Court File and Parties

COURT FILE NO.: 07-CV-334113 PD2 DATE: 20200810 ONTARIO SUPERIOR COURT OF JUSTICE

BETWEEN:

DARA FRESCO Plaintiff

- and –

CANADIAN IMPERIAL BANK OF COMMERCE Defendant

Proceeding under the Class Proceedings Act, 1992

BEFORE: Justice Edward P. Belobaba

COUNSEL: David O’Connor, Steven Barrett, Louis Sokolov, Adam Dewar, Peter Englemann and Louis Century for the Plaintiff Linda Plumpton, John C. Field, Sarah Whitmore, Ryan Lax, Lara Guest, Lauri A. Reesor and Elisha C. Jamieson-Davies for the Defendant

HEARD: June 29, 2020

Cross-motions for Summary Judgment – Damages Issues

[1] These motions for summary judgment in a class action for unpaid overtime are being heard in two stages. First, liability and then damages. The liability stage considered Common Issues 1 to 5 and ended with a finding in favour of the plaintiff. [1]

[2] In the Liability Decision released on March 30, 2020 I concluded as follows:

The defendant bank breached the overtime obligations prescribed by federal labour law. Rather than implementing a system to pay for all hours required or permitted (and to track actual hours worked so that this could be achieved), the bank failed to record actual hours worked and made overtime compensation contingent primarily on pre-approval. I have found on the record before me that some class members worked overtime hours that were not recorded and were not compensated in accordance with the requirements of the Code. I have also found, for the reasons set out above, that the defendant bank must be found to have permitted (or not prevented) all uncompensated overtime hours of the class members.

The bank’s unlawful overtime policies and hours-of-work recording practices were systemic or institutional impediments. That is, they were system-wide in nature and they impeded class member overtime claims that were otherwise compensable under the Code.

The plaintiff has established liability.

The appropriate remedies, including damages, will be addressed in the second part of this decision when I consider common issues 6 to 8 and the related question of aggregate damages … [2]

[3] I have now heard argument on common issues 6 to 8 and the related question of aggregated damages. I will set out my answers and analysis for each of these items in turn.

Common issue 6: Unjust Enrichment

Was the Defendant enriched by failing to pay Class Members appropriately for all their hours worked? If "yes", (a) Did the class suffer a corresponding deprivation? and (b) Was there no juristic reason for the enrichment?

Answer: Yes, to some extent.

[4] I say “to some extent” because of my finding on “an abundance of evidence” that “ some of the class members worked uncompensated overtime.” [3] (Emphasis added). Thus, I have no difficulty concluding that the defendant bank was enriched by failing to pay some class members appropriately for all their hours worked.

[5] As this court noted in Baroch v. Canada Cartage, 2015 ONSC 40: [4]

If it is determined that [the employer] had a policy or practice of avoiding or disregarding its overtime obligations to class members, then such a policy or practice … would likely result in some level of enrichment on the part of the defendant and a corresponding deprivation on the part of some or all of the class members.” [5]

[6] This applies here as well.

[7] I hasten to add, however, that a finding of unjust enrichment in the context of a case that is fundamentally about damages for breach of an employment contract adds little in the way of remedies. I expand on this point below.

Common issue 7: Remedies

If the answer to any of common issues 1, 2, 3, 5 or 6 is "yes", what remedies are Class Members entitled to?

Answer: The declarations as pleaded in the Statement of Claim’s paragraphs 1(e), (k), and (m), and damages.

[8] The class members are entitled to the declarations and orders as pleaded in paragraph 1(e) of the Statement of Claim [about the illegality of the 2006 Overtime Policy]; paragraph 1(k) [about the breaches of the employment contracts]; and paragraph 1(m) [about the breaches of statutory obligations under Part III, Division I of the Labour Code.] The class members are also entitled to damages arising out of the breaches of their employment contracts.

[9] The class members are not entitled to restitutionary relief for unjust enrichment because compensation in damages is already available under the breach of contract claim. The case law is clear that:

Contract trumps unjust enrichment in several respects. Most obviously, restitutionary relief is not available if the claimant possesses a right to contractual relief. [6]

[10] Thus, the available remedies are the declarations as noted and damages.

