Supreme Court of Canada

SUPREME COURT OF CANADA

Appeal Heard: November 29, 2017 Judgment Rendered: June 6, 2018 Docket: 37202

Between:

Haaretz.com, Haaretz Daily Newspaper Ltd., Haaretz Group,

Haaretz.co.il, Shlomi Barzel and David Marouani

Appellants

and

Mitchell Goldhar

Respondent

• and -

Samuelson-Glushko Canadian Internet Policy and Public Interest Clinic

Intervener

Coram: McLachlin C.J. and Abella, Moldaver, Karakatsanis, Wagner, Gascon, Côté, Brown and Rowe JJ.

Reasons: (paras. 1 to 98)

Côté J. (Brown and Rowe JJ. concurring)

Concurring Reasons: (paras. 99 to 103)

Karakatsanis J.

Concurring Reasons: (paras. 104 to 143)

Abella J.

Concurring Reasons: (paras. 144 to 150)

Wagner J.

Joint Dissenting Reasons: (paras. 151 to 240)

McLachlin C.J. and Moldaver and Gascon JJ.

Haaretz.com v. Goldhar, 2018 SCC 28, [2018] 2 S.C.R. 3

Haaretz.com, Haaretz Daily Newspaper Ltd.,

Haaretz Group, Haaretz.co.il, Shlomi Barzel and

David Marouani Appellants

v.

Mitchell Goldhar Respondent

and

Samuelson‑Glushko Canadian Internet Policy and

Public Interest Clinic Intervener

Indexed as: Haaretz.com v. Goldhar

2018 SCC 28

File No.: 37202.

2017: November 29; 2018: June 6.

Present: McLachlin C.J. and Abella, Moldaver, Karakatsanis, Wagner, Gascon, Côté, Brown and Rowe JJ.

on appeal from the court of appeal for ontario

Private international law — Choice of forum — Court having jurisdiction — Forum non conveniens — Libel action commenced in Ontario in respect of statements published in Israeli newspaper available electronically in Canada — Defendants bringing motion to stay action on grounds that Ontario court lacks jurisdiction or, alternatively, that Ontario court should decline to exercise its jurisdiction on basis of forum non conveniens — Whether situs of tort is reliable basis on which to presume real and substantial connection between chosen forum and subject matter of litigation in Internet defamation cases — If so, whether presumption of jurisdiction can be rebutted — Whether choice of law factor in forum non conveniens analysis for Internet defamation cases should be based on place where plaintiff suffered most substantial harm to reputation.

G is a prominent Canadian businessman who also owns one of the most popular professional soccer teams in Israel. H is Israel’s oldest daily newspaper, which is published in print and online. H published an article about G, which the latter alleges to be libellous. The main subject of the article is G’s ownership and management of his Israeli soccer team, but the article also references his Canadian business and his approach to management. While the article was not distributed in print form in Canada, it was available electronically. G commenced an action for libel in Ontario, alleging damage to his reputation. H brought a motion to stay the action, arguing that Ontario courts lacked jurisdiction or, alternatively, that Israel was a clearly more appropriate forum. The motion judge dismissed H’s motion, finding that Ontario courts had jurisdiction and refusing to decline to exercise this jurisdiction in favour of Israeli courts. A majority of the Ontario Court of Appeal dismissed H’s appeal.

Held (McLachlin C.J. and Moldaver and Gascon JJ. dissenting): The appeal should be allowed and the motion to stay the action granted.

Per Côté, Brown and Rowe JJ.: While multijurisdictional defamation claims are not new, the exponential increase in multijurisdictional publications over the Internet has led to growing concerns about libel tourism and the possible assumption of jurisdiction by an unlimited number of forums. The current rules for the assumption and exercise of jurisdiction are able to address these challenges so long as the underlying principles of stability and fairness are kept in mind. In this case, while the motion judge properly determined that he had jurisdiction (under the jurisdiction simpliciter test), he committed multiple errors in his forum non conveniens analysis. On a robust and careful assessment of the relevant factors tainted by these errors, Israel is a clearly more appropriate forum.

Central to a proper understanding of the conflicts rules of Canadian private international law is an appreciation of the distinct roles played by jurisdiction simpliciter and forum non conveniens , and how these must be understood and analysed as a cohesive whole. The jurisdiction simpliciter analysis is meant to ensure that a court has jurisdiction. This will be the case where a real and substantial connection exists between a chosen forum and the subject matter of the litigation. This test prioritizes order, stability and predictability by relying on objective connecting factors for the assumption of jurisdiction. The forum non conveniens analysis, on the other hand, is meant to guide courts in determining whether they should decline to exercise that jurisdiction in favour of a clearly more appropriate forum. This doctrine emphasizes fairness and efficiency by adopting a case-by-case approach to this question.

At the jurisdiction simpliciter stage, in determining whether a real and substantial connection exists between a chosen forum and the subject matter of the litigation, a court must first consider whether the existence of a recognized presumptive connecting factor has been established. The situs of the tort, which is one such recognized factor, is a reliable basis on which to presume a real and substantial connection, even in Internet defamation cases. Raising doubt as to the value of the situs of the tort as a presumptive connecting factor in such cases, because of the ease with which publication can be established, would significantly undermine the objectives of predictability and order at the jurisdiction simpliciter stage. Concerns relating to the insufficiency of a presumptive connecting factor should be addressed either at the rebuttal stage of the jurisdiction simpliciter analysis or during the forum non conveniens analysis.

In this case, the tort of defamation was committed in Ontario, and therefore a presumptive connecting factor has been established. As a result, the Court must consider whether H has successfully rebutted the presumption. The ability to rebut the presumption of jurisdiction where there is only a weak relationship between the subject matter of the litigation and the forum serves as an important check on jurisdiction. A careful examination of this question is therefore of particular importance in Internet defamation cases, where a presumptive connecting factor can easily be established. Presumptive connecting factors must not give rise to an irrebuttable presumption of jurisdiction. In order for a defendant to succeed in challenging jurisdiction, the circumstances must demonstrate that the relationship between the forum and the subject matter of the litigation is such that it would not be reasonable to expect that the defendant would be called to answer proceedings in that jurisdiction. Assuming that these principles are properly applied, the situs of the tort will not give rise to an irrebuttable presumption of jurisdiction in Internet defamation cases. In the case at bar, H could have reasonably expected to be called to answer a legal proceeding in Ontario. As such, the presumption of jurisdiction is not rebutted.

At the forum non conveniens stage, the burden is on the defendant to satisfy the motion judge that the alternative forum is clearly more appropriate. While the normal state of affairs favours exercising jurisdiction in the forum where it is properly assumed, this should never come at the cost of one party facing unfair or clearly inefficient proceedings. Given the ease with which jurisdiction may be established in a defamation case, in a motion for a stay, a judge must conduct a robust and carefully scrutinized review of the issue of forum non conveniens . The establishment of a presumptive connecting factor is virtually automatic in Internet defamation cases. As the rebuttal stage of the jurisdiction simpliciter analysis fails to address all the consequences of this fact, it is appropriate for motion judges to be particularly attuned to concerns about fairness and efficiency during the forum non conveniens analysis in these types of cases. This should not be understood as imposing a different standard or burden for defamation cases.

As the forum non conveniens analysis is inherently factual in nature, courts of appeal should not normally interfere with a motion judge’s factual findings. However, there are limits to deference. Where the motion judge has erred in principle, misapprehended or failed to take account of material evidence, or reached an unreasonable decision, courts of appeal may intervene. In the case at bar, the motion judge committed several errors, which tainted his forum non conveniens analysis on each of the factors they affected as well as his overall weighing of these factors. As a result, no deference should be afforded to these aspects of the motion judge’s analysis and the Court may intervene.

Ultimately, H has established that holding a trial in Israel would be fairer and more efficient. Israel is clearly the more appropriate forum. A robust and careful forum non conveniens analysis of the relevant factors indicates that H would face substantial unfairness and inefficiency if a trial were held in Ontario. Comparative convenience and expense for the parties and comparative convenience and expense for the witnesses favour Israel. Loss of legitimate juridical advantage favours a trial in Ontario, but this factor should not weigh too heavily in the analysis. Fairness favours Israel, namely in view of G’s significant business interest and reputation in that country and the significant unfairness that a trial in Ontario would impose on H. Enforcement slightly favours Israel as H has no presence or assets in Ontario. Finally, while applicable law, as determined by the lex loci delicti principle — the place where the tort occurs —, favours Ontario in this case, this factor should be accorded little weight in the forum non conveniens analysis in cases where jurisdiction is established on the basis of the situs of the tort. In those circumstances, lex loci delicti will inevitably also point to the chosen forum on the question of applicable law.

This would not be an appropriate case to adopt the place of most substantial harm to reputation test for choice of law instead of lex loci delicti . Although in Internet defamation actions, where a tort may have occurred in multiple jurisdictions, the lex loci delicti rule may allow courts in multiple forums to assume jurisdiction and apply their own law, the Court should be reluctant to make such changes to the existing private international law framework, as this may create legal uncertainty in a manner contrary to the objectives of conflicts rules.

Per Karakatsanis J.: There is agreement with Côté J.’s conclusion and much of her reasoning. However, there is disagreement with two aspects of her analysis relating to forum non conveniens . When considering the applicable law factor, assessing what law would apply in the alternative jurisdiction is not helpful, as the ultimate question that motivates this factor is whether the plaintiff’s chosen jurisdiction would be applying foreign law. Further, G’s Israeli reputation is not material to the fairness factor, which is concerned with the plaintiff’s interest in vindicating his reputation in the jurisdiction where he enjoys it. Ultimately, the overall conclusion reached by Côté J. on forum non conveniens does not turn on any of these elements, and therefore, the appeal should be allowed.

Per Abella J.: There is agreement with Côté J. that the appeal should be allowed. However, there is disagreement with her that the lex loci delicti rule should continue to serve as the basis for choice of law under the forum non conveniens analysis in cases of multijurisdictional Internet defamation. This standard approach to choice of law does not adequately respond to the unique issues and challenges raised by Internet defamation, where a single download can determine which law applies under a strict application of the lex loci delicti rule.

The framework for choice of law should therefore be modified by replacing lex loci delicti with a test based on the place where the most substantial harm to the plaintiff’s reputation occurred. This new approach would narrow the range of potentially applicable law in a rational way and would displace the law of the place of publication of the defamation with the law of the place with the most significant connection to the tort. It would also ensure that the choice of law rule reflects protection of reputation, which is at the core of the tort of defamation, and that the reasonable expectations of the publisher of the statement alleged to be defamatory as to where it could expect to be sued are properly considered, while at the same time striking a better balance between freedom of expression and harm to reputation.

Since there are symmetrical concerns between how the choice of law analysis proceeds and how jurisdiction is determined in Internet defamation cases, the same approach should be applied to determining jurisdiction. The current approach seems to make the assumption of jurisdiction automatic based on a single download. Since the essence of the harm in defamation is damage to reputation, the framework for determining jurisdiction should focus on where the plaintiff suffered the most substantial harm to his or her reputation. Such an approach allows the presumption of jurisdiction to be rebutted if the defendant can show that the most harm to the plaintiff’s reputation occurred elsewhere.

Adopting the most substantial harm test for determining the choice of law under the forum non conveniens analysis, the place of most substantial harm to G’s reputation is clearly Israel, and as a result, Israeli law should apply. The article in question is essentially about G and his conduct in Israel: it was about G’s soccer team, one of Israel’s most popular soccer teams, G’s involvement in his team’s management, and G’s relationship with his players, coaches and trainers in Israel. It was researched, written and edited in Israel, addressed to an Israeli audience, and focused on someone who is a public figure there. Although G spends most of his time in Canada, he maintains an apartment in Israel and his connection to Israel is significant. Accordingly, the article would have a far greater impact on his reputation in Israel than in Canada.

As for the rest of the forum non conveniens analysis, on the basis that Israeli law applies, there is agreement with Côté J. that Israel is the clearly more appropriate forum. All of the remaining factors — the comparative convenience/expense to the parties and witnesses, juridical advantage, fairness and enforcement — favour Israel.

Per Wagner J.: There is agreement with Côté J. that the appeal should be allowed. However, as set out in the reasons of Abella J., the choice of law rule during the forum non conveniens analysis should be modified for the tort of Internet defamation, from lex loci delicti to a test based on the place where the most substantial harm to the plaintiff’s reputation occurred. Although it may be that in certain cases it would be challenging to identify the place of most substantial reputational harm, the range of possibly applicable law for a given dispute would be much narrower than with lex loci delicti and would be determined on a more principled basis. Adopting this new test for choice of law would have several positive effects and would not result in a heavy evidentiary burden for the parties.

With respect to the jurisdiction simpliciter analysis, a Canadian court should not conclude that it does not have jurisdiction over a dispute with significant connections to Canada, including potentially significant reputational harm suffered in Canada, simply because greater reputational harm occurred elsewhere. As a result, concerns raised by the unique nature of Internet defamation are best addressed by changes to the choice of law rule, rather than by changes to the jurisdiction simpliciter stage of the analysis. The inquiry at that stage is simply whether there is a real and substantial connection between the dispute and the Canadian forum, not whether this connection is greater than that between the dispute and any other forum. There is no reason why this should be different in the context of Internet defamation.

In this case, when the most substantial harm test is applied to the facts, Israel is the clearly more appropriate forum.

Per McLachlin C.J. and Moldaver and Gascon JJ. (dissenting): The appeal should be dismissed. When a Canadian citizen is allegedly defamed for his Canadian business practices — in an article published online in his home province by a foreign newspaper — he is entitled to vindicate his reputation in the courts of the province where he lives and maintains his business, and where the sting of the article’s comments is felt.

The current rules that govern the application of the test for jurisdiction simpliciter readily accommodate multijurisdictional defamation cases, even in the Internet age. The commission of a tort in the jurisdiction remains a sound presumptive connecting factor on which to establish prima facie jurisdiction even in the context of Internet defamation cases, because the sting of the defamation is felt in the place where it is read. In this case, it is not contested that the allegedly libellous article was consulted by 200 to 300 people in Canada; therefore, a tort of defamation was committed in Ontario. There is no valid reason to reconsider or set aside this clearly established presumptive connecting factor.

While a presumptive connecting factor may be established virtually automatically in Internet defamation cases, a court does not necessarily assume jurisdiction. If there is no real and substantial connection between the action and the forum, the presumptive connecting factor would be rebutted. Reasonable foreseeability is central to the rebuttal step of the analysis: the strength of the relationship between the subject of the litigation and the forum is informed by the reasonable foreseeability of the claim proceeding in that jurisdiction. Without this important check of reasonable foreseeability of being sued in the jurisdiction, the presumptive connecting factor of the commission of a tort in the jurisdiction could raise concerns of forum shopping. Reasonable foreseeability is therefore an important limit on the ease with which jurisdiction can be presumptively assumed in defamation cases, especially over the Internet. In the present case, it was more than reasonably foreseeable that H would be sued in Ontario. The article was highly critical of G’s management style, allegedly imported from his Canadian business. Furthermore, H made the article readily available to readers worldwide through online publication. It is entirely foreseeable that a Canadian citizen and resident would want to vindicate his Canadian reputation as the owner of his Canadian businesses in a Canadian court. Therefore, the presumption of jurisdiction was not rebutted and Ontario courts have jurisdiction. The facts undeniably reveal a real and substantial connection between this case and Ontario.

If the analysis at the rebuttal stage is done properly, with an adequate consideration of reasonable foreseeability, there is no need to apply a robust and carefully scrutinized forum non conveniens analysis, as suggested by Côté J. This new standard would frustrate the predictability and stability that is at the core of the applicable framework. The basis of the forum non conveniens analysis is the clearly more appropriate forum test, which sets a high threshold for displacing the forum chosen by the plaintiff. This purposefully stringent and consistently upheld threshold should not be lowered, whether through lenient application or through a robust and carefully scrutinized review. Furthermore, a motion judge’s discretionary decision whether or not to decline jurisdiction on the basis of forum non conveniens is entitled to considerable deference, and having appellate courts apply the proposed robust and carefully scrutinized approach would disregard the discretionary nature of forum non conveniens decisions. A motion judge’s exercise of his discretionary power or assessment of the evidence should not be interfered with where it is not tainted by any error, or when only tainted by errors that have no impact on the result.

In this case, an assessment of the factors in the forum non conveniens analysis indicates that they do not meet the test of showing that Israel is a clearly more appropriate forum than Ontario. Only the factor of comparative convenience and expense for the parties and witnesses favours Israel, and this only slightly so with respect to the witnesses. The enforcement of judgment factor does not weigh heavily in the analysis. The factor of loss of legitimate juridical advantage weighs in favour of Ontario, and, most importantly, the key factors of applicable law and fairness to the parties weigh heavily in favour of Ontario.

With respect to the applicable law factor, the most substantial harm test to determine the applicable law in multijurisdictional Internet defamation cases should not be adopted in place of lex loci delicti . Such a rule is highly subjective, and will not reliably point to one jurisdiction. It does not provide a clear answer where a person lives and maintains an important reputation in one jurisdiction, but acts — and is the subject of defamatory statements — in another jurisdiction. It would also lead to complex preliminary motions requiring substantial evidence which would increase delay and expense. In terms of the proper approach to balancing this factor in the forum non conveniens analysis, it is entirely appropriate for courts to only look at the chosen forum in determining the applicable law. Requiring courts to assess the choice of law rules of a foreign jurisdiction may require extensive evidence, needlessly complicating the pre-trial motion stage of the proceedings. Where jurisdiction is based on the situs of the tort, the applicable law under lex loci delicti will indeed point to the forum. This does not mean that the applicable law factor should be granted little weight in the forum non conveniens analysis; rather, giving due weight to this factor reflects the notion that a case should proceed in a forum that properly has jurisdiction over the matter unless another forum is clearly more appropriate. Holding that the applicable law should be given little weight ignores the importance of the territorial jurisdiction of the chosen forum, and distorts the forum non conveniens analysis in favour of the foreign jurisdiction.

The lex loci delicti rule directs courts to apply their domestic law after having found that the tort of defamation occurred within their jurisdiction. Defamation law is directed to the protection of reputation. For choice of law purposes, it is therefore logical that a court of a jurisdiction where publication occurred is entitled to apply its own law. This remains true even if a tort took place simultaneously in another jurisdiction. In this case, as the applicable law is that of Ontario, this factor strongly favours Ontario over Israel.

With respect to fairness, this factor, along with the efficient resolution of disputes, is the cornerstone of forum non conveniens . The Court has repeatedly emphasized the importance of plaintiffs being allowed to sue for defamation in the locality where they enjoy their reputation. In the instant case, G has a real and long-standing reputational interest in Ontario. His reputation in Israel is not material to the analysis. It is therefore not unfair that Ontario be the forum deciding the dispute.

Following the forum non conveniens analysis, Israel has not emerged as a forum that would be more appropriate than Ontario to hear the case, much less a clearly more appropriate forum.

Cases Cited

By Côté J.

Applied: Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572; Tolofson v. Jensen , 1994 44 (SCC) , [1994] 3 S.C.R. 1022; considered : Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636; Moore v. Bertuzzi , 2014 ONSC 1318 , 53 C.P.C. (7th) 237; referred to: Breeden v. Black , 2012 SCC 19 , [2012] 1 S.C.R. 666; Lax Kw’alaams Indian Band v. Canada (Attorney General) , 2011 SCC 56 , [2011] 3 S.C.R. 535; Muscutt v. Courcelles (2002), 2002 44957 (ON CA) , 60 O.R. (3d) 20; Charron Estate v. Village Resorts Ltd. , 2010 ONCA 84 , 98 O.R. (3d) 721; Hunt v. T&N plc , 1993 43 (SCC) , [1993] 4 S.C.R. 289; Crookes v. Newton , 2011 SCC 47 , [2011] 3 S.C.R. 269; Crookes v. Holloway , 2007 BCSC 1325 , 75 B.C.L.R. (4th) 316, aff’d 2008 BCCA 165 , 77 B.C.L.R. (4th) 201; Barrick Gold Corp. v. Blanchard & Co. (2003), 2003 64238 (ON SC) , 9 B.L.R. (4th) 316.

By Karakatsanis J.

Referred to: Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636; Breeden v. Black , 2012 SCC 19 , [2012] 1 S.C.R. 666.

By Abella J.

