods that would otherwise apply to those class members' claims, and settlements.

[135] The respondents essentially argue that these provisions of the CPA alter the substantive foreign laws that would otherwise apply to AFCs by suspending foreign limitation periods and ignoring foreign res judicata principles, which limit the binding effect of judgments to claimants who ask to bring a claim.

[136] The Attorney General responds that, as a procedural statute, the CPA does not confer, modify, or create substantive rights. Subsections 27(3) and 29(3) of the CPA, which bind all class members to a settlement or common issues judgment, including AFCs who have not opted out, cannot prevent a claim from being brought in courts outside the province. The provisions only have res judicata effect outside Ontario if and when a foreign court recognizes and enforces a settlement or judgment. Similarly, s. 28(1) does not suspend limitation periods in a foreign jurisdiction.

[137] The motion judge noted, at para. 116 of her reasons, that the CPA is a procedural statute and, citing Bisaillon v. Concordia University, 2006 SCC 19, [2006] 1 S.C.R. 666, that class action procedure does not alter the jurisdiction of courts nor does it create new substantive rights. That said, at para. 204, she stated that she need not address the issue of the constitutional inapplicability of those subsections of the CPA given her decision on the absence of jurisdiction over the AFCs.

[138] While it was open to the respondents to advance this argument in the alternative before this court, they took the position that this issue was not under appeal: footnote 62 of the respondents' factum. In all of these circumstances, I decline to address the issue of the constitutional applicability of these sections of the CPA.

E. Disposition

[139] For these reasons, I would allow the appeal. As agreed, I would order the respondents to pay the appellants $75,000 in costs of this appeal, inclusive of disbursements and taxes and order the intervener to bear its own costs. I would order that the matter of costs of the jurisdiction motion be remitted to the motion judge for determination.

Released: October 17, 2017

"S.E. Pepall J.A."

"I agree. E.E. Gillese J.A."

"I agree. J. MacFarland J.A."

Appendix A

Class Proceedings Act, 1992, S.O. 1992, c. 6

Subsection 27(3) A judgment on common issues of a class or subclass binds every class member who has not opted out of the class proceeding, but only to the extent that the judgment determines common issues that,

(a) are set out in the certification order;

(b) relate to claims or defences described in the certification order; and

(c) relate to relief sought by or from the class or subclass as stated in the certification order.

Subsection 28(1) Subject to subsection (2), any limitation period applicable to a cause of action asserted in a class proceeding is suspended in favour of a class member on the commencement of the class proceeding and resumes running against the class member when,

(a) the member opts out of the class proceeding;

(b) an amendment that has the effect of excluding the member from the class is made to the certification order;

(c) a decertification order is made under section 10;

(d) the class proceeding is dismissed without an adjudication on the merits;

(e) the class proceeding is abandoned or discontinued with the approval of the court; or

(f) the class proceeding is settled with the approval of the court, unless the settlement provides otherwise.

Subsection 29(3) A settlement of a class proceeding that is approved by the court binds all class members.