Common issue 8: Punitive damages

If the answer to any of common issues 1, 2, 3, 5 or 6 is "yes", is the Class entitled to an award of aggravated, exemplary or punitive damages based upon the Defendant's conduct? If "yes" (i) Can these damages be determined on an aggregate basis? and (ii) What is the appropriate method or procedure for distributing any aggregate aggravated, exemplary or punitive damages to Class Members?

Answer: No.

[11] Punitive damages may be awarded in a breach of contract case if the plaintiff can show three things: (i) the defendant’s conduct was “malicious, oppressive and high-handed” and “a marked departure from ordinary standards of decent behaviour;” [7] (ii) when added to any compensatory award, a punitive damages award is rationally required to punish the defendant and to meet the objectives of retribution, deterrence and denunciation; [8] and (iii) the defendant committed an actionable wrong independent of the underlying claim for damages for breach of contract. [9]

[12] In the Liability Decision, I rejected the allegation that the defendant bank breached its obligation of good faith contractual performance. I concluded that at most the bank’s actions were careless and indifferent, even negligent, but did not amount to bad faith:

I can find on the evidence that CIBC was careless and indifferent, indeed negligent, about its obligation to comply with the requirements of the Code. I can also find that the bank should have known better. It is a multi-billion-dollar financial institution with an able legal staff that can easily advise on the requirements of federal labour law. For some reason this didn’t happen. The bank dropped the ball, to be sure. But I cannot find on the evidence before me that the defendant bank lied or knowingly misled its employees about the legality of its overtime policies. The “breach of good faith” allegation does not succeed. [10]

[13] Having made these findings in the Liability Decision, I would be hard pressed to now decide that the defendant bank’s overall conduct was in fact “malicious, oppressive and high-handed”. The plaintiff, however, argues otherwise and points to evidence that did not have to be considered in deciding the breach of good faith (in contractual performance) obligation.

[14] The plaintiff directs the court’s attention to three new points: that the defendant bank knowingly failed to produce any of the survey documents until ten years after the litigation began; that it filed sworn evidence at certification that contradicted the undisclosed survey evidence and falsely asserted that internal “reviews and inquiries” revealed no systemic overtime problems; and that it recklessly destroyed time-stamped data that was subject to a litigation hold and had relevance for the damages determination.

[15] The defendant bank responds to each of these “bad conduct” allegations with a detailed and credible rebuttal. There is no need for me to go through the competing points and counter-points. Suffice it to say that I am unable to conclude on the mix of evidence before me that the defendant bank’s conduct on the new points or even overall was so egregious that it justifies a punitive damages award. I cannot fairly and reasonably conclude that any of the impugned conduct was “malicious, oppressive and high-handed” and “a marked departure from ordinary standards of decent behaviour”. [11]

[16] I can, however, suggest that the impugned conduct, all of which occurred during the course of the litigation, may arguably merit some level of sanction by way of an elevated costs award if the plaintiff prevails in this action. I leave this for the cost submissions.

Adding common issue 9: Aggregate damages

[17] This last question – whether aggregate damages should be added as a common issue – generated the most debate. This was understandable. If aggregate damages are not allowed and class members are required to individually advance and prove claims (stretching over many years), the bank’s financial exposure, in practical terms, will probably be modest at best. If aggregate damages are permitted, the monetary liability of the defendant bank could well be in the tens of millions of dollars. Hence, the importance of this issue.

[18] Aggregate damages under s. 24(1)(c) of the Class Proceedings Act, 1992, S.O. 1992, c. 6 (“CPA”) are generally awarded if all or part of the claimed loss can be determined without proof by individual class members. In most “aggregate damage” cases, the court is able to determine all or part of the class members’ loss based on the defendant’s own records and individual hearings are not required.