Considered: Tolofson v. Jensen , 1994 44 (SCC) , [1994] 3 S.C.R. 1022; Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572; referred to: Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636; eDate Advertising GmbH v. X , C‑509/09, C‑161/10, [2011] E.C.R. I‑10302; Beals v. Saldanha , 2003 SCC 72 , [2003] 3 S.C.R. 416; Moran v. Pyle National (Canada) Ltd. , 1973 192 (SCC) , [1975] 1 S.C.R. 393; Morguard Investments Ltd. v. De Savoye , 1990 29 (SCC) , [1990] 3 S.C.R. 1077; Lapointe Rosenstein Marchand Melançon LLP v. Cassels Brock & Blackwell LLP , 2016 SCC 30 , [2016] 1 S.C.R. 851.

By Wagner J.

Considered: Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636; r eferred to: Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572.

By McLachlin C.J. and Moldaver and Gascon JJ. (dissenting)

Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572; Lax Kw’alaams Indian Band v. Canada (Attorney General) , 2011 SCC 56 , [2011] 3 S.C.R. 535; Morguard Investments Ltd. v. De Savoye , 1990 29 (SCC) , [1990] 3 S.C.R. 1077; Breeden v. Black , 2012 SCC 19 , [2012] 1 S.C.R. 666; Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636; Paulsson v. Cooper , 2011 ONCA 150 , 105 O.R. (3d) 28; Barrick Gold Corp. v. Blanchard & Co. (2003), 2003 64238 (ON SC) , 9 B.L.R. (4th) 316; Lapointe Rosenstein Marchand Melançon LLP v. Cassels Brock & Blackwell LLP , 2016 SCC 30 , [2016] 1 S.C.R. 851; Amchem Products Inc. v. British Columbia (Workers’ Compensation Board) , 1993 124 (SCC) , [1993] 1 S.C.R. 897; Egbert v. Short , [1907] 2 Ch. 205 ; St. Pierre v. South American Stores (Gath and Chaves), Limited , [1936] 1 K.B. 382 ; Rockware Glass Ltd. v. MacShannon , [1978] 2 W.L.R. 362 ; Spiliada Maritime Corporation v. Cansulex Ltd. , [1987] 1 A.C. 460 ; Spar Aerospace Ltd. v. American Mobile Satellite Corp. , 2002 SCC 78 , [2002] 4 S.C.R. 205; Unifund Assurance Co. v. Insurance Corp. of British Columbia , 2003 SCC 40 , [2003] 2 S.C.R. 63; Tolofson v. Jensen , 1994 44 (SCC) , [1994] 3 S.C.R. 1022; Bou Malhab v. Diffusion Métromédia CMR inc. , 2011 SCC 9 , [2011] 1 S.C.R. 214; Jenner v. Sun Oil Co. , 1951 129 (ON SC) , [1952] 2 D.L.R. 526.

Statutes and Regulations Cited

Defamation Act 2005 (N.S.W.), s. 11(3).

Rules of Civil Procedure , R.R.O. 1990, Reg. 194, rr. 1.08, 47.01, Tariff A.

Authors Cited

Australia. Law Reform Commission. Unfair Publication: Defamation and Privacy . Canberra, 1979.

Blom, Joost, and Elizabeth Edinger. “The Chimera of the Real and Substantial Connection Test” (2005), 38 U.B.C. L. Rev. 373.

Brown on Defamation: Canada, United Kingdom, Australia, New Zealand, United States , 2nd ed. by Raymond E. Brown. Toronto: Carswell, 1994 (loose‑leaf updated 2017, release 5).

Canadian Encyclopedic Digest , vol. 10, Ontario 4th ed. Toronto: Carswell, 2009 (loose-leaf updated 2018, release 2).

Castel, J.-G. “Multistate Defamation: Should the Place of Publication Rule be Abandoned for Jurisdiction and Choice of Law Purposes?” (1990), 28 Osgoode Hall L.J. 153.

Castel, Jean-Gabriel. “The Uncertainty Factor in Canadian Private International Law” (2007), 52 McGill L.J. 555.

Castel, Matthew. “Jurisdiction and Choice of Law Issues in Multistate Defamation on the Internet” (2013), 51 Alta. L. Rev. 153.

Cruz Villalón, Pedro. Opinion of Advocate General Cruz Villalón , C-509/09, C-161/10, [2011] E.C.R. I-10272.

Dicey, Morris and Collins on the Conflict of Laws , 15th ed. by Lord Collins of Mapesbury. London: Sweet & Maxwell/Thomson Reuters, 2012.

Downard, Peter A. The Law of Libel in Canada , 4th ed. Toronto: LexisNexis Canada, 2018.

Kain, Brandon, Elder C. Marques and Byron Shaw. “Developments in Private International Law: The 2011‑2012 Term — The Unfinished Project of the Van Breda Trilogy” (2012), 59 S.C.L.R. (2d) 277 .

Law Commission of Ontario. Defamation Law in the Internet Age: Consultation Paper , Toronto, 2017 (online: http://www.lco-cdo.org/wp-content/uploads/2017/12/Defamation-Consultation-Paper-Eng.pdf; archived version: http://www.scc-csc.ca/cso-dce/2018SCC-CSC28_1_eng.pdf).

Martin, Craig. “ Tolofson and Flames in Cyberspace: The Changing Landscape of Multistate Defamation” (1997), 31 U.B.C. L. Rev. 127.

Pitel, Stephen G. A., and Nicholas S. Rafferty. Conflict of Laws , 2nd ed. Toronto: Irwin Law, 2016.

Schmitz, Sandra. “From Where are They Casting Stones? — Determining Jurisdiction in Online Defamation Claims” (2012), 6 Masaryk U. J.L. & Tech. 159.

APPEAL from a judgment of the Ontario Court of Appeal (Simmons, Cronk and Pepall JJ.A.), 2016 ONCA 515 , 132 O.R. (3d) 331, 349 O.A.C. 132, 401 D.L.R. (4th) 634, [2016] O.J. No. 3471 (QL), 2016 CarswellOnt 10242 (WL Can.), affirming a decision of Faieta J., 2015 ONSC 1128 , 125 O.R. (3d) 619, [2015] O.J. No. 1084 (QL), 2015 CarswellOnt 3080 (WL Can.). Appeal allowed, McLachlin C.J. and Moldaver and Gascon JJ. dissenting.

Paul B. Schabas , Kaley Pulfer and Brittiny Rabinovitch , for the appellants.

William C. McDowell , Julian Porter , Q.C. , and Brian Kolenda , for the respondent.

Jeremy de Beer , Marina Pavlović and David Fewer , for the intervener.

The reasons of Côté, Brown and Rowe JJ. were delivered by

Côté J. —

I. Introduction

[ 1 ] This appeal has to do with the rules for the assumption and exercise of jurisdiction in the context of multijurisdictional defamation claims. While these types of claims are not new, the exponential increase in multijurisdictional publications over the Internet has led to growing concerns about libel tourism and the possible assumption of jurisdiction by an unlimited number of forums.

[ 2 ] For the reasons set out below, I find that the current rules are able to address these challenges so long as the underlying principles of stability and fairness are kept in mind.

[ 3 ] While the motion judge in this case properly determined that he had jurisdiction (under the jurisdiction simpliciter test), he committed multiple errors in his forum non conveniens analysis. On a robust and careful assessment of the relevant factors tainted by these errors, I conclude that Israel is a clearly more appropriate forum for this claim to be heard.

[ 4 ] The appeal should be allowed.

II. Background and Facts

[ 5 ] The respondent, Mitchell Goldhar, is a prominent Canadian businessman who owns and operates SmartCentres Inc. in Ontario. He also owns the Maccabi Tel Aviv Football Club (“Maccabi Tel Aviv”), one of the most popular professional soccer teams in Israel. Goldhar, who has been described as a celebrity in Israel, maintains a residence there and travels there every few months.

[ 6 ] The corporate appellants publish Israel’s oldest daily newspaper in both English and Hebrew, in print and online. It has a distribution of about 70,000 print copies in Israel. The individual appellants are, respectively, the newspaper’s former sports editor and the author of the article alleged to be libellous. Collectively, the appellants are referred to as “Haaretz”.

[ 7 ] On November 29, 2011, Haaretz published an article about Goldhar, which the latter alleges to be libellous. The main subject of the article is Goldhar’s ownership and management of Maccabi Tel Aviv. That being said, it also references his Canadian business and his approach to management, as follows:

Though he spends most of his time in Canada, Maccabi Tel Aviv owner Mitch Goldhar runs his club down to every detail. But could his penny pinching and lack of long term planning doom the team.

Crises are par for the course at Maccabi Tel Aviv, even when the club appears to be on an even keel. Most of the crises don’t make it onto the public’s radar, but they have one thing in common: their connection to [ sic ] way that Canadian owner Mitch Goldhar runs the club.

Goldhar’s management model was imported directly from his main business interest — a partnership with Wal-Mart to operate shopping centers in Canada.

Within the club, however, there are those who believe that Goldhar’s managerial culture is based on overconcentration bordering on megalomania, penny-pinching and a lack of long-term planning.

Goldhar boasts to his business contacts in Toronto that he is not only the owner of Maccabi Tel Aviv but also its soccer director.

(Reproduced in 2016 ONCA 515 , 132 O.R. (3d) 331, Appendix “A”.)

The article was researched, written and edited in Israel, primarily in reliance on Israeli sources.

[ 8 ] The article was published in print and electronically in Hebrew and English. While it was not distributed in print form in Canada, it was available electronically. The motion judge found it likely that 200 to 300 people in Canada read the article; by comparison, the evidence showed that approximately 70,000 people read the article in Israel. Two affiants, both employed by SmartCentres Inc., have stated that they read the article and that it came to the attention of most of their approximately 200 co-workers. There is no evidence that those who read the article thought less of Goldhar as a result.

[ 9 ] On December 29, 2011, Goldhar commenced an action for libel, alleging “damage to his reputation in his business and personal life”. His amended statement of claim states that “[t]he plaintiff conducts business in Israel, Canada and the United States, and will continue to suffer damages in these countries and elsewhere” (reproduced in A.R., vol. II, pp. 1-8, at para. 12).

[ 10 ] Haaretz brought a motion to stay the action, arguing that Ontario courts lacked jurisdiction or, alternatively, that Israel was a clearly more appropriate forum.

[ 11 ] The motion judge dismissed the motion, finding that Ontario courts had jurisdiction and refusing to decline to exercise this jurisdiction in favour of Israeli courts. In doing so, he relied on two undertakings by Goldhar’s counsel. First, Goldhar would not seek damages at the trial of the action for reputational harm suffered in Israel or anywhere else outside of Canada. Second, Goldhar would pay for the travel and accommodation expenses of Haaretz’s witnesses at the rates provided in the Rules of Civil Procedure , R.R.O. 1990, Reg. 194.

[ 12 ] The majority of the Ontario Court of Appeal dismissed the appeal.

III. Judicial History

A. Ontario Superior Court of Justice — 2015 ONSC 1128 , 125 O.R. (3d) 619, per Faieta J. (March 6, 2015)

[ 13 ] The motion judge dismissed the motion to stay the action and added that, in the event that the action proceeded in Ontario, Goldhar’s claim would be limited to damages for reputational harm suffered within Canada and he would pay for the travel and accommodation expenses of Haaretz’s witnesses at the rates provided in the Rules. He also expressed the view that the lawsuit was far from being an abuse of process by Goldhar.

[ 14 ] The motion judge found that he had jurisdiction. The parties agreed that, as the evidence established that the article had been read in Ontario, a presumptive connecting factor existed. The motion judge found that Haaretz had failed to rebut the presumption. In particular, he did not view the absence of substantial publication of the libellous material in Ontario as rebutting the presumption, and he considered proof of harm to reputation irrelevant for the purposes of determining whether a minor element of the tort had occurred in Ontario.

[ 15 ] Further, the motion judge refused to decline to exercise jurisdiction, finding that Israel was not a clearly more appropriate forum after weighing the following factors:

• Comparative convenience and expense for the parties favoured a trial in Israel. The Haaretz defendants were all based in Israel. Goldhar visited Israel regularly and there was no evidence that a trial in Israel would cause him inconvenience or expense.

• Comparative convenience and expense for the witnesses slightly favoured a trial in Israel. Goldhar had not filed any evidence regarding the witnesses that he would call at trial. Haaretz proposed to call 22 witnesses, 18 of whom lived in Israel. However, the relevance of the testimony of some of Haaretz’s witnesses was questionable. Compelling the attendance of these witnesses in Ontario could be accomplished through the use of letters rogatory, also called letters of request. Moreover, foreign witnesses could testify via videoconference. Finally, Goldhar’s undertaking to pay for the travel and accommodation expenses of Haaretz’s witnesses at the rates provided by the Rules addressed any additional expense.

• Applicable law favoured a trial in Ontario. Regardless of which choice of law rule applied, the lex loci delicti (the place where the tort is committed) rule or the “most substantial harm to reputation” rule, Ontario law was applicable to this case. Ontario was the locus delicti of the tort of libel. Further, there was no comparative evidence of reputational harm to Goldhar in Israel and Ontario as a result of the publication, and there was limited evidence regarding Goldhar’s reputation. In light of this evidence, Goldhar’s undertaking not to seek damages for reputational harm outside of Canada was a very significant factor which led to the conclusion that the most substantial harm to his reputation was in Ontario.

• Loss of juridical advantage favoured a trial in Ontario. The availability of a jury trial in Ontario was a juridical advantage of which Goldhar would be deprived if the case were tried in Israel. Any juridical advantage Goldhar might enjoy under Israeli defamation law was irrelevant, since the proper question was whether the plaintiff should be denied the benefits of his decision to select a forum that was appropriate under the conflicts rules.

• Fairness to the parties favoured a trial in Ontario. Given the importance of place of reputation in Canadian defamation law, there was no surprise or injustice to Goldhar’s attempt to vindicate his reputation in Ontario, where he lives and works.

B. Ontario Court of Appeal — 2016 ONCA 515 , 132 O.R. (3d) 331, per Simmons and Cronk JJ.A. (Pepall J.A. Dissenting) (June 28, 2016)

[ 16 ] The majority of the Court of Appeal dismissed Haaretz’s appeal. The majority was of the view that the motion judge had not erred in failing to find that Haaretz had successfully rebutted the presumption of jurisdiction. As the article “draws a link between Goldhar’s management model and his Canadian business” (para. 41), there was a significant connection between the subject matter of the action and Ontario, and it should not have come as a surprise to Haaretz that Goldhar would seek to vindicate his reputation in Ontario. The question at the rebuttal stage of the jurisdiction simpliciter analysis was whether, objectively speaking, Ontario had a real and substantial connection to the subject matter of the action, not whether there was another forum that could also assume jurisdiction over the action. In the absence of evidence demonstrating no reputational harm, evidence of actual reputational harm was not necessary to establish jurisdiction.

[ 17 ] The majority also found no basis on which to interfere with the motion judge’s conclusion that Israel had not been shown to be a clearly more appropriate forum. Its analysis focused on the following factors:

• It was reasonable for the motion judge to conclude that the convenience and expense for the witnesses slightly favoured a trial in Israel. The motion judge erred in law by suggesting that letters rogatory could be used to compel the attendance of Israeli witnesses in Ontario. Despite this error, the motion judge reasonably based his analysis on the availability of letters rogatory for compelling testimony from witnesses outside of Ontario via videoconferencing, Goldhar’s undertaking to fund foreign witnesses’ travel and accommodation expenses, and the lack of evidence concerning the likely testimony of Haaretz’s proposed witnesses.

• The motion judge reasonably found that applicable law favoured Ontario irrespective of which choice of law rule was applied — the lex loci delicti rule or the “most substantial harm to reputation” rule. Haaretz’s reliance on the extent of publication in Israel would turn the “most substantial harm” approach into a proxy for the “substantial publication” rule rejected in Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636. Furthermore, Goldhar’s undertaking not to seek damages for reputational harm sustained outside of Canada confirmed the significance to him of his reputation in Ontario and the importance to him of vindicating his reputation here.

• While loss of juridical advantage was a neutral factor rather than a factor that favoured a trial in Ontario, this error was not significant in terms of the motion judge’s overall conclusion. The motion judge erred in accepting that Goldhar would suffer a loss of juridical advantage. As he had failed to deliver a jury notice, Goldhar was not entitled to claim a loss of juridical advantage. However, the motion judge correctly found that potential juridical advantages to a plaintiff in the alternative forum are irrelevant to the forum non conveniens analysis.

• There was no basis on which to interfere with the motion judge’s conclusion on the question of fairness. The motion judge considered it important that Goldhar lives and works in Ontario and that Haaretz chose to write an article about him impugning his management of an Israeli soccer team in a manner that implicated his Canadian business practices.

[ 18 ] Pepall J.A. agreed with the majority on the issue of the jurisdiction simpliciter test but would have allowed the appeal and stayed the action, finding that Israel was clearly the more appropriate forum. Given the ease with which jurisdiction may be established in a defamation case, she expressed the view that a “robust and carefully scrutinized review” of the issue of forum non conveniens was required (para. 132). In light of the errors committed by the motion judge, as identified by the majority, the motion judge’s conclusion was unreasonable. Pepall J.A. weighed the following factors:

• Comparative convenience and expense for the parties clearly and overwhelmingly supported a trial in Israel. There was no evidence of any inconvenience or undue expense for Goldhar associated with a trial in Israel.

• Comparative convenience and expense for the witnesses overwhelmingly favoured a trial in Israel. The motion judge’s error on letters rogatory, his failure to consider the purport of Tariff A of the Rules — providing for rates significantly below the actual cost of travel and accommodation — when dealing with Goldhar’s undertaking, and his failure to consider the fact that Goldhar had not identified any prospective witnesses all served to cause him to erroneously conclude that this factor only slightly favoured Israel.

• Applicable law favoured a trial in Israel. As pleaded, the tort occurred in both Ontario and Israel. The most substantial harm test favoured a trial in Israel. The evidence was that the extent of publication and any harm suffered were much more significant in that forum. Furthermore, the article was written in Israel about an Israeli soccer team and was aimed at an Israeli audience. The motion judge erred by treating Goldhar’s undertaking to limit his claim for damages to Ontario as determinative. This undertaking was materially different than the one this Court considered in Breeden v. Black , 2012 SCC 19 , [2012] 1 S.C.R. 666.

• At most, juridical advantage was a neutral factor. The motion judge erred in accepting that Goldhar would suffer a loss of juridical advantage with regard to the availability of a jury trial in Ontario.

• Fairness clearly favoured a trial in Israel. The motion judge focused on vindication of Goldhar’s reputation while failing to mention the burden a trial in Ontario would impose on Haaretz or the ability of a trial in Israel to achieve the vindication sought by Goldhar.

• Enforcement favoured a trial in Israel. The motion judge said nothing about enforcement. The only evidence before him was that Haaretz had no assets in Ontario, whereas it could be inferred that Goldhar did have assets in Israel.

IV. Issues

[ 19 ] This appeal raises the following issues and sub-issues:

1. Did the motion judge err in assuming jurisdiction?

(a) Is the situs of the tort a reliable basis on which to presume a “real and substantial connection” in Internet defamation cases?

(b) Under what circumstances, if any, can the presumption of jurisdiction be rebutted?

2. Did the motion judge err in finding that Israel is not a clearly more appropriate forum than Ontario? Notably, should the “most substantial harm” test rather than the lex loci delicti test apply to determine the applicable law in defamation actions?

V. Analysis

A. The Scope of Goldhar’s Claim

[ 20 ] At the outset, it is important to define the scope of Goldhar’s claim; the relevance of any given consideration to the jurisdiction simpliciter and forum non conveniens analyses is dependent on the scope of the claim. In my view, a careful review of Goldhar’s amended statement of claim reveals that his action was never limited to libellous statements pertaining to his Canadian business or damage to his Canadian reputation — with respect, my colleagues McLachlin C.J. and Moldaver and Gascon JJ. err by limiting the claim in this way.

[ 21 ] It is well established that the statement of claim, which in this case was amended by experienced counsel, defines the issues and informs the opposing parties of the case they have to meet ( Lax Kw’alaams Indian Band v. Canada (Attorney General) , 2011 SCC 56 , [2011] 3 S.C.R. 535, at para. 41 ). It frames the action for the purposes of analysing the assumption and exercise of jurisdiction.

[ 22 ] I cannot conclude from Goldhar’s amended statement of claim, as my colleagues in dissent do, that Goldhar is particularly “concerned about the impact on his Canadian business reputation” or that the “sting” of libel underlying his claim relates to his reputation in Ontario (paras. 213-14). While the amended statement of claim states that he is a business owner and operator and an active community member in Toronto, it refers directly to only one of his business enterprises, Maccabi Tel Aviv, and does not even mention SmartCentres Inc. Similarly, para. 9 of his amended statement of claim, which sets out what he considers the natural and ordinary meaning of the article, fails to identify any connection to his Canadian business. Furthermore, para. 10, which lists alleged factual errors and fabrications in the article, does not identify any such errors or fabrications relating to Goldhar’s Canadian business practices, but it does specifically identify statements pertaining to his management of Maccabi Tel Aviv (for example, “Goldhar does not have a long term plan for the team”). Most notably, the amended statement of claim makes no mention of the article’s claim that “Goldhar’s management model was imported directly from his main business interest — a partnership with Wal-Mart to operate shopping centers in Canada”, even though this is the passage that is said to provide the connection between the allegedly libellous statements and Goldhar’s Canadian business reputation. Finally, in describing the damage he suffered, Mr. Goldhar clearly states that he “conducts business in Israel, Canada and the United States” (para. 12). Canada is never singled out as the forum where reputational harm has been suffered for the purposes of this action.

[ 23 ] For these reasons, I am satisfied that Goldhar’s action was never limited to damage to his reputation in Ontario or to statements pertaining to his business in Ontario. Moreover, his undertaking before the motion judge to seek damages only for reputational harm suffered within Canada should not be allowed to narrow the scope of his pleadings. This undertaking, which does not preclude a future action from being commenced in Israel to recover damages there, is materially different than the one considered in Black and, as observed by Pepall J.A., the failure to preclude other actions “detracts from one of the relevant factors for forum non conveniens enumerated in [ Black ]: the avoidance of a multiplicity of legal proceedings and conflicting decisions” (para. 162).

[ 24 ] In light of these comments, I am of the view that Goldhar’s Israeli reputation and statements identified in his amended statement of claim pertaining to his Israeli business are also relevant to the assumption and exercise of jurisdiction. To conclude that “his reputation in Israel is not material to the analysis” ignores the claim as formulated by Goldhar before Haaretz brought a motion to stay (McLachlin C.J. and Moldaver and Gascon JJ., at para. 218). This was the case Haaretz had to meet ( Lax Kw’alaams , at para. 43). Neither Goldhar nor my colleagues McLachlin C.J. and Moldaver and Gascon JJ. (see notably paras. 151, 172, 213, 214 and 225) may now redefine Goldhar’s action so that it better responds to Haaretz’s motion to stay.

[ 25 ] I now turn to the principles underlying the assumption and exercise of jurisdiction.

B. Fundamental Principles Underlying the Conflict of Laws: Balancing Order and Fairness

[ 26 ] In Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572, LeBel J., for a unanimous Court, carefully explained the jurisdiction simpliciter analysis, which applies to the assumption of jurisdiction, as well as the forum non conveniens doctrine, which is meant to guide courts in deciding whether to exercise their jurisdiction. These principles, along with those relating to the recognition of foreign judgments, represent the common law conflicts rules of Canadian private international law and must be understood and analysed as a cohesive whole ( Van Breda , at para. 16).

[ 27 ] Central to a proper understanding of the conflicts rules of Canadian private international law, and to the resolution of this appeal, is an appreciation of the distinct roles played by jurisdiction simpliciter and forum non conveniens ( Van Breda , at paras. 46 and 56, affirming the reasoning of Sharpe J.A. in Muscutt v. Courcelles (2002), 2002 44957 (ON CA) , 60 O.R. (3d) 20 (C.A.), and Charron Estate v. Village Resorts Ltd. , 2010 ONCA 84 , 98 O.R. (3d) 721). The jurisdiction simpliciter analysis is meant to ensure that a court has jurisdiction. This will be the case where a “real and substantial connection” exists between a chosen forum and the subject matter of the litigation. The forum non conveniens analysis, on the other hand, is meant to guide courts in determining whether they should decline to exercise that jurisdiction in favour of a “clearly more appropriate” forum.

[ 28 ] The importance of maintaining this distinction flows from the discrete concerns underlying each analysis and the nature of the relevant factors at each stage. The “real and substantial connection” test at the jurisdiction simpliciter stage prioritizes order, stability and predictability by relying on objective connecting factors for the assumption of jurisdiction. Conversely, the forum non conveniens analysis emphasizes fairness and efficiency by adopting a case-by-case approach to identify whether an alternative jurisdiction may be “clearly more appropriate”. I will briefly elaborate on the principles underlying each analysis.

[ 29 ] In defining the content of the “real and substantial connection” test for the assumption of jurisdiction, this Court was faced with a choice between an approach based on objective connecting factors and a case-by-case approach ( Van Breda , at para. 30). This choice was characterized by the tension between predictability and consistency, on the one hand, and fairness and efficiency, on the other ( Van Breda , at para. 66). Ultimately, the Court decided to prioritize order and predictability at the jurisdiction simpliciter stage, in the following terms:

Given the nature of the relationships governed by private international law, 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. As La Forest J. wrote in Morguard , there must be order in the system, and it must permit the development of a just and fair approach to resolving conflicts. Justice and fairness are undoubtedly essential purposes of a sound system of private international law. But they cannot be attained without a system of principles and rules that ensures security and predictability in the law governing the assumption of jurisdiction by a court. Parties must be able to predict with reasonable confidence whether a court will assume jurisdiction in a case with an international or interprovincial aspect.

( Van Breda , at para. 73)

To achieve this order and predictability, the Court opted to rely on a set of defined presumptive connecting factors at the jurisdiction simpliciter stage ( Van Breda , at para. 78).

[ 30 ] This objectively ascertainable and relatively low bar to establishing that a chosen forum has jurisdiction, on a prima facie basis, reflects the constitutional imperative underlying the jurisdiction simpliciter stage, as described in Van Breda :

From a constitutional standpoint, the Court has, by developing tests such as the real and substantial connection test, sought to limit the reach of provincial conflicts rules or the assumption of jurisdiction by a province’s courts . . . . In its constitutional sense, it places limits on the reach of the jurisdiction of a province’s courts and on the application of provincial laws to interprovincial or international situations. [Emphasis added; para. 23.]

The constitutional purpose of the jurisdiction simpliciter test is to establish a minimum threshold for the assumption of jurisdiction in order to prevent improper assumptions of jurisdiction ( Van Breda , at para. 26; see also Hunt v. T&N plc , 1993 43 (SCC) , [1993] 4 S.C.R. 289, at p. 325). Its objective is to delineate circumstances in which a court has jurisdiction, not circumstances in which it should exercise it (which is the purpose of forum non conveniens ). The prioritization of order and predictability at the jurisdiction simpliciter stage is also consistent with the principle of comity, which is central to Canadian private international law ( Van Breda , at para. 74).

[ 31 ] This prioritization of order and stability at the jurisdiction simpliciter stage, through the adoption of objective presumptive connecting factors, is meant to work in tandem with a flexible case-by-case approach to forum non conveniens . Once it is established that a court has jurisdiction, the forum non conveniens doctrine requires a court to determine whether it should exercise such jurisdiction.

[ 32 ] The purpose of the forum non conveniens analysis is to temper any potential rigidity in the rules governing the assumption of jurisdiction and “to assure fairness to the parties and the efficient resolution of the dispute” ( Van Breda , at para. 104). This is necessary given this Court’s recognition that jurisdiction “may sometimes be established on a rather low threshold” ( Van Breda , at para. 109). By focusing “on the contexts of individual cases”, the forum non conveniens stage plays an important role in striking a balance between order and fairness ( Van Breda , at para. 105).

[ 33 ] Bearing these principles in mind, I turn to the case at bar.

C. Did the Motion Judge Err in Assuming Jurisdiction?

[ 34 ] In determining whether a “real and substantial connection” exists between a chosen forum and the subject matter of the litigation, courts are required to consider two issues. First, a court must consider whether the existence of a recognized presumptive connecting factor has been established ( Van Breda , at para. 80). If so, the court must consider whether the party challenging the assumption of jurisdiction has successfully rebutted the presumption ( Van Breda , at para. 81).

(1) Existence of a Presumptive Connecting Factor

[ 35 ] The judges in the courts below agreed that a presumptive connecting factor had been established. Haaretz, however, submits that the situs of the tort is an unreliable basis on which to presume a “real and substantial connection” in Internet defamation cases. In its view, the ease with which publication can be established in such cases gives rise to only a “weak relationship” with the chosen forum.

[ 36 ] As previously discussed, in Van Breda , the Court stressed the importance of determining jurisdiction “on the basis of objective factors” establishing a relationship between the subject matter of the litigation and the chosen forum (para. 82). The Court identified the following presumptive connecting factors grounding a court’s assumption of jurisdiction:

(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. [Emphasis added; para. 90.]

The tort of defamation, which is a tort of strict liability, is committed where material has been “communicated” to, that is, conveyed to and received by, at least one person other than the plaintiff ( Crookes v. Newton , 2011 SCC 47 , [2011] 3 S.C.R. 269, at paras. 1 and 16 ). This was recognized by this Court in Banro : “. . . a single instance of publication is sufficient for the tort to crystallize” (para. 55). In the case of Internet communications, the publication of defamatory statements occurs when they are read or downloaded by the recipient ( Black , at para. 20 ; see also P. A. Downard, The Law of Libel in Canada (4th ed. 2018); Brown on Defamation: Canada, United Kingdom, Australia, New Zealand, United States (2nd ed. (loose-leaf)), by R. E. Brown, at pp. 7‑17 to 7-25). Accordingly, the situs of Internet-based defamation is the place where the defamatory statements are read, accessed or downloaded by the third party ( Crookes v. Holloway , 2007 BCSC 1325 , 75 B.C.L.R. (4th) 316, at para. 26 , aff’d 2008 BCCA 165 , 77 B.C.L.R. (4th) 201; Brown, at pp. 7-122 to 7-126; M. Castel, “Jurisdiction and Choice of Law Issues in Multistate Defamation on the Internet” (2013), 51 Alta. L. Rev. 153, at p. 156).

[ 37 ] Insofar as it attempts to raise doubt as to the validity of the presumptive connecting factors identified in Van Breda , Haaretz’s argument must be rejected. This Court has found that “[t]he situs of the tort is clearly an appropriate connecting factor” and that there is no difficulty “in acknowledging the validity of this factor once the situs has been identified” ( Van Breda , at para. 88 (emphasis added)). Raising doubt as to the value of the situs of the tort as a presumptive connecting factor would significantly undermine the above-noted objectives of predictability and order at the jurisdiction simpliciter stage. Indeed, courts should be cautious in carving out exceptions to conflicts rules, as “[a]ny exception adds an element of uncertainty” ( Tolofson v. Jensen , 1994 44 (SCC) , [1994] 3 S.C.R. 1022, at p. 1061). It is therefore preferable to address any concerns relating to the insufficiency of a presumptive connecting factor either at the rebuttal stage of the jurisdiction simpliciter analysis or at the forum non conveniens stage.

[ 38 ] For these reasons, I conclude that a presumptive connecting factor has been established in the circumstances of this case, and I turn to the issue of whether the presumption has been rebutted.

(2) Rebutting the Presumption

[ 39 ] At this stage, it is appropriate to take into account the legitimate concerns raised by Haaretz about the ease with which a presumptive connecting factor may be established in Internet defamation cases. This Court has previously recognized the risk of jurisdictional overreach in these types of cases:

The tort of defamation presents an interesting challenge for the principles underlying the assumption of jurisdiction. At common law, the tort of defamation crystallizes upon publication of the libellous material . . . . This also raises difficult issues when publication occurs through the Internet . . . .

( Banro , at para. 3)

Pepall J.A., dissenting at the Court of Appeal, expressed her reluctance to accept the motion judge’s assumption of jurisdiction in light of similar concerns:

To succeed in an action for defamation, the plaintiff must prove on a balance of probabilities that the defamatory words were communicated to at least one person other than the plaintiff: see Crookes v. Newton , [2011] 3 S.C.R. 269, [2011] S.C.J. No. 47, 2011 SCC 47 , at para. 1 . As well, at the jurisdiction stage of the proceedings, the plaintiff’s pleadings are accepted as true unless contradicted by evidence adduced by the defendants: see Banro , at para. 38. Accordingly, all that is needed for the presumptive connecting factor to be found is for the plaintiff to plead that the alleged defamatory material was communicated to at least one person in Ontario other than the plaintiff. While this is easy to establish in any defamation case, it is virtually automatic in a case of defamation on the Internet , where online publications are readily shared and accessed by users across the world. [Emphasis added; para. 127.]

[ 40 ] The ability to rebut the presumption of jurisdiction where there is only a weak relationship between the subject matter of the litigation and the forum serves as an important check on jurisdiction ( Van Breda , at para. 95). A careful examination of this question is therefore of particular importance in Internet defamation cases, where a presumptive connecting factor can easily be established.

[ 41 ] Having recognized the importance of the ability to rebut the presumption of jurisdiction, I turn to consider Haaretz’s submission that, based on the analyses of the motion judge and the majority below, rebutting the presumption of jurisdiction does not seem possible at all in these types of cases.

[ 42 ] This Court has recognized that presumptive connecting factors must not give rise to an irrebuttable presumption of jurisdiction. A defendant may argue that a given connection is inappropriate in the circumstances of a particular case:

The presumption of jurisdiction that arises where a recognized connecting factor — whether listed or new — applies is not irrebuttable. The burden of rebutting the presumption of jurisdiction rests, of course, on the party challenging the assumption of jurisdiction. That party must establish facts which demonstrate that the presumptive connecting factor does not point to any real relationship between the subject matter of the litigation and the forum or points only to a weak relationship between them.

( Van Breda , at para. 95; see also para. 81)

[ 43 ] In order for a defendant to succeed in showing that “a given connection is inappropriate in the circumstances of the case”, the circumstances must demonstrate that the relationship between the forum and the subject matter of the litigation is such that it would “not be reasonable to expect that the defendant would be called to answer proceedings in that jurisdiction” ( Van Breda , at paras. 81 and 97; see also para. 92). To satisfy this test, the party challenging the court’s jurisdiction should rely on factors other than those considered at the forum non conveniens stage: “. . . the factors that would justify a stay in the forum non conveniens analysis should not be worked into the jurisdiction simpliciter analysis . . .” ( Van Breda , at para. 56).

[ 44 ] Assuming that these principles are properly applied, the situs of the tort will not give rise to an irrebuttable presumption of jurisdiction in Internet defamation cases. While it is not appropriate to propose an exhaustive list of factors that can rebut the presumption of jurisdiction in these types of cases, it is not difficult to imagine circumstances in which it would not be reasonable to expect that the defendant would be called to answer a legal proceeding in a chosen forum. For example, evidence that a plaintiff has no reputation in the chosen forum may be a factor tending to rebut the presumption of jurisdiction in a defamation action. As the protection of reputation is the primary purpose of defamation law ( Banro , at paras. 57-58), absence of reputation would tend to point to a weak relationship between the forum and the subject matter of the litigation. Indeed, this Court, in Banro , relied in part on the plaintiff’s reputation in the chosen forum to conclude that it would be inappropriate to find that the presumption of jurisdiction had been rebutted in the circumstances of that case (para. 38).

[ 45 ] In the case at bar, the evidence fails to establish that Haaretz could not have reasonably expected to be called to answer a legal proceeding in Ontario. The pleadings indicate that Goldhar lives and operates his businesses in Ontario. Haaretz had knowledge of this fact, and the allegedly libellous article directly references Goldhar’s Canadian residency and Canadian business practices. As such, this is not a case where the presumption of jurisdiction is rebutted.

D. Did the Motion Judge Err in Finding That Israel Is Not a Clearly More Appropriate Forum Than Ontario?

[ 46 ] Having established jurisdiction (pursuant to the jurisdiction simpliciter analysis), the motion judge properly considered the question of forum non conveniens . At the forum non conveniens stage, the burden is on the defendant to satisfy the motion judge that the alternative forum is “clearly more appropriate” by establishing that it would be fairer and more efficient to proceed in that forum:

The use of the words “clearly” and “exceptionally” should be interpreted as an acknowledgment that the normal state of affairs is that jurisdiction should be exercised once it is properly assumed. The burden is on a party who seeks to depart from this normal state of affairs to show that, in light of the characteristics of the alternative forum, it would be fairer and more efficient to do so and that the plaintiff should be denied the benefits of his or her decision to select a forum that is appropriate under the conflicts rules. The court should not exercise its discretion in favour of a stay solely because it finds, once all relevant concerns and factors are weighed, that comparable forums exist in other provinces or states. It is not a matter of flipping a coin. A court hearing an application for a stay of proceedings must find that a forum exists that is in a better position to dispose fairly and efficiently of the litigation. But the court must be mindful that jurisdiction may sometimes be established on a rather low threshold under the conflicts rules. Forum non conveniens may play an important role in identifying a forum that is clearly more appropriate for disposing of the litigation and thus ensuring fairness to the parties and a more efficient process for resolving their dispute. [Emphasis added.]

( Van Breda , at para. 109)

[ 47 ] While the normal state of affairs favours exercising jurisdiction in the forum where it is properly assumed, this should never come at the cost of one party facing unfair or clearly inefficient proceedings. The purpose of forum non conveniens , as discussed above, is to temper any potential rigidity in the rules governing the assumption of jurisdiction and to “assure fairness to the parties and the efficient resolution of the dispute” ( Van Breda , at para. 104). Where the evidence indicates that the alternative forum is in a better position to dispose fairly and efficiently of the litigation, the court should grant the stay ( Van Breda , at para. 109). This is especially true in cases where the evidence raises doubt as to whether proceeding in the chosen forum will provide the defendant with a fair opportunity to present its case.

[ 48 ] In light of the purpose of forum non conveniens , I agree with Pepall J.A. that, “given the ease with which jurisdiction simpliciter may be established in a defamation case, in a motion for a stay, a motion judge must conduct a robust and carefully scrutinized review of the issue of forum non conveniens ” (para. 132). It is true that defamation cases involve a particularly rigid application of the rules governing the assumption of jurisdiction. As discussed above, the establishment of a presumptive connecting factor is “virtually automatic” in Internet defamation cases (Pepall J.A., at para. 127). Where there is no “real and substantial connection” to the chosen forum, a proper analysis at the rebuttal stage will alleviate some of the consequences of the rigid application of the rules governing the assumption of jurisdiction. That being said, there are some other consequences to the rigid application of these rules that can only be addressed in the forum non conveniens analysis. For example, where a plaintiff enjoys a reputation in multiple forums, publication may allow jurisdiction to be properly assumed in all of them, without regard to how fair or efficient it may be to proceed in the chosen forum. This is to be expected as, again, “the factors that would justify a stay in the forum non conveniens analysis should not be worked into the jurisdiction simpliciter analysis” ( Van Breda , at para. 56). As the rebuttal stage fails to address all the consequences of the “virtually automatic” presumption of jurisdiction in defamation actions, it is appropriate for motion judges to be particularly attuned to concerns about fairness and efficiency at the forum non conveniens stage in these types of cases. This should not be understood as imposing a different standard or burden for defamation cases.

[ 49 ] I acknowledge that a motion judge’s decision on a stay motion is entitled to deference:

The application of forum non conveniens is an exercise of discretion reviewable in accordance with the principle of deference to discretionary decisions: an appeal court should intervene only if the motion judge erred in principle, misapprehended or failed to take account of material evidence, or reached an unreasonable decision (see Young v. Tyco International of Canada Ltd. , at para. 27).

( Banro , at para. 41)

As the forum non conveniens analysis is inherently factual in nature, courts of appeal should not normally interfere with a motion judge’s factual findings. That being said, there are limits to deference, as recognized in Banro . Where the motion judge has “erred in principle, misapprehended or failed to take account of material evidence, or reached an unreasonable decision”, courts of appeal may intervene.

[ 50 ] Bearing these principles in mind, and for the reasons set out below, I conclude that this Court may intervene in the case at bar. The motion judge committed the following errors (with references to each factor impacted by the error, as discussed below):

• He erred in finding that letters rogatory could be used to compel Israeli witnesses to testify in Ontario (Comparative Convenience and Expense for the Witnesses, and Fairness).

• He erred by giving significant weight to Goldhar’s undertaking to fund the travel and accommodation expenses of the foreign witnesses in accordance with the rates provided in the Rules (Comparative Convenience and Expense for the Witnesses).

• He erred by unreasonably discounting Haaretz’s proposed witnesses and the relevance of their evidence (Comparative Convenience and Expense for the Witnesses).

• He erred by failing to consider Goldhar’s significant reputation in Israel (Fairness).

• He erred by failing to weigh Goldhar’s interest in vindicating his reputation in Ontario against the significant unfairness that a trial in Ontario would impose on Haaretz (Fairness).