[19] The wrinkle here is that the defendant bank, in violation of federal law, failed to make and keep records of the “hours worked” [13] – the very records that the plaintiff would have used to determine damages on an aggregate basis. Having failed to keep the required records, the defendant bank now says that every aggrieved class member should be required to prove their case, and if it takes some 30,000 or more individual trials, so be it. The plaintiff says individual proof is not only unrealistic (few if any employees ever made and kept such records) but profoundly unfair and could well result in a significant wrong eluding an effective remedy. The plaintiff proposes an alternative methodology for assessing aggregate damages that she says, in the circumstances, is reasonable and reliable. Not surprisingly, the defendant submits it is neither.

[20] The other wrinkle is that when the Court of Appeal certified this action as a class proceeding eight years ago, it refused to certify the aggregate damages question. [14] Winkler C.J.O., writing for the Court, concluded that the “sampling” methodology proposed by the plaintiff’s expert in both Fulawka v. Bank of Nova Scotia, 2012 ONCA 443, a parallel bank overtime case, and Fresco v. Canadian Imperial Bank of Commerce, 2012 ONCA 444 could not be used to determine aggregate damages under s. 24(1)(c). [15]

[21] In a decision released in 2013, I suggested that Winkler C.J.O.’s analysis on this narrow but important point about sampling was probably wrong. [16] The Court’s interpretation of ss. 23(1) and 24(1)(c) of the CPA in Fulawka (and Fresco) was contrary to the plain meaning of these provisions and to the decisions of other panels of the Court of Appeal in Markson v. MBNA Canada Bank, 2007 ONCA 334, [17] Cassano v. Toronto Dominion Bank, 2007 ONCA 781 [18] and Cloud v. Canada (Attorney General). [19] Indeed, in Markson, the Court of Appeal expressly confirmed that:

Statistical sampling - as provided for in s. 23 - can be employed to determine the aggregate or part of the defendant's liability without proof of individual claims.” [20]

[22] The Court of Appeal’s (sensible) interpretation of ss. 23 and 24 in Markson was repeated by the Divisional Court in 2038724 Ontario Ltd. v. Quizno’s Canada Restaurant Corporation [21] and in numerous decisions of this court. [22] The error in Fulawka (and here in Fresco), in my respectful opinion, is an unfortunate outlier that should be corrected by the Court of Appeal at the first opportunity.

[23] However, there is no need to belabour this point because the sampling methodology that was rejected by Winkler C.J.O. is markedly different from the methodology that is now being proposed by the plaintiff and that is not based on sampling. I will discuss the proposed methodology shortly. But it is enough to note that the Court of Appeal’s reason for refusing to certify the aggregate damages question herein eight years ago can easily be distinguished. The defendant’s res judicata submission has no application and does not succeed.

[24] I also note, in any event, that this court has ample jurisdiction to add an aggregate damages question even where this very question was rejected at certification. As the Supreme Court of Canada noted in Pro-Sys Consultants Ltd. v. Microsoft Corporation, 2013 SCC 57: “[t]he failure to propose or certify aggregate damages ... as a common issue does not preclude a trial judge from invoking the provision if considered appropriate once liability is found.” [23]

[25] This court can add an aggregate damages question at this stage if the three well-known requirements set out in s. 24(1)(a) to (c) of the CPA are satisfied. The first two requirements are not in dispute: (i) monetary relief is being claimed on behalf of some of the class members, and (ii) the defendant’s liability has been established. The dispute between the parties involves the third requirement as set out in s. 24(1)(c): whether some or all of the defendant’s monetary liability can reasonably be determined in the aggregate without proof by individual class members.

[26] Typically, aggregate damages are determined in two steps.

[27] The first step, generally litigated at certification when the availability of aggregate damages is being proposed as a common issue, is whether there is a reasonable likelihood that the methodology suggested by the plaintiff’s expert can determine damages in the aggregate without proof by individual class members. [24] When Justice Cullity first coined the phrase “reasonable likelihood” in Vezina v. Loblaw Companies Ltd., [2005] O.J. No. 1974 (S.C.J.), [25] some 15 years ago, he intended this to be a relatively low hurdle - more akin to “possibility” than to actual likelihood or probability. [26]

[28] Thus, the first-step question for me is whether there is a reasonable possibility that the methodology proposed by the plaintiff’s expert can determine damages in the aggregate without proof by individual class members. If so, the aggregate damages question should be certified. Questions relating to the reliability of the plaintiff’s proposed methodology and overall fairness to the defendant are relevant at this first step but only to a limited extent. The primary focus is on the “reasonable possibility” hurdle.