• He erred by failing to consider the question of enforcement (Enforcement).

In committing the first error, the motion judge misapprehended the role of letters rogatory entirely. As I will explain, the remaining errors involved a complete misapprehension of, or failure to consider, material evidence, and not merely, as my colleagues in dissent argue, an unsatisfactory weighing of that evidence. These errors tainted the motion judge’s forum non conveniens analysis on each of the factors they affected as well as his overall weighing of these factors. As a result, no deference should be afforded to these aspects of the motion judge’s analysis.

[ 51 ] I turn then to an assessment of each factor raised by Haaretz.

(1) Comparative Convenience and Expense for the Parties

[ 52 ] The motion judge concluded that the comparative convenience and expense for the parties favoured a trial in Israel (para. 36). Neither party disputes this.

[ 53 ] The motion judge reached this conclusion on three bases. First, there was no evidence that a trial in Israel would cause any inconvenience or expense to Goldhar (para. 35). Second, holding a trial in Ontario would place a strain on the Israeli defendants (paras. 31-33). And finally, portions of the trial might need to be conducted in Hebrew with interpreters (para. 34). I would not disturb the motion judge’s conclusion on this factor.

(2) Comparative Convenience and Expense for the Witnesses

[ 54 ] Goldhar did not file any evidence regarding the witnesses he would call at trial, while Haaretz filed a list of 22 witnesses and described, in its factum, what each of the witnesses “may speak to” (motion judge, at para. 41). Furthermore, a supplemental affidavit described the evidence that eight of these witnesses could give to assist Haaretz at trial (motion judge, at para. 43). Of the 22 witnesses, 18 resided in Israel.

[ 55 ] The motion judge concluded that the comparative convenience and expense for the witnesses, a factor distinct from the comparative convenience and expense for the parties, only slightly favoured a trial in Israel (para. 45). This conclusion was wholly unreasonable in light of the evidence before him.

[ 56 ] While the motion judge did not specifically identify the basis on which he concluded that the comparative convenience and expense for the witnesses only slightly favoured a trial in Israel, he did highlight four considerations. First, he considered that Goldhar had filed an expert opinion to the effect that “many of the witnesses do not have relevant evidence” (para. 41). Second, he dismissed Haaretz’s concern about being unable to compel unwilling witnesses living outside of Ontario. He found that these witnesses would remain unwilling to testify even if the trial were held in Israel and that “compelling the attendance of these witnesses to a court in Ontario can be accomplished through the use of letters rogatory” (para. 42). Third, the motion judge considered that arrangements could be made to have foreign witnesses testify by videoconferencing technology, pursuant to rule 1.08 of the Rules (para. 44). Finally, he found that Goldhar’s undertaking to pay for the travel and accommodation costs of foreign witnesses in accordance with the rates provided in the Rules addressed any additional expense related to holding a trial in Ontario (para. 44).

[ 57 ] The Court of Appeal correctly recognized that the motion judge had erred in law by suggesting that letters rogatory could be used to compel the attendance of Haaretz’s witnesses in Ontario, but the majority concluded that this error did not make the motion judge’s overall assessment of this factor unreasonable. It gave three reasons for reaching this conclusion. First, videoconferencing could be used to obtain the testimony of witnesses who were unwilling or unable to come to Ontario:

Contrary to Haaretz’s arguments, in my view, the motion judge was entitled to accept that reluctant foreign witnesses could be compelled to provide evidence in Israel through the use of letters of request and that videoconferencing was a potential means of obtaining the evidence of any witnesses unwilling to come to Ontario.

These are available methods, under the Rules of Civil Procedure , for dealing with witnesses outside the jurisdiction. Haaretz led no evidence to undermine Goldhar’s submissions that these methods would be available in this case. Haaretz bore the burden of demonstrating that Israel is the clearly more appropriate forum. On this record, it was not unreasonable for the motion judge to accept that Ontario letters of request would be honoured by Israel and that videoconferencing would be available in that jurisdiction.

Further, in the absence of evidence or adverse judicial commentary, the use of technology and interpreters cannot be viewed as undermining the fairness of a civil trial. We live in an age of international communication and commerce. Multi-jurisdictional parties — and witnesses who do not speak either of Canada’s official languages — are to be expected. Courtroom procedures must accommodate testimony by videoconferencing. Interpreters have long been a common feature of the Canadian judicial system. The motion judge’s implicit conclusion that using these procedures would not undermine the fairness of the trial was not unreasonable. [Footnote omitted; paras. 69-71.]

Second, the majority of the Court of Appeal was of the view that Goldhar’s undertaking to fund travel and accommodation expenses relieved any additional expense of holding the trial in Ontario. Finally, there was a lack of evidence concerning the likely testimony of Haaretz’s proposed witnesses:

While many of Haaretz’s proposed witnesses could have information about relevant matters, the record contains scant information about what particular witnesses are actually likely to say. Importantly, Mr. Marouani, the reporter who wrote the article, did not provide an affidavit on the motion. Nor did Haaretz produce any witness statements or even any notes of conversations with the proposed witnesses. In these circumstances, the motion judge was entitled to treat Haaretz’s proposed witness list with caution. [para. 73]

[ 58 ] There is no doubt that the motion judge erred in finding that letters rogatory could be used to compel Israeli witnesses to testify in Ontario.

[ 59 ] First, by dismissing evidence that Haaretz’s witnesses would not testify voluntarily on the basis that “[t]his concern will exist even if the trial is held in Israel” (para. 42), the motion judge ignored the very concern raised by Haaretz, namely that Israeli witnesses, while compellable in Israel, could not be effectively compelled to testify if the trial proceeded in Ontario. Second, as the Court of Appeal found, he erred by considering that Israeli witnesses could be compelled in Ontario through the use of letters rogatory.

[ 60 ] The analysis of the majority of the Court of Appeal on this point only compounded the motion judge’s errors.

[ 61 ] First, the majority of the Court of Appeal erroneously found that the motion judge “was entitled to accept that reluctant foreign witnesses could be compelled to provide evidence in Israel through the use of letters of request and that videoconferencing was a potential means of obtaining the evidence of any witnesses unwilling to come to Ontario” (para. 69). The motion judge never made such a finding. He found that Israeli witnesses could be compelled to testify in Ontario by letters rogatory, which was incorrect. He never found that letters rogatory could be used to compel an Israeli witness to testify by videoconference at a trial taking place in Ontario.

[ 62 ] Second, the majority found that there were “available methods, under the Rules of Civil Procedure , for dealing with witnesses outside the jurisdiction” and that “Haaretz led no evidence to undermine Goldhar’s submissions that these methods would be available in this case” (para. 70).

[ 63 ] It was not up to Haaretz to lead evidence that videoconferencing would not be an available means of compelling the testimony of Israeli witnesses. Haaretz had relied on evidence to the effect that many of its witnesses would not voluntarily testify. Given the fact that these witnesses could not be compelled directly, Haaretz met its burden in establishing a concern as to the fairness of a trial in Ontario. As such, it was up to Goldhar to respond with evidence that these witnesses could, in fact, be compelled in Ontario, thus addressing any concern relating to fairness. Furthermore, Goldhar raised the availability of videoconferencing only in oral submissions. Haaretz could not be required to lead evidence to respond to unsupported representations. Finally, Haaretz could not be required to prove a negative: that compelled testimony by videoconference would not be available.

[ 64 ] This conclusion is supported by Moore v. Bertuzzi , 2014 ONSC 1318 , 53 C.P.C. (7th) 237, on which Goldhar relied in establishing the availability of videoconferencing as a means to compel foreign witnesses to testify. That case involved a motion to issue a letter of request to the judicial authorities in the state of Washington to compel testimony by videoconference, not a forum non conveniens application. The Ontario Superior Court, at para. 86, found that expert evidence proving that a foreign jurisdiction would actually enforce a request to compel testimony by videoconference was unnecessary in the context of determining whether an Ontario court should grant a motion for a letter of request. However, the court agreed that “expert evidence would be necessary to prove U.S. laws to determine whether Washington courts would enforce such a request ” (para. 86 (emphasis added)). While proving this fact was unnecessary in Bertuzzi , determining whether it is likely that Israel would actually enforce such a letter of request is crucial to ensuring the fairness of a potential trial in Ontario. It was not found in Bertuzzi that such letters of request will generally be enforced by foreign jurisdictions. This fact must be proven by expert evidence led by the party seeking to establish it; in this case, Goldhar.

[ 65 ] For all these reasons, the evidence did not allow the courts below to ensure that Haaretz would be able to compel its witnesses to testify if the trial proceeded in Ontario. Being unable to do so would affect Haaretz’s ability to defend itself in Ontario, which would be significantly unfair.

[ 66 ] The motion judge also erred by giving significant weight to Goldhar’s undertaking to fund the travel and accommodation expenses of the foreign witnesses in accordance with the rates provided in the Rules. The motion judge placed significant weight on this undertaking, finding that it addressed the additional expense of the Ontario forum. He went so far as to include this undertaking as a condition in his order. The majority of the Court of Appeal ignored the fact that the undertaking was relevant only in light of the motion judge’s finding that letters rogatory could be used to compel testimony in Ontario. If testimony is now to take place via videoconference, this undertaking is of no significant value. Furthermore, consideration of such an undertaking would allow a wealthy plaintiff to sway the forum non conveniens analysis, which would be inimical to the foundational principles of fairness and efficiency underlying this doctrine.

[ 67 ] Finally, the motion judge erred by unreasonably discounting Haaretz’s proposed witnesses and the relevance of their evidence. Haaretz had, in its factum, described what these 22 witnesses “may speak to” and had filed a supplemental affidavit briefly describing the evidence that 8 of the witnesses might give to assist it at trial. Goldhar had, in comparison, no evidence concerning the witnesses he might call and what those witnesses would speak to.

[ 68 ] The opinion of an expert cannot serve to raise doubt as to the relevance of a proposed witness’ testimony. Only a motion judge can make such determinations. In a forum non conveniens analysis, an expert’s opinion as to the relevance of certain testimony should not be permitted to minimize the testimony of possible witnesses such that their inability to participate in a trial is seen as any less unfair.

[ 69 ] The evidence of Haaretz’s Israeli witnesses was clearly relevant. The statements alleged by Goldhar to be libellous were, in large part, derived from information obtained from informants within Maccabi Tel Aviv. One notable example is the following statement:

Within the club , however, there are those who believe that Goldhar’s managerial culture is based on overconcentration bordering on megalomania, penny-pinching and a lack of long-term planning. [Emphasis added.]

This statement, the natural and ordinary meaning of which is raised at para. 9 of the amended statement of claim, is specifically said to be based on information obtained by Haaretz from Maccabi Tel Aviv insiders. The motion judge erred by discounting the relevance of their testimony.

[ 70 ] For all these reasons, I conclude that this factor weighs heavily in favour of a trial in Israel. Haaretz was the only party to provide evidence of the witnesses it might call. The testimony of those witnesses was clearly relevant to this action. Even so, the courts below never satisfied themselves that these witnesses could be compelled to testify if the action proceeded in Ontario, despite the fact that it would be significantly unfair for Haaretz to be unable to compel them.

(3) Loss of Legitimate Juridical Advantage

[ 71 ] The motion judge relied on the fact that a jury trial was available in Ontario, but not in Israel, to conclude that loss of juridical advantage favoured a trial in Ontario (paras. 55 and 61-63). The majority of the Court of Appeal found that the motion judge had erred in accepting that Goldhar would suffer a loss of juridical advantage, as he had not delivered a jury notice prior to the motion (paras. 92-94). As a result, it concluded that this was a neutral factor rather than one that favoured a trial in Ontario (para. 99).

[ 72 ] Haaretz submits that the majority of the Court of Appeal erred by finding (at para. 100) that it was “confident that [the motion judge] was aware of [the] cautions” against giving this factor too much weight. Further, it submits that the motion judge erred by discounting the relevance of juridical advantages available to Goldhar in Israel. The expert evidence was that Israeli defamation law is more plaintiff-friendly because truth is not an absolute defence, a successful plaintiff is entitled to statutory damages and courts may order a defendant to publish a correction or retraction.

[ 73 ] Conversely, Goldhar submits that the motion judge properly concluded that access to a jury trial was an important juridical advantage that favoured a trial in Ontario (R.F., at para. 105). I agree.

[ 74 ] The right to a jury trial is a substantive right of particular importance in defamation cases. As any party in Ontario may deliver a jury notice before the close of pleadings (rule 47.01 of the Rules), this was a juridical advantage still available to Goldhar at the time of the stay motion.

[ 75 ] With regard to the relevance of any juridical advantage available to Goldhar in Israel, in my view, the motion judge correctly held as follows:

In my view, any juridical advantages to the plaintiff under Israeli defamation law are irrelevant as a comparative analysis at this stage is not required. Given that the context for this analysis is whether the plaintiff should be denied the benefits of his decision to select a forum that is appropriate under the conflicts rules, then the measure is whether there is a loss, rather than a calculation of the net loss, of legitimate juridical advantage for the plaintiff if this action were to proceed in Israel. [para. 62]

[ 76 ] As a result, I would not disturb the motion judge’s conclusion on this factor. That being said, for the reasons set out by this Court at para. 27 of Black , this factor “should not weigh too heavily in the forum non conveniens analysis”.

(4) Fairness

[ 77 ] The motion judge found that fairness favoured a trial in Ontario, as there “is no surprise or injustice to the plaintiff’s attempt to vindicate his reputation in Ontario, where he lives and works” (para. 65). The majority of the Court of Appeal saw no reason to interfere with this finding (para. 104). In my view, the motion judge committed two errors in considering this important factor.

[ 78 ] First, he erred by failing to consider Goldhar’s significant reputation in Israel. As discussed above, Goldhar’s action was never limited to damage to his reputation in Ontario or to statements pertaining to his business in Ontario. While I agree with my colleagues McLachlin C.J. and Moldaver and Gascon JJ. that fairness “supports allowing Mr. Goldhar to vindicate his reputation in the jurisdiction where he maintains his reputation, and where the sting of the article was felt by him” (para. 214), in the circumstances of this case, fairness must be analysed in light of Goldhar’s multijurisdictional reputation. As my colleague Abella J. correctly notes at para. 141 of her concurring reasons, focusing solely on Ontario ignores the reality of Goldhar’s significant business interest and reputation in Israel. Not only is his reputation in Israel established by the evidence, but importantly, the amended statement of claim confirms that he saw himself as enjoying a significant reputation in Israel. While the motion judge correctly found that there “is no surprise or injustice to the plaintiff’s attempt to vindicate his reputation in Ontario, where he lives and works” (para. 65), Goldhar would suffer no significant unfairness by having to bring a libel claim in Israel for comments that were written and researched in Israel and that pertain primarily to his reputation and business in that jurisdiction.

[ 79 ] Second, the motion judge erred by failing to weigh Goldhar’s interest in vindicating his reputation in Ontario against the significant unfairness that a trial in Ontario would impose on Haaretz. As discussed above, the evidence did not allow the courts below to ensure that Haaretz would be able to compel its witnesses to testify if the trial were to proceed in Ontario. This raises doubt as to whether Haaretz would have a fair opportunity to defend itself if a trial were held in Ontario. The prospect of such a circumstance, which would be significantly unfair to Haaretz, outweighs Goldhar’s interest in vindicating his reputation in Ontario rather than Israel. By not considering this, the motion judge failed to carry out his duty “to ensure that both parties are treated fairly” ( Van Breda , at para. 105 (emphasis added)).

[ 80 ] For these reasons, I conclude that fairness favours a trial in Israel.

(5) Enforcement

[ 81 ] As noted by Pepall J.A., the question of enforcement was argued before the motion judge but was not addressed in his decision or by the majority of the Court of Appeal (para. 192). The motion judge erred in failing to address this question.

[ 82 ] As Haaretz has no presence or assets in Ontario, any order against it will have to be enforced by Israeli courts. My colleague Abella J., at para. 142 of her concurring reasons, also recognizes that this raises concerns about the multiplicity of proceedings that may arise from a trial in Ontario, and thus it slightly favours a trial in Israel.

[ 83 ] I respectfully disagree with my colleagues McLachlin C.J. and Moldaver and Gascon JJ. that the focus on the vindication of a plaintiff’s reputation “often renders the enforcement of the final judgment irrelevant to the forum non conveniens analysis in defamation cases” (para. 236). This Court, in Van Breda , identified problems related to the recognition and enforcement of judgments as a factor that a court might consider in deciding whether to apply forum non conveniens (para. 110). In the case at bar, Goldhar specifically claimed general damages in the amount of $600,000 and punitive damages in the amount of $100,000 in his amended statement of claim. He claimed he would “continue to suffer damage, and in particular financial loss ” in Israel, Canada and the United States and elsewhere (A.R., vol. II, pp. 1-8, at para. 12 (emphasis added)). There is no basis upon which to now claim that, if Goldhar succeeds, he will not seek to enforce a final judgment. In fact, even if Goldhar were to seek and obtain an order requiring Haaretz to correct or remove the offending article, which would serve to vindicate his reputation, the order would need to be enforced in Israel. While I agree with Nordheimer J., as he then was, that “this factor alone is not determinative” ( Barrick Gold Corp. v. Blanchard & Co. (2003), 2003 64238 (ON SC) , 9 B.L.R. (4th) 316, at para. 40 ), it is not irrelevant, and it is properly weighed in the forum non conveniens analysis, as this Court did in Black , also an action for defamation (para. 35).

(6) Applicable Law

[ 84 ] This Court, in Tolofson , established lex loci delicti , or the place where the tort occurs, as the general principle for determining choice of law (p. 1050). This rule is meant to ensure “certainty, ease of application and predictability” ( Tolofson , at p. 1050).

[ 85 ] This Court did, however, leave the door open to carefully defined exceptions to this rule, particularly if the place where the tort occurs differs from the place where its consequences are felt. La Forest J., in Tolofson , considered that the tort of libel may possibly be such a case (pp. 1042 and 1050; see also Banro , at paras. 50-51). This led LeBel J., in Banro , to note that a possible alternative approach to choice of law in defamation cases may be the place of most substantial harm to reputation (para. 56).

[ 86 ] The motion judge found that the locus delicti of the tort was Ontario. Based on the limited comparative evidence regarding Goldhar’s reputation in Ontario and Israel, as well as Goldhar’s undertaking not to seek at the trial of the action to recover damages for reputational harm outside of Canada, the motion judge found that the most substantial harm to his reputation was also in Ontario. The majority of the Court of Appeal agreed.

[ 87 ] Pepall J.A., dissenting at the Court of Appeal, concluded that “ lex loci delicti is too thin a strand on which to anchor choice of law in an Internet defamation case such as this one” (para. 179) and that, under the most substantial harm test, the law of Israel should govern the dispute. She would have found that the motion judge erred on the basis that he did not consider that the tort occurred in both Ontario and Israel, that there was no evidence of substantial harm to Goldhar’s reputation in Ontario and that he did not consider the principle of comity.

[ 88 ] As a tort has occurred in Ontario, Ontario law applies to the present action under the lex loci delicti rule. If, however, the action were to proceed in Israel, we can infer, relying upon the evidence of Dr. Tamar Gidron, a law professor at the Haim Striks School of Law in Israel, that Israeli courts would also apply their own law. As each forum would apply its own law, the applicable law factor cannot aid Haaretz in showing that it would be fairer and more efficient to proceed in the alternative forum.

[ 89 ] I recognize that, in Black and Banro , this Court considered only the applicable law in the chosen forum. I am concerned that disregarding the applicable law in the alternative forum is inconsistent with the comparative nature of the forum non conveniens analysis:

In many cases, including multi-jurisdiction defamation actions, different choice of law rules in each forum may well lead to different jurisdictions applying different substantive law. If the applicable law to the dispute is going to be used as a factor in the forum non conveniens analysis, then these different choice of law rules should be considered in order to properly determine whether in fact they can be said to favour one forum over the other.

(B. Kain, E. C. Marques and B. Shaw, “Developments in Private International Law: The 2011-2012 Term — The Unfinished Project of the Van Breda Trilogy” (2012), 59 S.C.L.R. (2d) 277, at p. 293)

[ 90 ] In any event, it is my view that applicable law, as determined by the lex loci delicti principle, should be accorded little weight in the forum non conveniens analysis in cases where jurisdiction is established on the basis of the situs of the tort. In circumstances where the situs of the tort leads to the assumption of jurisdiction in the chosen forum, lex loci delicti will inevitably also point to the chosen forum on the question of applicable law. This could be problematic, as this Court has clearly directed that the jurisdiction simpliciter and forum non conveniens analyses should be based on different factors ( Van Breda , at para. 56; see also J.-G. Castel, “Multistate Defamation: Should the Place of Publication Rule be Abandoned for Jurisdiction and Choice of Law Purposes?” (1990), 28 Osgoode Hall L.J. 153, at p. 163, and M. Castel, at pp. 154-55 and 160). Accordingly, applicable law is of little value in determining whether an alternative forum is clearly more appropriate in cases where jurisdiction is established on the basis of the situs of the tort. As such, while I would not disturb the motion judge’s conclusion that applicable law favours Ontario, this factor should be given little weight in the ultimate balancing.

[ 91 ] This would not be an appropriate case for this Court to adopt the place of most substantial harm test proposed in Banro , since, in my view, the submissions on this issue provide an insufficient basis upon which to create such an exception. Indeed, this Court should be reluctant to make such changes to the existing private international law framework as they may create legal uncertainty in a manner contrary to the objectives of conflicts rules ( Tolofson , at p. 1061).

[ 92 ] I recognize that in Internet defamation actions, where a tort may have occurred in multiple jurisdictions, the lex loci delicti rule may allow courts in multiple forums to assume jurisdiction and apply their own law. In an interconnected world where international players with global reputations are defamed through global publications, this is unsurprising.

[ 93 ] While I do not wish to discourage this Court from taking up this issue in a future case, it should do so only where this is necessary for the determination of the specific case before it and where appropriate evidence and argument are presented as to the impact of such a change.

[ 94 ] In concluding on this point, I would note that, in this case, the most substantial harm test would not have clearly favoured either forum. This is not a case such as the one contemplated in Tolofson , where the tort occurred in a different place than its consequences. The evidence is that Goldhar has a substantial reputation in Ontario, where his primary business interests lie, as well as a substantial reputation in Israel, where he enjoys a certain celebrity status by virtue of his ownership of a popular soccer team. While these reputations are qualitatively different, the evidence before me does not allow for a determination as to where Goldhar enjoyed the most substantial reputation or where the most substantial harm to that reputation occurred.

(7) Conclusion: Israel Is a Clearly More Appropriate Forum Than Ontario

[ 95 ] A robust and careful forum non conveniens analysis indicates that Haaretz would face substantial unfairness and inefficiency if a trial were held in Ontario. Goldhar’s interest in vindicating his reputation in Ontario fails to outweigh these concerns.

[ 96 ] A summary of my conclusions on each of the above elements of the forum non conveniens analysis is as follows:

(1) Comparative Convenience and Expense for the Parties favours Israel;

(2) Comparative Convenience and Expense for the Witnesses heavily favours Israel;

(3) Loss of Legitimate Juridical Advantage, while favouring Ontario, should not weigh heavily in the analysis;

(4) Fairness favours Israel;

(5) Enforcement slightly favours Israel; and

(6) Applicable law, while favouring Ontario, should be given little weight.

[ 97 ] Haaretz has established that holding a trial in Israel would be fairer and more efficient. Israel is clearly the more appropriate forum.

VI. Conclusion

[ 98 ] For all these reasons, I would allow the appeal and grant Haaretz’s motion to stay the action, with its costs in this Court and throughout.

The following are the reasons delivered by

[ 99 ] Karakatsanis J. — I agree with much of the reasoning of Côté J. and the conclusion she reaches. I write briefly to indicate my disagreement with certain aspects of her analysis relating to forum non conveniens .

[ 100 ] Under the applicable law factor, Côté J. considers the law that would apply if the action proceeded in Israel, as well as the law that would apply in Ontario (paras. 88-89). As my colleagues McLachlin C.J. and Moldaver and Gascon JJ. indicate, this approach is inconsistent with this Court’s jurisprudence and risks lengthening the forum non conveniens analysis (para. 207). It is also untethered from the rationale underlying the applicable law factor. The ultimate question that motivates this factor is whether the plaintiff’s chosen jurisdiction would be applying foreign law, which may diminish efficiency and raise a risk of forum shopping (see Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636, at para. 49 ). Assessing what law would apply in the alternative jurisdiction is not helpful to answering this question. That said, I agree with Côté J. that this factor holds little weight here, where jurisdiction and applicable law are both established on the basis of where the tort was committed.

[ 101 ] Further, my colleague Côté J. finds that Mr. Goldhar’s reputation in Israel is relevant to the exercise of jurisdiction (para. 24). She thus concludes that the motion judge erred in failing to take his reputation in Israel into account as part of the fairness factor (paras. 50 and 78). In my view, Goldhar’s Israeli reputation is not material to this factor, which is concerned with the plaintiff’s interest in vindicating his reputation in the jurisdiction where he enjoys it ( Banro , at para. 58; Breeden v. Black , 2012 SCC 19 , [2012] 1 S.C.R. 666, at para. 36 ; see also McLachlin C.J. and Moldaver and Gascon JJ., at para. 212). By bringing his claim in Ontario and undertaking to limit his claim to his Canadian reputation, Goldhar establishes that Ontario is where he enjoys and wishes to vindicate his reputation. Thus, I cannot agree that he would suffer no unfairness if he were forced to bring his claim in Israel. However, I agree with Côté J. that any unfairness to Goldhar is outweighed by fairness concerns over Haaretz’s ability to compel its witnesses’ testimony if the claim proceeds in Ontario (para. 79).

[ 102 ] On a related note, I cannot agree that Goldhar’s undertaking to limit his claim to his Canadian reputation and not bring a claim in another jurisdiction “should not be allowed to narrow the scope of his pleadings” (Côté J., at para. 23). Parties can and do narrow their claims as proceedings progress.

[ 103 ] In my view, however, the overall conclusion reached by Côté J. in her forum non conveniens analysis does not turn on any of the above elements. Like her, I would allow the appeal.

The following are the reasons delivered by

[ 104 ] Abella J. — Like Justice Côté, I would allow the appeal, but for somewhat different reasons.

[ 105 ] Haaretz argued that this case demonstrates why the standard approach to choice of law — as well as to jurisdiction — does not adequately respond to the unique issues and challenges raised by Internet defamation. It has, as a result, urged us to modify the test for choice of law and jurisdiction. In particular, it has invited the Court to take up LeBel J.’s invitation in Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636, that this Court modify the choice of law framework by replacing lex loci delicti with a test based on the place where the most substantial harm to the plaintiff’s reputation occurred. Since the very framework and almost all subsequent modifications have come from the courts, it is appropriate to respond positively to the invitation.

[ 106 ] The basis for choice of law in Canada is lex loci delicti , that is, where the tort occurred. The tort of defamation occurs when the alleged defamation is “published”. Publication occurs when material is read or downloaded by a third party. In the case of Internet defamation, therefore, a single download can determine which law applies. When combined with the standard framework for jurisdiction, which is also based on where the alleged defamation is published, this gives a plaintiff in Ontario an almost automatic entitlement to having an Ontario court assume jurisdiction over, and apply Ontario law to, an Internet defamation claim, regardless of the strength of the connection to Ontario.

[ 107 ] It seems to me that a more realistic approach would be one that narrows the range of potentially applicable law in a rational way (J.-G. Castel, “Multistate Defamation: Should the Place of Publication Rule be Abandoned for Jurisdiction and Choice of Law Purposes?” (1990), 28 Osgoode Hall L.J. 153, at p. 168).

[ 108 ] It is true that an Ontario court could always decline jurisdiction on the basis of forum non conveniens if the defendant is able to demonstrate that an alternative forum is “clearly more appropriate”. But whether another forum is clearly more appropriate depends, in part, on the law to be applied. And the law to be applied is, at the moment, governed by the choice of law rule for torts, namely, where the tort occurred.

[ 109 ] I agree that the standard framework for choice of law should be modified in a way that incorporates “most substantial harm to reputation”. This new approach would displace the law of the place of publication of the defamation with the law of the place with the most significant connection to the tort. In the case of Internet defamation, that will be the place where the plaintiff suffered the greatest harm to his or her reputation.

[ 110 ] A strict adoption of the lex loci delicti rule that makes each “publication” its own cause of action, contradicts La Forest J.’s acknowledgment in Tolofson v. Jensen , 1994 44 (SCC) , [1994] 3 S.C.R. 1022, that the overarching principle is that the applicable law is the one most closely connected to the wrong. La Forest J.’s focus, however, on “order and fairness” (p. 1058) emphasized order and, notably, predated the global reach of the Internet. The rigidity of that approach is hard to justify in circumstances where the applicable law could be the law of any country in which damage is suffered because the information is downloaded there (B. Kain, E. C. Marques and B. Shaw, “Developments in Private International Law: The 2011-2012 Term — The Unfinished Project of the Van Breda Trilogy” (2012), 59 S.C.L.R. (2d) 277, at p. 301) .

[ 111 ] It is worth remembering that even before Banro , this Court considered the possibility of establishing an exception to the general rule of lex loci delicti in the choice of law analysis. In Tolofson itself, La Forest J. acknowledged that the Court could establish exceptions involving acts that occurred in one place with the consequences being directly felt elsewhere, as well as situations where the wrong directly emerged from a transnational or interprovincial activity (p. 1050).

[ 112 ] Multijurisdictional Internet defamation fits squarely within this discussion.

[ 113 ] And adopting “most substantial harm” for choice of law would ensure that the choice of law rule reflects what is at the core of the tort of defamation — protection of reputation (J.-G. Castel (1990), at p. 160; see also C. Martin, “ Tolofson and Flames in Cyberspace: The Changing Landscape of Multistate Defamation” (1997), 31 U.B.C. L. Rev. 127, at p. 158). The practical implication of this approach is that the applicable law is restricted “to one, rather than potentially dozens [of laws] under a rule focusing on wherever the effects of the tort are felt” (Martin, at p. 158).

[ 114 ] In Banro , LeBel J. conceptualized the “most substantial harm” test according to the factors endorsed by the Australian Law Reform Commission in its 1979 Unfair Publication: Defamation and Privacy report (and codified in s. 11(3) of Australia’s Defamation Act 2005 (N.S.W.)). These factors include:

(a) the place at the time of publication where the plaintiff was ordinarily resident or, in the case of a corporation, the place where the corporation had its principal place of business at that time;

(b) the extent of publication in each relevant jurisdiction;

(c) the extent of harm sustained by the plaintiff in each relevant jurisdiction; and

(d) any other matter that the court considers relevant. [1]

[ 115 ] These Australian factors are helpful in outlining the kinds of considerations that can assist a court in determining the applicable law in cases of multijurisdiction defamation. I stress that they are only illustrative, and do not serve as a formulaic template.

[ 116 ] I also think the “centre of gravity” approach set out in the Opinion of the Advocate General Pedro Cruz Villalón of the European Court of Justice is helpful in Internet defamation cases. [2] Although the approach is applied in determining jurisdiction, its nuanced framework is useful in assessing where the harm occurred. The Advocate General articulated his approach in two joined references from France and Germany dealing with the interpretation of Article 5(3) of Council Regulation (EC) No. 44/2001 of 22 December 2000 on jurisdiction and the recognition and enforcement of judgments in civil and commercial matters , [2001] O.J. L. 12/1, which governs jurisdiction in civil matters in the European Community. [3] The conclusion was that a court may assume jurisdiction when it is the “centre of gravity of the dispute”. This will arise when two elements coincide. The first element concerns the individual’s reputation. The court will determine where the plaintiff has his or her “centre of interests”, that is, the state where “the victim is known [and] essentially carries out his life plan” (Opinion, at para. 59). The second element concerns the nature of the information published. The court will identify the state where the information is “objectively relevant”. This focuses on the place where the information is of greatest interest to readers. This is likely to be the place where it inflicts the most damage (Opinion, at paras. 60-61).

[ 117 ] Looking at both the Australian and European approaches, I think they address the harms of Internet defamation more realistically than our current approach to choice of law — or jurisdiction — according to which a single download is sufficient to establish the applicable law. It seems to me to be inherently un reasonable for an action to be heard where, relatively speaking, the harm to reputation was minor when the substantially greater harm to reputation occurred elsewhere. Applying the law where the most substantial harm occurred to the plaintiff’s reputation ensures respect for the purpose of defamation laws — freedom from reputational harm — as well as the reasonable expectations of parties — where did the publisher of the material expect any dispute to be resolved.

[ 118 ] Modifying the choice of law analysis to incorporate the “most substantial harm” test would not only render the extent of the harm to reputation relevant, it would also ensure that the reasonable expectations of the publisher of the statement alleged to be defamatory as to where it could expect to be sued are properly considered. It seems to me to be beyond the reasonable expectations of a publisher to answer to the laws of all of the states in which the plaintiff enjoys some reputation. A focus on where the most substantial harm occurred would address the potential unfairness of being held liable under the law of the plaintiff ’s chosen forum when the publisher complied with the law of a different forum (J.-G. Castel, “The Uncertainty Factor in Canadian Private International Law” (2007), 52 McGill L.J. 555, at p. 559, citing American Law Institute, Restatement of the Law, Second: Conflict of Laws (1971), at §6(g)).

[ 119 ] Adopting the place of most substantial harm to the plaintiff’s reputation when deciding the applicable choice of law would also arguably strike a better balance between freedom of expression and harm to reputation concerns. If choice of law rules are designed primarily to reflect the most characteristic element of the tort of defamation, namely the protection of reputation, then choice of law rules should “focus squarely on the law of the place where the reputation of the plaintiff has been most injured” (M. Castel, “Jurisdiction and Choice of Law Issues in Multistate Defamation on the Internet” (2013), 51 Alta. L. Rev. 153, at p. 160). While material posted on the Internet may harm an individual’s reputation in many places, there will only be one place where that harm hurts the most (M. Castel, at p. 161).

[ 120 ] Similar issues arise in connection with jurisdiction. It seems apparent to me, as it was to Haaretz, that there are symmetrical concerns between how the choice of law analysis proceeds and how jurisdiction is determined in cases of Internet defamation. In my view, while not strictly necessary to decide in this case, going forward it is worth considering whether the same approach should be applied to determining jurisdiction as the one I propose for choice of law.

[ 121 ] Jurisdiction is concerned with which court will hear the action. The purpose of the jurisdiction inquiry is to identify where the “real and substantial connection” between the subject matter of the litigation and the forum is strong enough such that the parties could reasonably have expected to sue or be sued there ( Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572, at para. 97 ).

[ 122 ] The first step in the standard framework for determining jurisdiction is to determine whether there is a “real and substantial connection” between the claim and Ontario, that is, whether the strength of the connection is sufficiently strong for a court to exercise authority over the claim. This is based on whether one of the four presumptive, rebuttable connecting factors set out in Van Breda is present. [4]

[ 123 ] The rationale LeBel J. offered for adopting the four presumptive connecting factors was to protect the reasonable expectations of the parties, namely, where the defendant would reasonably have expected to defend an action. As he indicated, ensuring fairness and protecting reasonable expectations means looking at the substance of the connections, a concern he acknowledged in Beals v. Saldanha , 2003 SCC 72 , [2003] 3 S.C.R. 416:

The test should ensure that, considering the totality of the connections between the forum and all aspects of the action, it is not unfair to expect the defendant to litigate in that forum . . . . There are situations where, given the other connections between the forum and the proceeding, it is a reasonable place for the action to be heard and the defendant can fairly be expected to go there even though he personally has no link at all to that jurisdiction. [Emphasis added; para. 182.]

[ 124 ] The inquiry is focused on the “reasonable expectations of the parties”, not of a “reasonable person”. As Joost Blom and Elizabeth Edinger explain:

What distinguishes the “real and substantial connection” test for jurisdiction from a concept like negligence or foreseeability is that it lacks . . . a clear psychological standpoint. Asking what a reasonable person would do in these circumstances, or what a reasonable person could foresee, requires the judge to put her or himself in the position of the mythical, but understandable, reasonable person and assess the facts from that point of view. The “real and substantial connection” test, however, requires the judge to adopt the view, not of a hypothetical person viewing the facts, but of an administrator whose mandate is to balance fairly the interests of the parties and legal systems involved.

(“The Chimera of the Real and Substantial Connection Test” (2005), 38 U.B.C. L. Rev. 373, at p. 416)

[ 125 ] The “reasonable contemplation of the parties” was set out by Dickson J. as the basis for the operative test, namely whether it was “inherently reasonable” for the action to be brought in a particular jurisdiction ( Moran v. Pyle National (Canada) Ltd. , 1973 192 (SCC) , [1975] 1 S.C.R. 393, at pp. 408-9). This turned into La Forest J.’s “real and substantial connection” test in Morguard Investments Ltd. v. De Savoye , 1990 29 (SCC) , [1990] 3 S.C.R. 1077, at p. 1108, which then morphed into the four presumptive, rebuttable factors in Van Breda . And even as this evolution took place, the basic principle never changed: Was it “inherently reasonable” for the action to proceed in that forum, that is, was it the forum the defendant/publisher ought reasonably to have had in its contemplation when it published the article that caused the injury/harm?

[ 126 ] The operative presumptive connecting factor in this case is that the tort was committed in Ontario. The tort of defamation is deemed to have occurred where the allegedly defamatory material was published. And, the standard test for publication is where the allegedly defamatory content is “read or downloaded by someone other than the plaintiff or the publisher” ( Banro , at para. 57). When defamation occurs on the Internet, where all it takes is one download, the tort is theoretically committed all over the world. As one academic commentator noted:

The problem with internet related torts and the determination of the place of commitment is the variety of available connecting factors: There are . . . the place where the information was generated, where it was uploaded, where it was downloaded, where the information was read or where the server hosting the information is located.

(S. Schmitz, “From Where are They Casting Stones? — Determining Jurisdiction in Online Defamation Claims” (2012), 6 Masaryk U. J.L. & Tech. 159, at p. 163)

[ 127 ] How then does one rebut the presumption when the tort is theoretically committed everywhere? How does a defendant show that the connection between the tort and Ontario is insufficient when all it takes to create the connection is one download in Ontario? And, if all it takes to create the connection is one download in Ontario, what does it take to rebut the presumption? The challenge raised by these questions is that the current approach seems to make the assumption of jurisdiction in Ontario automatic based on a single download.

[ 128 ] LeBel J. did not comment in Van Breda on what circumstances would rebut the presumption of jurisdiction. Instead, he merely noted that it remained open to the defendant to “establish facts” pointing to either no “real relationship” or a “weak relationship”, between the forum and the subject matter of the litigation, including the fact that only a “minor element” of the tort was committed in the jurisdiction (paras. 95-96).

[ 129 ] Since the “essence” of the harm in defamation is damage to reputation, [5] this leads me to conclude that the framework for determining jurisdiction should focus on where the plaintiff suffered the most substantial harm to his or her reputation. Such an approach, in my view, leaves room for concluding that the presumption can be rebutted if the defendant can show that the most harm to the plaintiff’s reputation occurred elsewhere. The inquiry into the most significant connection logically zeroes in on the severity of the harm to that reputation.

[ 130 ] This new approach also means that the choice of what test to apply for “real and substantial connection” no longer comes down to a choice between the attribution of a real and substantial connection wherever a defamatory act occurs, versus the fairness of ensuring that the dispute is resolved where there is in fact a real and substantial connection.

[ 131 ] Adopting the “most substantial harm” test for determining the choice of law under the forum non conveniens analysis, I am of the view that the place of most substantial harm to Mr. Goldhar’s reputation is clearly Israel, and that, as a result, Israeli law should apply.

[ 132 ] In his amended statement of claim, Mr. Goldhar refers to six statements as defamatory:

(a) To get a car, Mr. Angelides had to go to a team sponsor, behind Mr. Goldhar’s back.

(b) “Goldhar boasts to his business contacts in Toronto that he is not only the owner of Maccabi Tel Aviv but also its soccer director.”

(c) “He rented a dingy apartment for himself and drives nothing more than a Hyundai Getz.”

(d) “[Goldhar] cut out a cartoon of him that appeared in one paper [in the Greek press], asking all his employees whether it was flattering.”

(e) “Goldhar plays soccer at least once a week in Toronto with Ilan Sa’adi, a former professional player and close friend.”

(f) “Goldhar does not have a long term plan for the team.”

(A.R., vol. II, p. 5)

Of these, five focus on events and circumstances that concern Mr. Goldhar’s conduct and reputation in Israel, not Canada.