[29] The second step, addressed at the trial or summary adjudication of the certified common issues, requires the court to answer the certified question in the context of all of the evidence – that is, the court considers the proposed methodology as applied to the defendant-based data. The court must determine whether aggregate damages should actually be awarded and, if so, in what amount. This is a merit-based analysis that asks whether, on balance, aggregate damages can be fairly and reasonably determined without proof by individual members. At this second step, questions about reliability and overall fairness to the defendant are paramount.

[30] In Ramdath v. George Brown College, 2014 ONSC 3066, [27] a class action where the common issues were decided at a trial, I discussed the importance of aggregate damages in class action litigation, the practical dimensions of the reasonableness standard and the need to ensure that both sides are treated fairly by the aggregate damages approach. I restated the overarching proposition as follows:

The court may award aggregate damages under s. 24(1)(c) of the CPA if the evidence put forward by class counsel is sufficiently reliable to permit a just determination of all or part of the defendant’s monetary liability without proof by individual class members. [28] (Emphasis added.)

[31] This court’s analysis of aggregate damages in Ramdath was affirmed on appeal. [29] Feldman J.A., writing for the Court of Appeal, said this:

I endorse the trial judge’s view that it is desirable to award aggregate damages where the criteria under s. 24(1) are met in order to make the class action an effective instrument to provide access to justice. I also agree with his focus, at para. 47, on the legislature’s choice of a ‘reasonableness’ standard for determining aggregate liability and with the three criteria he sets out to ensure that both sides are treated fairly by the assessment:

(1) whether the non-individualized evidence presented by the plaintiff is sufficiently reliable;

(2) whether use of the evidence will result in unfairness or injustice to the defendant, such as overstatement of its liability; and

(3) whether the denial of an aggregate approach will result in a “wrong eluding an effective remedy” and denial of access to justice. [30]

[32] It is important to remember that just because the aggregate damages question is certified as a common issue does not mean that the same question will be answered in favour of the plaintiff when it is considered on the merits. It may well be that when the plaintiff’s proposed methodology is tested on the basis of a more complete evidentiary record against the criteria of reasonableness, reliability and fairness to the defendant that an aggregate damages award will not be made.

[33] Here, I have two distinct obligations. First, because the aggregate damages question has not yet been certified as a common issue, I must decide whether the plaintiff has established a reasonable possibility that the methodology proposed by its expert can determine damages in the aggregate without proof by individual class members. Second, I will have to decide whether the plaintiff’s methodology as applied to the evidence presented actually results in a fair and sufficiently reliable determination of the defendant’s monetary liability without proof by individual class members.

[34] Returning then to the first step of the analysis. Here, as already noted, the “hours worked” records do not exist in the hands of the defendant bank because, in violation of federal law, the bank never made or kept such records. The plaintiff’s expert, Stefan Boedeker, proposes what the plaintiff submits is a reasonably equivalent methodology that can fairly and reliably determine the hours worked and, from this, the unpaid overtime.

[35] The proposed methodology is based on a review of the defendant bank’s electronic records (currently housed in nine internal computer systems) that contain time-stamped data showing, among other things, the daily start and stop times of the employee’s computer. The plaintiff says that these time-stamps can be used as a means of reasonably gauging the time worked by all or even some of the class members. In addition, the defendant bank has relevant paper records, namely the blotters maintained by the tellers. These blotters have been used by the bank in the past to resolve employee overtime claims.

[36] If the time-stamped data reveals gaps in the evidence, where complete data cannot be obtained, then statistical sampling or extrapolation (back-casting and forecasting) would be used to fill in the gaps.

[37] Mr. Boedeker, a statistician and economist, has used this particular “time-stamped” methodology in more than 100 “off the clock” overtime class actions in the U.S. and in at least one class action in Canada. [31] He is reasonably confident that a fair and reasonable aggregate damages determination can be achieved in this case as well.