[ 133 ] The article [6] is essentially about Mr. Goldhar and his conduct in Israel. It was about an Israeli soccer team owned by Mr. Goldhar, his involvement in his own team’s management and his relationship with his players, coaches and trainers in Israel . It was researched, written and edited in Israel , addressed to an Israeli audience, and focused on someone who is a public figure there. Any information written about the team and Mr. Goldhar would have a far greater impact on his reputation in Israel than in Canada.

[ 134 ] Although Mr. Goldhar spends most of his time in Canada, he maintains an apartment in Israel which he visits “about five or six times per year”, and his connection to Israel is significant. He is the owner of Maccabi Tel Aviv Football Club, one of, if not the most popular soccer teams in Israel. The many articles which have been written about Maccabi Tel Aviv refer to, or feature Mr. Goldhar. They also form part of the broader Israeli media landscape in which Mr. Goldhar has a very high public profile.

[ 135 ] The extent of the article’s publication in Israel clearly overshadowed the extent of the article’s publication in Ontario. The record showed that between 200 and 300 people in Canada read the article online whereas approximately 70,000 people read the article in Israel. It is obvious from these numbers too that any reputational harm to Mr. Goldhar was overwhelmingly greater in Israel.

[ 136 ] This brings me to the rest of the forum non conveniens analysis. On the basis that Israeli law applies, I agree with Côté J. that Haaretz has successfully demonstrated that Israel is the “clearly more appropriate” forum ( Van Breda , at para. 108).

[ 137 ] The forum non conveniens analysis authorizes courts to “go beyond a strict application of the test governing the recognition and assumption of jurisdiction” ( Van Breda , at para. 104). This not only assures fairness to the parties; it also guarantees the efficient resolution of the dispute ( Van Breda , at para. 104). In Van Breda , the Court set out several non-exhaustive factors as being relevant to determining whether forum non conveniens applies (para. 110). And, these factors may vary depending on the context ( Lapointe Rosenstein Marchand Melançon LLP v. Cassels Brock & Blackwell LLP , 2016 SCC 30 , [2016] 1 S.C.R. 851, at para. 53 ).

[ 138 ] I am of the view that all of the remaining factors raised before the Court — the comparative convenience/expense to the parties and witnesses, juridical advantage, fairness and enforcement — favour Israel.

[ 139 ] A libel trial in Ontario would place a significant financial strain on the newspaper and the other defendants, all of whom are based in Israel and have no assets in Canada. Given the absence of evidence regarding any inconvenience or undue expense for Mr. Goldhar, the factor of comparative convenience/expense clearly supports a trial in Israel.

[ 140 ] Similarly, considerations relating to the comparative convenience/expense to the witnesses point to the trial taking place in Israel. Mr. Goldhar filed no information before the motion judge about the witnesses he would call at trial. Haaretz, on the other hand, filed a list of 22 witnesses, 18 of whom live in Israel. As for Mr. Goldhar’s undertaking to fund the travel and accommodation costs of Haaretz’s foreign witnesses in accordance with the rates provided in the Rules of Civil Procedure , R.R.O. 1990, Reg. 194, Tariff A, I think it would be tantamount to permitting parties with greater resources to tip the scales in their favour by “buying” a forum. While the individual circumstances of each party are clearly relevant to any balancing, it is their actual circumstances, and not artificially created ones, that should be weighed. I am also of the view that the absence of a jury trial in Israel has no effect on the quality of civil justice available in Israel.

[ 141 ] Turning to the factor of fairness, a singular focus on Ontario as being the place where Mr. Goldhar lives and works overlooks the reality that he owns a prominent business in Israel that attracts significant public attention in Israel and brings him to Israel several times per year. Regardless of where he spends most of his time, he still spends much of it in Israel and is a known and active participant in Israeli life and society. On the other hand, a trial in Ontario would put significant financial burdens on Haaretz.

[ 142 ] It is also clear to me that enforcement concerns would favour a trial in Israel, in large part because Haaretz’s lack of assets in Ontario would mean that any order made against it would have to be enforced by Israeli courts, thereby raising concerns about a multiplicity of proceedings.

[ 143 ] I therefore agree that the appeal should be allowed.

The following are the reasons delivered by

[ 144 ] Wagner J . — Haaretz urged this Court to modify the choice of law rule for the tort of Internet defamation, from lex loci delicti to a test based on the place where the most substantial harm to the plaintiff’s reputation occurred. For substantially the reasons given by my colleague Abella J., I agree that this Court should make such a modification to the choice of law rule in the specific context of Internet defamation. As private international law in common law Canada is almost entirely judge-made law, I see no need to wait for legislative initiative in this area or for the completion by the Law Commission of Ontario of its reform project. In Éditions Écosociété Inc. v. Banro Corp ., 2012 SCC 18 , [2012] 1 S.C.R. 636, at paras. 58-62 , LeBel J. noted that the importance of place of reputation has long been recognized in Canadian defamation law. Despite citing several commentators in favour of the idea, LeBel J. nonetheless left the question of whether to modify the choice of law rule for multijurisdictional defamation cases for “another day”. That day has now arrived.

[ 145 ] I agree with Abella J. that the factors which have been codified in Australia, and the “centre of gravity” approach set out in an opinion for the European Court of Justice, provide useful guidance as to how the place of most substantial harm to reputation test is to be applied in practice. It may be that in certain cases it will be challenging to identify the place of most substantial reputational harm. However, the range of possibly applicable law for a given dispute will be much narrower than with lex loci delicti and will be determined on a more principled basis.

[ 146 ] I further agree that adopting this new test for choice of law would have several positive effects. As discussed in more detail by my colleague Abella J., these positive effects include ensuring that the reasonable expectations of the publisher of the statement alleged to be defamatory are properly considered, striking a better balance between freedom of expression and harm to reputation concerns, and ensuring that choice of law will reflect the purpose of defamation laws. Adopting this new choice of law test will not result in a heavy evidentiary burden for the parties. The plaintiff’s reputation is already a relevant consideration during the rebuttal stage of the analysis and in relation to several factors other than choice of law during the forum non conveniens analysis.

[ 147 ] Academic commentators and this Court, in its past decisions, have discussed the “most substantial harm” test in the context of choice of law, not jurisdiction simpliciter . This is because case law from this Court has established that more than one forum may have jurisdiction over a given dispute. The inquiry at the jurisdiction simpliciter stage of the analysis is simply whether there is a “real and substantial connection” between the dispute and the Canadian forum, not whether the “real and substantial connection” between the dispute and the Canadian forum is greater than that between the dispute and any other forum. I see no reason why this should be different in the context of Internet defamation. I cannot agree with an approach whereby a Canadian court would conclude that it does not have jurisdiction over a dispute with significant connections to Canada, including potentially significant reputational harm suffered in Canada, simply because greater reputational harm occurred elsewhere.

[ 148 ] The advantage of adopting the new test solely for choice of law purposes is that choice of law is just one amongst a range of factors considered during the forum non conveniens analysis. If the applicable law is that of another forum where more substantial reputational harm has been suffered, this will support a finding that the other forum is clearly more appropriate for litigating the dispute. However, after considering all the factors, a court may conclude that the other forum is not clearly more appropriate than the Canadian forum. In such cases, although it will retain jurisdiction, the Canadian court will nonetheless apply the law of that other forum where greater reputational harm occurred.

[ 149 ] It is entirely consistent with Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572, for Canadian courts to find the presumption of jurisdiction rebutted where there are no connections between the plaintiff and the Canadian forum beyond a small number of acts of publication. It is true that, despite the importance of the rebuttal stage, this Court recognized in Van Breda , at para. 109, that “jurisdiction may sometimes be established on a rather low threshold under the conflicts rules”. This is simply inherent to private international law. In my view, any concerns raised by the unique nature of Internet defamation are best addressed by changes to the choice of law rule.

[ 150 ] In this case, as set out in the reasons of my colleague Abella J., when the “most substantial harm” test is applied to these facts, Israel is the “clearly more appropriate” forum. Accordingly, I would allow the appeal.

The following are the reasons delivered by

[ 151 ] The Chief Justice and Moldaver and Gascon JJ. (dissenting) — Distilled to its essence, this case boils down to a single question. When a Canadian citizen is allegedly defamed for his Canadian business practices — in an article published online in his home province by a foreign newspaper — is he entitled to vindicate his reputation in the courts of the province where he lives and maintains his business, and where the sting of the article’s comments is felt? The answer of the motion judge and of the majority in the Court of Appeal was yes. We agree, and would accordingly dismiss the appeal.

I. C ontext

[ 152 ] Mr. Goldhar is a prominent Canadian businessman who lives in Toronto. For about 20 years, he has operated a real estate business in Ontario and participated actively in the Toronto community. Since 2009, he has also owned the Israeli-based Maccabi Tel Aviv Football Club, one of Israel’s most popular professional soccer teams.

[ 153 ] In November 2011, Haaretz, an Israeli newspaper, published an article that included disparaging statements about Mr. Goldhar. The main topic of the article was the way he runs Maccabi Tel Aviv. For reasons best known to itself, in that context, Haaretz chose to publish gratuitous comments about Mr. Goldhar’s Canadian business enterprises and his management of them. The article identified Mr. Goldhar as the “Canadian owner” of the soccer club, and claimed that his “management model was imported directly from his main business interest — a partnership with Wal-Mart to operate shopping centers in Canada” (reproduced in 2016 ONCA 515 , 132 O.R. (3d) 331, Appendix “A”). The article went on to suggest that this managerial culture — allegedly imported from his Canadian businesses — was “based on overconcentration bordering on megalomania, penny-pinching and a lack of long-term planning”.

[ 154 ] This article came to the attention of 200 to 300 Canadian readers — including employees of Mr. Goldhar’s business in Ontario — through Haaretz’s English language website. In December 2011, Mr. Goldhar commenced a libel action in Ontario against Haaretz. The newspaper countered the proceedings with a motion to stay the action for lack of jurisdiction 7 or, alternatively, to stay the proceedings on the ground of forum non conveniens .

II. Judicial History

A. Ontario Superior Court of Justice, 2015 ONSC 1128 , 125 O.R. (3d) 619 (Faieta J.)

[ 155 ] The motion judge dismissed Haaretz’s motion. He determined that Ontario courts have jurisdiction, finding that the two-step test established in Club Resorts Ltd. v. Van Breda , 2012 SCC 17 , [2012] 1 S.C.R. 572, was met. Because 200 to 300 people in Canada had read the article, the tort of defamation was committed in Ontario, presumptively establishing the jurisdiction of the Ontario courts. This presumption was not rebutted.

[ 156 ] The motion judge further found that Haaretz had not established that Israel was a clearly more appropriate forum than Ontario, weighing the following factors: comparative convenience and expense for the parties (which favoured Israel); comparative convenience and expense for the witnesses (which slightly favoured Israel); choice of law (which favoured Ontario); loss of legitimate juridical advantage (which favoured Ontario); and fairness to the parties (which favoured Ontario).

[ 157 ] Finally, he found that Mr. Goldhar’s action was “far from being an abuse of process” (para. 76).

B. Court of Appeal for Ontario, 2016 ONCA 515 , 132 O.R. (3d) 331 (Simmons, Cronk and Pepall JJ.A.)

(1) Majority Reasons of Simmons and Cronk JJ.A.

[ 158 ] The majority in the Court of Appeal dismissed Haaretz’s appeal. First, it agreed with the motion judge on the question of jurisdiction simpliciter . Second, it found that the motion judge made two errors in his forum non conveniens analysis regarding the relevance of Mr. Goldhar’s intent to have a jury trial and regarding the effect of letters rogatory. The majority was, however, persuaded that these errors were not significant to the overall conclusion. It therefore agreed that Israel was not a clearly more appropriate forum than Ontario.

(2) Dissenting Reasons of Pepall J.A.

[ 159 ] The dissenting judge, Pepall J.A., agreed with the majority on jurisdiction simpliciter , but found that Israel was a clearly more appropriate forum than Ontario. She held that the ease with which jurisdiction can be established in Internet defamation cases requires a “robust and carefully scrutinized review” at the forum non conveniens stage (para. 132). Applying this robust approach, she concluded that the motion judge’s analysis was “infected by errors” and found that all factors except one favoured Israel (para. 137).

[ 160 ] Importantly, the dissenting judge stated that the test for choice of law — a factor in the forum non conveniens analysis — should be modified. She was of the view that the law of the place of the most substantial harm to reputation should apply to defamation cases, rather than the law of the place where the tort was committed ( lex loci delicti ). She found that, according to the place of most substantial harm rule, Israeli law would be applicable in this case.

III. Issues

[ 161 ] Our analysis is divided into two parts: jurisdiction simpliciter and forum non conveniens . In the first part, we explain why the test for jurisdiction simpliciter is met here, and how the current rules that govern its application accommodate multijurisdictional defamation cases, with no need to apply a robust review at the forum non conveniens stage. In the second part, we explain why the high threshold set by the “clearly more appropriate” test is not met in this case, and the reasons why this Court should not adopt a place of most substantial harm rule for the applicable law in multijurisdictional defamation cases.

IV. Analysis

[ 162 ] A preliminary issue must be addressed regarding the characterization of this proceeding. The statement of claim defines what is at issue in a given case (see Lax Kw’alaams Indian Band v. Canada (Attorney General) , 2011 SCC 56 , [2011] 3 S.C.R. 535, at para. 41 ). For the purposes of analysing the assumption and exercise of jurisdiction, however, it is not considered in isolation. While the statement of claim defines the issues in the case, subsequent representations and undertakings that limit the scope of the plaintiff’s action are relevant to the overall determination.

[ 163 ] Justice Côté parses each line of Mr. Goldhar’s claim in an effort to show that his concern about his business reputation in Canada is simply an afterthought. This ignores the paragraphs of the statement of claim that pertain to the connections of Mr. Goldhar and of the article with Ontario. More fundamentally, that approach disregards his position, taken early in the process, that the sting of the libel is felt in the province where he lives and maintains his business. With respect, in focussing so intently on the broader wording of Mr. Goldhar’s statement of claim, our colleague takes a formalistic approach that bears little relationship to how this case has been fought in the courts below and in this Court. It is simply not true to state that Mr. Goldhar’s “action was never limited to damage to his reputation in Ontario” (Côté J., at paras. 23 and 78). Mr. Goldhar has undertaken to limit his claim to his Canadian reputation. He has gone so far as to state, on the record before the motion judge, that he has no interest in seeking damages for any loss of reputation that he may have suffered in Israel, and his counsel has affirmed before this Court that it would be an abuse of process if he were to pursue a damages claim in Israel.

A. Jurisdiction Simpliciter

[ 164 ] To assume jurisdiction over a case, a court must be satisfied that a “real and substantial connection” exists between the subject matter of the litigation and the forum in which the proceeding is brought ( Morguard Investments Ltd. v. De Savoye , 1990 29 (SCC) , [1990] 3 S.C.R. 1077; Van Breda , at paras. 22-34; Canadian Encyclopedic Digest (Ont. 4th ed. (loose-leaf)), vol. 10, at §62.1). The analysis has two stages: first, identifying whether a presumptive connecting factor exists that prima facie entitles the court to assume jurisdiction over the dispute, and second, determining whether the presumption of jurisdiction is rebutted on the facts of that case.

(1) The Presumptive Connecting Factors

[ 165 ] In Van Breda , this Court sought to respond to dissatisfaction with the “real and substantial connection” test by bringing “greater clarity and predictability to the analysis of the problems of assumption of jurisdiction” (para. 78; see also para. 67). To that end, it listed a number of presumptive connective factors that prima facie entitle a court to assume jurisdiction over a dispute (para. 90).

[ 166 ] Here, Mr. Goldhar established the existence of such a presumptive connecting factor: the commission of a tort in Ontario. In Canadian law, “the tort of defamation occurs upon publication of a defamatory statement to a third party” ( Breeden v. Black , 2012 SCC 19 , [2012] 1 S.C.R. 666, at para. 20 ; see also Éditions Écosociété Inc. v. Banro Corp. , 2012 SCC 18 , [2012] 1 S.C.R. 636, at para. 34 ). Contrary to Haaretz’s submissions, there is no valid reason to reconsider or set aside this presumptive connecting factor. As stressed in Van Breda , “[t]he situs of the tort is clearly an appropriate connecting factor” (para. 88). Commission of a tort in the jurisdiction remains a sound basis on which to establish prima facie jurisdiction even in the context of Internet defamation cases, because the sting of the defamation is felt in the place where it is read. The framework recently established by this Court in Van Breda cannot ensure the clarity and predictability it is meant to achieve unless it is applied consistently.

[ 167 ] In this case, it is not contested that the allegedly libellous article was consulted by 200 to 300 people in Canada, including employees of Mr. Goldhar’s business in Ontario. It is apparent that a tort of defamation was committed in Ontario, clearly establishing a presumptive connecting factor. There was therefore a presumption that Ontario courts could properly assume jurisdiction over the dispute.

(2) Rebuttal of the Presumption of Jurisdiction

[ 168 ] In Van Breda , this Court took care to explain that the presumption of jurisdiction is not irrebuttable (paras. 81 and 95-100). The burden of rebutting this presumption rests, however, on the party challenging the court’s jurisdiction. To be successful, it must “establish facts which demonstrate that the presumptive connecting factor does not point to any real relationship between the subject matter of the litigation and the forum or points only to a weak relationship between them” (para. 95).

[ 169 ] The foreseeability of the defendant being called to answer proceedings in the forum is key to whether a court may properly assume jurisdiction over the case. As indicated in Van Breda , this informed the selection of the presumptive connecting factors:

All presumptive connecting factors generally point to a relationship between the subject matter of the litigation and the forum such that it would be reasonable to expect that the defendant would be called to answer legal proceedings in that forum . [Emphasis added; para. 92.]

[ 170 ] Reasonable foreseeability is similarly central at the rebuttal step of the analysis. Notwithstanding the existence of a presumptive connecting factor, the defendant may establish a lack of jurisdiction where there is no “real and substantial connection”, but only a weak relationship between the subject of the litigation and the forum. The strength of that relationship is informed by the reasonable foreseeability of the claim proceeding in that jurisdiction. After providing a number of examples where the presumption of jurisdiction would be rebutted, the Court in Van Breda explained:

In each of [these] examples, it is arguable that the presumptive connecting factor points to a weak relationship between the forum and the subject matter of the litigation and that it would accordingly not be reasonable to expect that the defendant would be called to answer proceedings in that jurisdiction . In such circumstances, the real and substantial connection test would not be satisfied and the court would lack jurisdiction to hear the dispute. [Emphasis added; para. 97.]

[ 171 ] Without this important check of reasonable foreseeability of being sued in the jurisdiction, the application of the presumptive connecting factor of the commission of a tort in the jurisdiction could raise concerns of forum shopping ( Banro , at para. 34). The reasonable foreseeability of being sued in a jurisdiction where the impugned statements have caused harm is therefore an important limit on the ease with which jurisdiction can be presumptively assumed in defamation cases (see Paulsson v. Cooper , 2011 ONCA 150 , 105 O.R. (3d) 28, at para. 34 ; Barrick Gold Corp. v. Blanchard & Co. (2003) 2003 64238 (ON SC) , 9 B.L.R. (4th) 316 (Ont. S.C.J.), at paras. 42-45 ). This is especially true with respect to the publication of allegedly defamatory statements over the Internet, where such publications can often be accessed worldwide. Assumption of jurisdiction is therefore far from being “ automatic based on a single download” (Abella J., at para. 127 (emphasis in original)).

[ 172 ] In the present case, it was more than reasonably foreseeable that Haaretz would be sued in Ontario. The newspaper published an article attacking a Canadian who lives and does business in Ontario. We do not have to decide at this stage whether the statements published by Haaretz are libellous. We simply have to locate where the sting of the article truly is. In this respect, one must not be distracted by the remainder of the article; the heart of the dispute at hand is the corrosive and highly critical comments about Mr. Goldhar’s management style, allegedly imported from his Canadian business.

[ 173 ] Furthermore, Haaretz made the article readily available not only to readers in Israel, but also to readers worldwide through online publication on its website. While it is true that defamation cases may raise forum shopping concerns, especially in the Internet context, the present case is clearly not one of forum shopping. It is entirely foreseeable that a Canadian citizen and resident would want to vindicate his Canadian reputation as the owner of his Canadian businesses in a Canadian court. The facts undeniably reveal a real and substantial connection between this case and Ontario. Therefore, the presumption of jurisdiction was not rebutted.

[ 174 ] That said, contrary to what the dissenting judge in the Court of Appeal suggested, jurisdiction simpliciter readily accommodates multijurisdictional defamation cases, even in the Internet age. While it may be true that a presumptive connecting factor may be established “virtually automatically” in Internet defamation cases, a court does not necessarily assume jurisdiction. If there is no real and substantial connection between the action and the forum, the presumptive connecting factor would be rebutted, which is likely to be the case where the plaintiff is forum shopping. If the analysis at the rebuttal stage is done properly, with an adequate consideration of reasonable foreseeability, there is no need to apply a “robust and carefully scrutinized” forum non conveniens analysis, as suggested by the dissenting judge in the Court of Appeal and by our colleague Justice Côté (paras. 3, 48 and 95). As we will explain in the following section, this new standard would frustrate the predictability and stability that is at the core of the Van Breda framework.

[ 175 ] As indicated, in this case, the presumption has not been rebutted, and the Ontario court has jurisdiction. However, even when jurisdiction is established, the court retains the discretion to decline jurisdiction under the doctrine of forum non conveniens . In this case, the Ontario courts did not exercise this discretion, correctly, in our view. We turn to this next.

B. Forum Non Conveniens

[ 176 ] The doctrine of forum non conveniens relates to a court’s discretionary power not to exercise its jurisdiction, in certain circumstances, in order to assure fairness to the parties and the efficient resolution of the dispute ( Van Breda , at para. 104).

[ 177 ] A motion judge’s discretionary decision whether or not to decline jurisdiction on the basis of forum non conveniens is entitled to considerable deference on appeal ( Lapointe Rosenstein Marchand Melançon LLP v. Cassels Brock & Blackwell LLP , 2016 SCC 30 , [2016] 1 S.C.R. 851, at para. 54 ; Van Breda , at para. 112; Banro , at para. 41). It is true that errors of law may present grounds for intervention. However, having appellate courts apply the new “robust and carefully scrutinized” approach would disregard the discretionary nature of forum non conveniens decisions and the applicable standard of review.

[ 178 ] Admittedly, the motion judge here made two legal errors in his forum non conveniens analysis: first, regarding the relevance of Mr. Goldhar’s intent to have a jury trial and second, regarding the effect of letters rogatory. Normally, no deference would be owed to the motion judge’s decision on the factors affected by these errors — i.e., comparative convenience and expense for witnesses, and loss of legitimate juridical advantage — and on the overall balancing. That said, the two errors made by the motion judge have no impact on the result in this case. Indeed, as will be explained in further detail below, the first has been remedied and the second is immaterial since Haaretz did not meet its burden of proof.

[ 179 ] At para. 50 of her reasons, Justice Côté lists six “errors” made by the motion judge. However, four of these “errors” are merely points where our colleague would have weighed the evidence differently had she been the motion judge. Justice Côté criticizes the motion judge for not placing significant weight on some factors and for discounting others (see para. 50). This is not the role of an appellate court, and is at odds with the deferential standard of review.

[ 180 ] Our concern with Justice Côté’s approach is that it undermines stability and increases costs and uncertainty for parties. Professors Pitel and Rafferty stress that motion judges’ decisions on forum non conveniens are often reversed on appeal, which

indicates how different judges can reach different results on the same facts. This reduces predictability and confidence in the litigation process. Ultimately, the doctrine has led to parties spending more time on preliminary jurisdictional issues, which delays moving the dispute forward and addressing the merits.

(S. G. A. Pitel and N. S. Rafferty, Conflict of Laws (2nd ed. 2016), at pp. 118-19)

With respect, by interfering with the motion judge’s assessment of the evidence on each and every factor of the analysis, our colleague lends support to this problematic approach.

[ 181 ] This Court must refrain from interfering with the motion judge’s exercise of his discretionary power or assessment of the evidence where it is not tainted by any error. In this regard, as we will explain, we agree with the majority in the Court of Appeal that the two errors he made were not significant to the overall conclusion on forum non conveniens .

[ 182 ] As the party seeking to stay the proceedings on the ground of forum non conveniens , Haaretz bears the burden of demonstrating that Israel has a real and substantial connection with the case, and that it is a clearly more appropriate forum than Ontario ( Lapointe , at para. 52; Van Breda , at paras. 102-3; Pitel and Rafferty, at pp. 121-22). As the history and application of the forum non conveniens doctrine show, the “clearly more appropriate” standard was intended as a high threshold for displacing the plaintiff’s chosen forum.

(1) Origin of the “Clearly More Appropriate” Test

[ 183 ] The basis of the forum non conveniens analysis is the “clearly more appropriate” test. As indicated, it sets a high threshold for displacing the forum chosen by the plaintiff. To understand how the test operates, it is important to return to the historical reasons that led courts to adopt such a stringent standard.

(a) The Scottish and English Roots of the “Clearly More Appropriate” Test

[ 184 ] Historically, English courts applied a two-part test for jurisdiction, where the defendant who opposed the plaintiff’s choice of forum had to establish: “. . . (1) that the continuation of the action would cause an injustice to him or her because it would be oppressive or vexatious or constitute an abuse of the process, and (2) that [a] stay would not cause an injustice to the plaintiff” ( Amchem Products Inc. v. British Columbia (Workers’ Compensation Board) , 1993 124 (SCC) , [1993] 1 S.C.R. 897, at p. 915; see e.g . Egbert v. Short , [1907] 2 Ch. 205 ; St. Pierre v. South American Stores (Garth and Chaves), Limited , [1936] 1 K.B. 382 (C.A.), at p. 398) .

[ 185 ] In Rockware Glass Ltd. v. MacShannon , [1978] 2 W.L.R. 362 , the House of Lords discarded this test in favour of an approach substantially similar to the one that had originated in Scotland ( Spiliada Maritime Corporation v. Cansulex Ltd. , [1987] 1 A.C. 460 (H.L.), at p. 474) . Scottish courts could “decline to exercise jurisdiction, after giving consideration to the interests of the parties and the requirements of justice, on the ground that the case [could not] be suitably tried in the Scottish court nor full justice be done there, but only in another court” ( Dicey, Morris and Collins on the Conflict of Laws (15th ed. 2012), by Lord Collins of Mapesbury, at §12-007). In Spiliada , the House of Lords refined this test and adopted the current language. This test was meant to take into account the fact that the jurisdiction of English courts had already been established at this stage:

In my opinion, the burden resting on the defendant is not just to show that England is not the natural or appropriate forum for the trial, but to establish that there is another available forum which is clearly or distinctly more appropriate than the English forum. In this way, proper regard is paid to the fact that jurisdiction has been founded in England as of right. [Emphasis added; p. 477.]

(b) The “Clearly More Appropriate” Test in Canadian Law

[ 186 ] The use of the “clearly more appropriate” test in Canadian law originated in Amchem . In that decision, this Court adopted and built on the forum non conveniens test as articulated by the House of Lords in Spiliada (p. 921). After a careful review of the international jurisprudence on this topic, this Court stated that it “agree[d] with the English authorities that the existence of a more appropriate forum must be clearly established to displace the forum selected by the plaintiff” ( ibid. (emphasis in original)). Later, it stated: “. . . the court must determine whether there is another forum that is clearly more appropriate ” (p. 931 (emphasis added)).

[ 187 ] The Court highlighted the increasingly international character of business and litigation. It expressed the view that “it has become more difficult to identify one clearly appropriate forum for this type of litigation” (p. 911). It added that, “[f]requently, there is no single forum that is clearly the most convenient or appropriate for the trial of the action but rather several which are equally suitable alternatives” (p. 912).

[ 188 ] More recently, in Van Breda , the Court reiterated the “clearly more appropriate” test. It stressed that the expression “clearly more appropriate” was chosen instead of “more appropriate” in order to emphasize the exceptional character of the circumstances that would warrant a court to decline to exercise its jurisdiction over a case (paras. 108-9). It opined that “the normal state of affairs is that jurisdiction should be exercised once it is properly assumed”, and added “[i]t is not a matter of flipping a coin” (para. 109). The inclusion of the adverb “clearly” in the test was not a stylistic caprice. It serves the key purpose of indicating the high threshold the Court wanted to establish, categorically rejecting the notion that a court should stay a proceeding where another forum is merely more appropriate.

[ 189 ] In Van Breda , this Court emphasized that the doctrine of forum non conveniens comes into play after the plaintiff establishes that the forum has jurisdiction (para. 101). The Court stressed that this doctrine “is based on a recognition that a common law court retains a residual power to decline to exercise its jurisdiction in appropriate, but limited , circumstances in order to assure fairness to the parties and the efficient resolution of the dispute” (para. 104 (emphasis added)). The forum that has jurisdiction will exercise it in the absence of a compelling reason not to do so.

[ 190 ] Since Amchem , this Court has constantly and consistently reiterated the “clearly more appropriate” forum test and the high threshold it implies ( Spar Aerospace Ltd. v. American Mobile Satellite Corp. , 2002 SCC 78 , [2002] 4 S.C.R. 205, at para. 70 ; Unifund Assurance Co. v. Insurance Corp. of British Columbia , 2003 SCC 40 , [2003] 2 S.C.R. 63, at para. 137 ; Black , at para. 37 ; Banro , at para. 64; Lapointe , at para. 52). As stated in Unifund , “[i]f neither forum is clearly more appropriate, the domestic forum wins by default” (para. 137).

[ 191 ] Given this history and the consistent application of the test, we should not lower the purposefully stringent threshold set by the “clearly more appropriate” test, whether through lenient application or through a “robust and carefully scrutinized review” such as the one suggested by the dissenting judge in the Court of Appeal.

(2) Factors to Consider in the Forum Non Conveniens Analysis

[ 192 ] In Van Breda , this Court highlighted the contextual nature of the forum non conveniens analysis: “. . . the factors that a court may consider in deciding whether to apply forum non conveniens may vary depending on the context” (para. 110). In this case, the motion judge considered the following factors, which closely mirror the ones considered by this Court in Black , another multijurisdictional defamation case: applicable law; fairness to the parties; comparative convenience and expense for the parties and witnesses; and loss of legitimate juridical advantage. In Black , this Court also considered two further factors: avoidance of a multiplicity of proceedings and conflicting decisions; and enforcement of the judgment (paras. 34-35).

[ 193 ] We will discuss these factors in turn and explain why they do not meet the test of showing that Israel is a clearly more appropriate forum than Ontario.

(a) Applicable Law

[ 194 ] The applicable law in tort cases is generally the lex loci delicti — the law of the place where the tort occurred ( Tolofson v. Jensen , 1994 44 (SCC) , [1994] 3 S.C.R. 1022, at p. 1050). In defamation cases, this means the jurisdiction where the impugned statements were published to a third party. In this case, there is no dispute that the article was accessed by hundreds of readers in Canada, including several people in Ontario.

[ 195 ] Here, two issues arise with respect to applicable law. The first is whether this Court should adopt the place of most substantial harm rule for the applicable law in multijurisdictional defamation cases. The second is the proper approach to balancing the applicable law factor in the forum non conveniens analysis.

(i) This Court Should Not Adopt the Place of Most Substantial Harm Rule

[ 196 ] Haaretz argues that this Court should adopt the place of most substantial harm rule for the applicable law in defamation cases. It relies on Banro , where this Court declined to change the choice of law rule, but noted, in obiter , that the place of most substantial harm rule could potentially be an alternative to lex loci delicti in defamation cases (para. 56).

[ 197 ] Contrary to the motion judge and the majority in the Court of Appeal, the dissenting judge agreed with Haaretz and concluded that the place of most substantial harm rule should replace lex loci delicti as the choice of law rule applicable to defamation cases. She concluded that the place of most substantial harm in this case was Israel — therefore, in her view, this factor weighed heavily in favour of Israel. This is the only basis upon which she found that the applicable law would not favour Ontario.

[ 198 ] Despite the views of the dissenting judge in the Court of Appeal and of our colleagues Justices Abella and Wagner, we consider it both unwise and unnecessary for this Court to adopt the place of most substantial harm rule for the applicable law in place of lex loci delicti . We reject the place of most substantial harm rule for four reasons: it does not point predictably to one jurisdiction, it would lead to complex preliminary motions, it received only limited support in the Canadian academic literature and jurisprudence, and its adoption in Australia is an insufficient basis for overhauling our own law in this area.

[ 199 ] First, the place of most substantial harm rule is highly subjective, and will not reliably point to one jurisdiction. This rule does not provide a clear answer where a person lives and maintains an important reputation in one jurisdiction, but acts — and is the subject of defamatory statements — in another jurisdiction. The place of most substantial harm is indeed opaque in this case, particularly when one considers Mr. Goldhar’s undertaking to limit his claim to his Canadian reputation. It is telling that the judges below disagreed on the jurisdiction where the place of most substantial harm was felt (see motion judge, at para. 47; majority in the Court of Appeal, at paras. 86-87; Pepall J.A., at para. 181). The factual finding of the motion judge that the most substantial harm to the plaintiff’s reputation is in Ontario cannot be displaced absent palpable and overriding error.

[ 200 ] A similar concern is raised, in the context of multijurisdictional defamation cases, with respect to lex loci delicti : it may not point to a single law and will therefore fail to curb forum shopping ( Banro , at paras. 49 and 60; Australia, Law Reform Commission, Unfair Publication: Defamation and Privacy (1979), at para. 339). But, as indicated, that rationale has no application in this case. Forum shopping is simply not an issue here. In any event, as discussed above, the place of most substantial harm rule would also be of limited use in curbing forum shopping — the inquiry is highly subjective, and it would not, in many cases, point predictably to one law.

[ 201 ] Second, the place of most substantial harm rule would result in the proliferation of “mini-trials” requiring substantial evidence at this preliminary stage of the proceedings. We should be cautious about adding elements that must be proved at the jurisdictional stage, as this would increase delay and expense. Adopting the place of most substantial harm rule would require parties to defamation actions to establish the extent of the plaintiff’s reputation in both jurisdictions and the harm done to this reputation as a result of the allegedly defamatory statements. That hardly makes sense when one considers that onerous evidence of this sort is not even required at a defamation trial if the plaintiff only claims general damages, much less at the preliminary stages of the proceedings (see Brown on Defamation: Canada, United Kingdom, Australia, New Zealand, United States (2nd ed. (loose-leaf)), by R. E. Brown, at pp. 25-16 to 25-25). In addition, adopting a test that requires the submission of this evidence will further complicate and increase the unpredictability of the analysis for what is only one factor in the forum non conveniens analysis.

[ 202 ] Third, there is limited doctrinal support in the Canadian academic literature in favour of adopting the place of most substantial harm rule in defamation cases. To our knowledge, only three articles have suggested this approach (see J.-G. Castel, “Multistate Defamation: Should the Place of Publication Rule be Abandoned for Jurisdiction and Choice of Law Purposes?” (1990), 28 Osgoode Hall L.J. 153; C. Martin, “ Tolofson and Flames in Cyberspace: The Changing Landscape of Multistate Defamation” (1997), 31 U.B.C. L. Rev. 127; M. Castel, “Jurisdiction and Choice of Law Issues in Multistate Defamation on the Internet” (2013), 51 Alta. L. Rev. 153) — and for good cause. In our view, lex loci delicti adequately accommodates the concerns raised in them: it puts the individual’s reputation squarely at the core of the applicable law because harm to reputation occurs at the place of publication, where the tort occurs. Further, we would also note that apart from the dissenting judge in the Court of Appeal and our colleagues Justices Abella and Wagner, no other Canadian judge has showed an interest in adopting the place of most substantial harm rule.

[ 203 ] Last, Haaretz refers this Court to only one jurisdiction that has adopted this rule: Australia. In our view, it would be unwise for this Court to rely on the Australian approach as a basis for overhauling the choice of law rules in this area for three reasons: first, in Australia, the modification of the choice of law rule was made in the context of legislative reform of the law of defamation; second, even in Australia, the place of most substantial harm rule is confined to cases involving multiple domestic jurisdictions, not international defamation cases like this one; and third, Australia adopted a multi-factor test for choice of law in defamation cases, where the place of most substantial harm is merely one of the many factors to consider ( Defamation Act 2005 (N.S.W.), s. 11(3)). The Law Commission of Ontario is currently working on a reform project regarding defamation law in the Internet context ( Defamation Law in the Internet Age: Consultation Paper (2017) (online)). In light of this, it would not be appropriate for this Court to overhaul the choice of law rules in this area.

[ 204 ] For these reasons, we are not satisfied that the law should be changed. The lex loci delicti rule, which would find Ontario law to be the applicable law, governs and should continue to govern. Respectfully, in our view, the prominence given to this factor by the dissenting judge in the Court of Appeal taints her ultimate finding that Israel was a clearly more appropriate forum.

(ii) The Proper Approach to Balancing the Applicable Law Factor

[ 205 ] Applicable law is an important factor in the forum non conveniens analysis. Fairness and efficiency — as well as concerns of cost, convenience, and accuracy — militate in favour of resolving a dispute in a forum familiar with the applicable law (Pitel and Rafferty, at p. 126). In this case, lex loci delicti indicates that the applicable law is that of Ontario. It is therefore fairer and more efficient for this dispute to be heard by the courts in Ontario.

[ 206 ] Our colleague Justice Côté agrees that Ontario law is applicable under the lex loci delicti rule (para. 88). However, she concludes that this factor “cannot aid Haaretz in showing that it would be fairer and more efficient to proceed in the alternative forum” because if the action were to proceed in Israel, Israeli law would apply ( ibid. ). She further suggests that this factor should be granted little weight in the analysis because if jurisdiction is established on the basis of the situs of the tort, the lex loci delicti analysis will inevitably point to the chosen forum, making it of little value in the comparative forum non conveniens analysis (para. 90).

[ 207 ] With respect, we disagree. This Court has considered the applicable law in multijurisdictional defamation cases, and in each case it has identified a single applicable law and weighed this factor accordingly (see, for example, Black , at para. 33 ; Banro , at para. 62). It is entirely appropriate, in our view, for courts to only look at the chosen forum in determining the applicable law. Requiring courts to assess the choice of law rules of a foreign jurisdiction may require extensive evidence, needlessly complicating the pre-trial motion stage of the proceedings.

[ 208 ] Where jurisdiction is based on the situs of the tort, the applicable law (under lex loci delicti ) will indeed point to the forum. However, this does not mean, as Justice Côté suggests, that the applicable law factor should be granted little weight in the forum non conveniens analysis. Rather, giving due weight to this factor reflects the notion that a case should proceed in a forum that properly has jurisdiction over the matter unless another forum is clearly more appropriate. Holding that the applicable law should be given little weight ignores the importance of the territorial jurisdiction of the chosen forum, and distorts the forum non conveniens analysis in favour of the foreign jurisdiction.

[ 209 ] There is a compelling reason why the lex loci delicti rule directs courts to apply their domestic law after having found that the tort of defamation occurred within their jurisdiction. Defamation law is directed to the protection of reputation. For choice of law purposes, it is therefore logical that a court of a jurisdiction where publication occurred — and where harm to reputation consequently occurred — is entitled to apply its own law. This remains true even if a tort took place simultaneously in another jurisdiction.

[ 210 ] As the applicable law is that of Ontario, this factor strongly favours Ontario over Israel.

(b) Fairness to the Parties

[ 211 ] Fairness to the parties, along with the efficient resolution of disputes, is the cornerstone of the doctrine of forum non conveniens ( Van Breda , at para. 104; Black , at para. 36 ). The motion judge and the majority in the Court of Appeal found that this factor favoured Ontario, while the dissenting judge concluded that it favoured Israel. We agree with the motion judge and the majority in the Court of Appeal.

[ 212 ] This Court has repeatedly emphasized the importance of plaintiffs being allowed to sue for defamation in the locality where they enjoy their reputation, recognizing the value of the plaintiff’s subjective conception of his or her reputation ( Banro , at para. 58; Black , at para. 36 ). As the majority of this Court recently stated, “[t]he right to the protection of reputation, which is the basis for an action in defamation, is an individual right that is intrinsically attached to the person” ( Bou Malhab v. Diffusion Métromédia CMR inc. , 2011 SCC 9 , [2011] 1 S.C.R. 214, at para. 46 ). In Banro , this Court approved the decision of the Ontario High Court in Jenner v. Sun Oil Co. , 1951 129 (ON SC) , [1952] 2 D.L.R. 526, wherein the judge “found that the plaintiff would not be able to satisfactorily ‘clear his good name of the imputation made against him’ other than by suing for defamation in the locality where he enjoyed his reputation — that is, where he lived and had his place of business and vocation in life” ( Banro , at para. 58, quoting Jenner , at pp. 538 and 540; see also Paulsson , at paras. 29-30 ).

[ 213 ] In the instant case, Mr. Goldhar has a real and long-standing reputational interest in Ontario. This is where he lives and works, and it is where he has his main business interests. In the context of an article about Mr. Goldhar’s management of an Israel soccer team, Haaretz chose to publish disparaging comments about his Canadian business enterprises and his management of them. The sting of the article relates to his reputation in Ontario .

[ 214 ] Because Mr. Goldhar is concerned about the impact on his Canadian business reputation, it does not matter that only a relatively small section of the article refers to his Canadian business practices; libellous statements may well be buried in lengthy materials. In our view, it is immaterial that Mr. Goldhar also owned a business in Israel, or that the article pertained primarily to his Israeli business. The fact that the other parts of the article relate to another topic has no bearing on the reputational harm at stake. In other words, there is no merit to the quantitative argument. We must look at the allegedly libellous statements themselves in order to identify the reputational harm at stake. Fairness strongly supports allowing Mr. Goldhar to vindicate his reputation in the jurisdiction where he maintains his reputation, and where the sting of the article was felt by him.

[ 215 ] While the plaintiff in this case is wealthy, access to justice concerns are implicated when considering fairness, and must be considered. For many non-wealthy plaintiffs, being denied access to the courts of a particular jurisdiction — typically their home forum — means being denied justice altogether. In those cases, fairness would weigh even more heavily in favour of the plaintiff’s choice of forum.

[ 216 ] While fairness to both parties must be considered, the motion judge expressly considered and rejected Haaretz’s submission that it was unreasonable for it to defend this action in Ontario; instead, he pointed to the fact that it should come as no surprise to Haaretz that Mr. Goldhar would seek to “vindicate his reputation in Ontario, where he lives and works” (paras. 64-65).

[ 217 ] It is true that Haaretz does not have any connection with Ontario. But that does not matter here. As indicated, for reasons best known to itself, Haaretz’s article made gratuitous reference to Mr. Goldhar’s Canadian businesses. Moreover, the newspaper allowed the article to be freely accessed online in Canada. Given Haaretz’s course of action, it is not unfair that Ontario be the forum deciding the dispute.

[ 218 ] Our colleague Justice Côté suggests that in assessing fairness to the parties, we cannot ignore the reality of Mr. Goldhar’s significant business interest and reputation in Israel (para. 78). However, in doing so, she effectively seeks to portray Mr. Goldhar’s claim that the sting of the libel is felt in Ontario as an afterthought, if not a disingenuous maneuver to force Haaretz to come to Canada and bear the added costs associated with this. As indicated, the focal point of Mr. Goldhar’s claim before this Court was tied to Haaretz’s gratuitous reference to the way he runs his Canadian businesses. In our respectful view, therefore, his reputation in Israel is not material to the analysis.

[ 219 ] Finally, it is important to recognize that this case was found not to be an abuse of process by the motion judge. Although she did not expressly dissent on this issue, Pepall J.A. seemed very much influenced by the “competing theory that this lawsuit was instituted in Ontario with a view not to protect a reputation, but to burden a foreign newspaper . . . or . . . to muzzle the newspaper” (para. 191). But this theory was rejected by the motion judge who found that this action was “far from being an abuse of process” (para. 76). The majority of the Court of Appeal saw no error in this conclusion, and Haaretz does not challenge it before this Court. We are therefore faced with a legitimate claim, brought by a long-time resident of Ontario in the jurisdiction where he lives and has his main business interest, in a manner that was reasonably foreseeable when Haaretz decided to publish the impugned statements.

[ 220 ] The fairness factor weighs heavily in favour of Ontario.

(c) Comparative Convenience and Expense for the Parties and Witnesses

[ 221 ] The motion judge found that the comparative convenience and expense for the parties favoured Israel, and that comparative convenience and expense for the witnesses slightly favoured Israel. The majority in the Court of Appeal upheld this conclusion, but the dissenting judge found that this factor overwhelmingly favoured Israel. Again, we agree with the conclusion of the majority.

[ 222 ] With respect to the comparative convenience and expense for the parties, it was relevant that Mr. Goldhar, despite living in Canada, has an apartment in Israel, often travels to Israel, and has strong connections with that jurisdiction. The courts below correctly found that the comparative convenience and expense for the parties favours Israel.

[ 223 ] With respect to the comparative convenience and expense for the witnesses, Mr. Goldhar did not file evidence regarding the witnesses that he would call to testify at trial. For its part, Haaretz listed 22 witnesses that could be called, 18 of whom are in Israel. We agree with the majority in the Court of Appeal that “the motion judge was entitled to treat Haaretz’s proposed witness list with caution” (para. 73). Indeed, it is not clear what these proposed witnesses would speak to, particularly since the allegedly libellous statements complained of concern Mr. Goldhar’s business practices in Canada.

[ 224 ] The cautionary approach of the motion judge was in our view proper for two reasons. First, when considering the factor of comparative convenience and expenses for the witnesses, courts should be mindful that at this preliminary stage of the proceedings, parties may not yet have decided which witnesses will actually be called to testify at trial. The timing of the stay motion “increases the difficulty for the court in identifying the forum conveniens ” (Pitel and Rafferty, at p. 118).

[ 225 ] Second, a defendant cannot change the nature of the plaintiff’s action and answer the claim it would rather have brought against it. In this case, Mr. Goldhar’s claim pertains to the statements made in relation to the management of his Canadian business. Haaretz cannot reshape this action into one concerning the management of the Maccabi Tel Aviv soccer team. Haaretz only provided particulars with respect to what eight witnesses could speak to, despite being aware that the relevance of the testimony of every proposed witness was at issue. All but one of these eight witnesses are former or current Maccabi Tel Aviv staff whose evidence relates to the management of the club in Israel. The motion judge was therefore left with eight Israeli witnesses who could speak of an issue that is merely incidental to the heart of the dispute. Neither Haaretz nor our colleague Justice Côté has identified a single one of these witnesses who has said anything or can say anything about Mr. Goldhar’s business practices in Ontario. Given this, the motion judge was justified in not giving undue weight to Haaretz’s preliminary list of Israeli witnesses, and he certainly did not make a legal error in this respect (Côté J., at para. 50).

[ 226 ] As mentioned earlier, appellate courts must not interfere with a motion judge’s exercise of discretionary power if the judge has not “erred in principle, misapprehended or failed to take account of material evidence, or reached an unreasonable decision” ( Lapointe , at para. 54, quoting Banro , at para. 41). Here, there are no grounds upon which to interfere with the motion judge’s finding regarding the likely relevance of the proposed witnesses. With respect to our colleague Justice Côté, in our view, this Court should show restraint in revisiting this finding.

[ 227 ] In fact, the effect of her approach is to make a long list of foreign witnesses practically determinative in the forum non conveniens analysis. We disagree. One party should not be permitted to manipulate this factor simply by listing numerous witnesses in its jurisdiction of choice — without providing any further indication of relevance. In multijurisdictional cases, it is almost certain that there will be parties and witnesses in different jurisdictions. In such circumstances, it is virtually inevitable that some parties and witnesses will incur travel expenses.

[ 228 ] Fortunately, in our era of mobility and interconnectivity, we are well equipped to face these challenges. There are many procedural tools to mitigate the practical inconvenience arising in cases where parties are in multiple jurisdictions: written affidavits, testimony through videoconference, examinations before trial, rogatory commissions, etc. Modern communication technologies and methods of transportation have rendered these kinds of arrangements much more practicable than in the past. In this regard, the majority in the Court of Appeal was right to stress that “the use of technology and interpreters cannot be viewed as undermining the fairness of a civil trial” (para. 71). This is especially true when the witnesses who testify via videoconference and/or interpreters are only called to speak to a secondary issue.

[ 229 ] In the face of inconclusive evidence as to the state of Israeli law, Justice Côté proceeds on the basis that the Israeli witnesses could not be effectively compelled to testify if the trial were to proceed in Ontario (paras. 59, 65 and 79). She recognizes that, at the forum non conveniens stage, the burden is on the defendant (para. 46). However, despite the fact that Haaretz provided no evidence regarding the impossibility of compelling Israeli witnesses, she finds that “Haaretz met its burden in establishing a concern as to the fairness of a trial in Ontario” (para. 63). On the basis of this reasoning, she states that the burden of establishing the fairness of the trial in Ontario rested on Mr. Goldhar. With respect, we cannot agree with this change in the law. If we follow this logic, courts in a particular jurisdiction would always need to presume that their own rules of civil procedure dealing with witnesses outside of the jurisdiction would be ineffective. The burden of establishing the fairness of the trial would therefore shift to the plaintiff in all cases involving witnesses outside of the jurisdiction.

[ 230 ] Before this Court’s decision in Van Breda , the onus of proof in motions to stay based on the doctrine of forum non conveniens varied from province to province and depended on how the defendant had been served. In Van Breda , this Court indicated that the burden is always on the defendant (para. 103; Pitel and Rafferty, at pp. 121-22). This rule is consistent with the underlying idea that jurisdiction should be exercised once it is properly assumed. In addition, our colleague Justice Côté’s approach — according to which the burden shifts where there is a “ prima facie concern” — would unnecessarily complicate the handling of forum non conveniens motions.

[ 231 ] For these reasons, the motion judge and the majority of the Court of Appeal were correct to find that this factor, at best, slightly favours Israel.

(d) Loss of Legitimate Juridical Advantage

[ 232 ] The motion judge found that the juridical advantage factor favoured Ontario, as Mr. Goldhar would have access to a jury trial in Ontario. Both the majority and the dissent in the Court of Appeal thought that it was a neutral factor because prior to the motion, Mr. Goldhar had not delivered a jury notice.

[ 233 ] However, as our colleague Justice Côté notes, any party in a proceeding in Ontario may deliver a jury notice before the close of pleadings (para. 74). After the Court of Appeal dismissed Haaretz’s appeal, Mr. Goldhar promptly delivered a jury notice. There was therefore a juridical advantage still available to Mr. Goldhar in Ontario. This factor favours Ontario.

(e) Multiplicity of Proceedings and Conflicting Decisions

[ 234 ] This factor was not considered by the motion judge or the majority in the Court of Appeal. For her part, the dissenting judge pointed out that Mr. Goldhar’s undertaking to limit his claim to his Canadian reputation did not prevent him from bringing an action in another jurisdiction, and that there was a risk of multiplicity of proceedings and conflicting decisions. We disagree. When properly assessed, that risk does not exist in this case.

[ 235 ] Mr. Goldhar’s undertaking limiting his claim to his Canadian reputation ensures that there will be no conflicting decisions. There is also no risk of multiple proceedings. Before this Court, Mr. Goldhar took the position that it would be an abuse of process for him to sue in another jurisdiction. Therefore, no weight should be granted to this factor.

(f) Enforcement of Judgment

[ 236 ] Finally, the “enforcement of judgment” factor was not considered by the motion judge or the majority in the Court of Appeal. The dissenting judge found that this factor favoured Israel. We disagree. In defamation cases, vindication of the plaintiff’s reputation is often a primary concern, if not the primary concern. This stance often renders the enforcement of the final judgment irrelevant to the forum non conveniens analysis in defamation cases. As the Ontario Superior Court of Justice stated in Barrick Gold Corp. :

It is recognized in defamation cases that the vindication of one’s reputation is as important as any monetary award of damages that might be obtained. For its purposes, Barrick may be quite content with a declaration by a court in Ontario that the statements made by the defendants are untrue even if it cannot recover any damages that might be awarded to it as a consequence. [para. 40]

[ 237 ] Contrary to the dissenting judge’s opinion, this factor does not weigh heavily in the analysis in this case.

(3) Conclusion on Forum Non Conveniens

[ 238 ] In sum, the key factors of applicable law and fairness to the parties weigh heavily in favour of Ontario, while the factor of loss of legitimate juridical advantage also weighs in favour of Ontario. Only the factor of comparative convenience and expense for the parties and witnesses favours Israel, and this only slightly so with respect to the witnesses. The enforcement of judgment factor does not weigh heavily in the analysis. As this Court stated in Black , “[t]he forum non conveniens analysis does not require that all the factors point to a single forum or involve a simple numerical tallying up of the relevant factors. However, it does require that one forum ultimately emerge as clearly more appropriate” (para. 37 (emphasis in original)).

[ 239 ] Admittedly, it would not be unreasonable to hold a trial to settle this dispute in Israel. But that is not the point. Ontario courts have jurisdiction. Following the forum non conveniens analysis, Israel has not emerged as a forum that would be more appropriate than Ontario to hear the case, much less a clearly more appropriate forum. This was the high threshold that Haaretz was required to meet in order to displace the forum chosen by the plaintiff and to convince the Ontario courts — whose jurisdiction has been properly assumed — not to exercise their jurisdiction over this matter. Haaretz has not displaced the normal state of affairs, which is that jurisdiction should be exercised once it is properly assumed. Above all else, fairness concerns militate in favour of Mr. Goldhar being able to vindicate his reputation in the place where his Canadian business practices were impugned and the sting of the article was felt by him.

[ 240 ] This Court should not lower, through a relaxed application of the “clearly more appropriate” test, the high threshold that it has consistently upheld since Amchem . For these reasons, we would dismiss the appeal.

APPENDIX

Soccer / Profile / Long-distance operator

Though he spends most of his time in Canada, Maccabi Tel Aviv owner Mitch Goldhar runs his club down to every detail. But could his penny pinching and lack of long term planning doom the team.

by David Marouani

Crises are par for the course at Maccabi Tel Aviv, even when the club appears to be on an even keel. Most of the crises don’t make it onto the public’s radar, but they have one thing in common: their connection to [ sic ] way that Canadian owner Mitch Goldhar runs the club.

Just over a year ago, Goldhar’s representative in Israel, Jack Angelides, complained about the job that Clarice Zadikov, the long-time CFO of the team, was doing. Goldhar’s immediate response was to suggest appointing someone to do an identical job, with a slightly different title — but reporting back to the owner. So Tomer Shmuel was appointed commercial manager and Zadikov’s authority was slowly eroded. Two months ago, the policy had the desired effect and Zadikov reached an agreement with Angelides over her retirement.

“Mitch’s game plan is to wear down anybody who he wants to get rid of, until they’ve had enough and decide to leave of their own accord,” one club insider told Haaretz this week.

The departure of CEO Uzi Shaya, following the gradual erosion of his powers, is a case in point. “The dismissal of Avi Nimni is the exception that proves the rule,” the same insider said. “For the most part, [Goldhar is] supremely patient. One could even say he’s cold and calculated.”

Goldhar is also playing with time in the battle between coach Moti lvanir and star striker Barak Yitzhaki. Goldhar landed in Israel on Friday, but he opted not to address the spat until Monday evening.

According to club sources, the owner is currently observing the situation and has not yet decided how he will handle this latest crisis. “Whatever happens,” one source said, “he will be remembered as the knight in shining armor who came in and saved the day.”

Goldhar’s management model was imported directly from his main business interest — a partnership with Wal-Mart to operate shopping centers in Canada. He even spelled out his managerial vision in a leaflet distributed to fans ahead of Sunday night’s derby against Hapoel Tel Aviv.

“By dealing with disciplinary matters, commitment and the right approach,” he wrote, “we are now at the dawn of a cultural revolution — a process of building a new sporting culture.”

Within the club, however, there are those who believe that Goldhar’s managerial culture is based on overconcentration bordering on megalomania, penny-pinching and a lack of long-term planning.

“With all due respect to ‘cultural revolutions’, the gap between Maccabi Tel Aviv and Maccabi Haifa is getting wider since he arrived,” said one team insider.

And with all due respect to Angelides, everyone at Maccabi knows that it’s a one-man show. Anything that Goldhar’s Cypriot lieutenant says to the players or to the coaching staff is prefixed by the words “Mitch says. . . ”

When Ivanir read the riot act to his players at a meeting in Caesarea last week, almost every sentence included the phrase, “the owner told me that. . . ”

Despite running the club from afar, decisions are only made once Goldhar has given them the green light. He was even involved in the minute details of the search for a location for the club’s new souvenir shop.

“I want to invest in branding the store,” he told his employees over a year ago. For months, he was presented with dozens of potential locations for the store in north Tel Aviv, but rejected them all. In the end, he decided to renovate the mobile home in the south of the city where the store is currently located.

Do as your boss says

Goldhar boasts to his business contacts in Toronto that he is not only the owner of Maccabi Tel Aviv but also its soccer director. The last time he was in Israel, he brought Ivanir into his office and tried to tell him how the team should be playing. “[Haris] Medunjanin should be playing in the same position that he plays for the [Bosnian] national team,” Goldhar reportedly told his coach. In fact, it was at Goldhar’s suggestion that Medunjanin was returned to the starting line-up at the expense of Gal Alberman. “Ivanir doesn’t know how to respond in these situations,” says a club source. “But he believes that he really should do as his boss suggested — even if that boss knows nothing about soccer.”

This week, too, in the aftermath of the defeat in Sunday’s derby match, Goldhar got involved.

“You showed that you’ve got the ability,” he told the players, “but you seem to have misplaced the character that you showed at the start of the season. I am convinced that you still have that character and now’s the time that you have to show it.”

Goldhar has invested hundreds of millions of shekels in Maccabi since he arrived on scene some two and a half years ago, but club sources say that he borders on the frugal when it comes to the managerial side of the club. When Angelides was first offered a job, for example, Goldhar did not see fit to offer him a company car. Angelides complained bitterly but silently about this, until he eventually persuaded one of the team’s sponsors to provide him with a vehicle — without Goldhar’s knowledge.

In an interview with Yedioth Ahronoth’s Nahum Barnea, Goldhar spoke about how much he values the work done behind the scenes by the club’s equipment manager, David Zachi, who earns a fraction of the salary of the players. What he failed to point out, however, is that he has steadfastly refused to raise Zachi’s measly pay by just a few hundred shekels. To Goldhar’s credit, it should be noted that, when it comes to frugality, he practices what he preaches: he rented a dingy apartment for himself in Tel Aviv and he drives nothing more fancy than a Hyundai Getz.

Goldhar, according to club insiders, thrives on the media attention that Maccabi brings him. Despite the fact that he planned his latest visit to Israel well in advance, for example, and the crew aboard his private jet was briefed a week in advance, he made sure that the media were kept in the dark, in order to create an aura of expectation.

When Maccabi played against Panathinaikos earlier this season, he read everything that was written about him [ sic ] the Greek press and even cut out a cartoon of him that appeared in one [ sic ] the paper, asking all his employees whether it was flattering. He also has articles in which his name appears translated into English.

Despite his many statements, Goldhar does not have a long-term plan for the team. The only plan he has presented so far has been to upgrade the club’s training facility, but that still hasn’t happened. The only changes he has made have been to the youth team set-up, and he often boasts about that team’s accomplishments.

This has become a sore point with former owner Alex Shnaider, who complained that Goldhar was taking credit for a five-year plan that was implemented before he even arrived at the club.

As for his long-term future, Goldhar says that he’s here to stay. “He is so keen to prove to everybody that his business model can work that he won’t leave until he’s won at least a league championship,” according to one of his close associates.

There are those, however, who see things differently. Goldhar plays soccer at least once a week in Toronto with Ilan Sa’adi, a former professional player and close friend. One of the people who plays with them says that, between the lines, there are clear signs that Goldhar is getting frustrated with Maccabi.

“He’s very distressed at the way the team is playing,” the source says. “If I understand him correctly, he will give the team until the end of this season to win the championship and then he’ll start looking for someone to take Maccabi off his hands.”

Goldhar declined to comment for this article.

Appeal allowed with costs throughout, McLachlin C.J. and Moldaver and Gascon JJ. dissenting .

Solicitors for the appellants: Blake, Cassels & Graydon, Toronto.

Solicitors for the respondent: Lenczner Slaght Royce Smith Griffin, Toronto; Julian Porter, Q.C., Toronto.

Solicitor for the intervener: University of Ottawa, Ottawa.

[1] See also J.-G. Castel (1990), at p. 173.

[2] Opinion of Advocate General Cruz Villalón , C-509/09, C-161/10, [2011] E.C.R. I-10272.

[3] eDate Advertising GmbH v. X , C-509/09, C-161/10, [2011] E.C.R. I-10302.

[4] The factors are: “. . . (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” ( Van Breda , at para. 90).

[5] J.-G. Castel (1990), at p. 164, citing A. M. Linden, Canadian Tort Law (4th ed. 1988), at pp. 627-29.

[6] The Haaretz article which is the subject of the defamation claim is set out in full in the Appendix.

7 Note that, procedurally, this should have been a motion to dismiss (see S. G. A. Pitel and N. S. Rafferty, Conflict of Laws (2nd ed. 2016), at pp. 119-21).