[38] In essence, the proposed methodology would reconstruct timesheets for class members not by using random sampling but by reviewing and using all the relevant time-stamped data that is available in the bank’s computer systems. As Mr. Boedeker explained:

Faced with a lack of accurate timesheets or time records showing actual hours worked, we are left to reconstruct hours worked as best we can, based on available data. Time-stamped records are used as a proxy for actual hours worked, which CIBC failed to track. My methodology offers a reasonable, statistically reliable way estimating uncompensated time worked, which I have applied in numerous cases.

[39] The fact that the proposed methodology has been used in scores of other unpaid overtime cases – I assume with some degree of success – is enough to clear the “reasonable possibility” hurdle. I therefore have no difficulty concluding, as a first step, that it is reasonably possible that Mr. Boedeker’s proposed methodology, based mainly on the defendant bank’s time-stamped computer data, can fairly determine all or part of the bank’s monetary liability without proof by individual members.

[40] The defendant bank’s vigorous submissions to the contrary, mainly about the methodology’s unreliability, do not succeed at this first step – at least not to a point that can show that the relatively modest “reasonable likelihood” or “reasonable possibility” hurdle has demonstrably not been cleared.

[41] The defendant bank’s submissions about the unreliability of the “time-stamped” methodology may be more effective at the second step when the aggregate damages question will be considered on the merits and in full. The bank will then be able to revisit its submission that computer log in and log out times are not necessarily the same as actual hours worked - that some employees may attend to personal matters even after logging on to their computer.

[42] Of course, the same point can be made about any system that records the hours worked by any employee. In my view, one must begin with a neutral assumption that any record of “hours worked” are just that – unless there is credible evidence to the contrary. If the defendant bank intends to argue the unreliability of time-stamps as a reasonable measure of hours worked, it should provide actual evidence to support this submission, not speculation. For example, the bank may be able to provide evidence that as a matter of accepted routine or bank culture, many employees would log in and then take personal time to have breakfast, speak by phone to family or friends or play on-line computer games, before actually starting to work. I frankly doubt that any such evidence will be forthcoming but it’s possible. In any event, no such evidence was presented to the court at this stage.

[43] The defendant bank also tried to impugn the reliability of the time-stamp approach by submitting that Mr. Boekeker’s methodology could not account for the bank’s “flex-time” system which allowed employees to take employees extra breaks, arrive late or leave the office to attend to personal business. However, as Mr. Boedeker explained in his reply report, the proposed methodology may well be able to take flex-time into account:

My methodology may account for a pattern or practice of "permitted flexibility in the workday" … A standard practice of flexible scheduling will leave traceable patterns in the data that can be detected and quantified using statistical techniques such as linear and non-linear regression modeling. Once these patterns have been identified and quantified, they can then be taken into account in the estimation of unpaid hours. …Flexibility in scheduling practices can be taken into account by building assumptions into my methodology and adjusting algorithms accordingly. That is reflected when they arrive late and leave early much like an accurate timesheet would record the actual hours of work.

[44] To repeat, the defendant bank will have ample opportunity to challenge the reliability of the “time-stamped data” approach, and if there are evidentiary gaps, to contest the statistical integrity of the suggested “extrapolation” techniques or the legality of random sampling. These arguments can be made at the so-called “second step” – when the aggregate damages question is answered on the merits, the proposed methodology is actually applied to the gathered evidence and overall reliability and fairness is fully considered.

[45] At this point, however, I am satisfied that the “reasonable possibility” hurdle is cleared and that an aggregate damages question should be added as Common Issue 9:

9. Can the defendant’s monetary liability be determined on an aggregate basis? If so, in what amount?

[46] This new common issue, however, cannot be answered today. Whether aggregate damages can reasonably and fairly be determined, and if so in what amount, cannot be decided until the plaintiff’s expert reviews the bank’s time-stamped data and completes his proposed damages report.

[47] The final wrinkle in this case is that the plaintiff’s expert has not yet been given access to the time-stamped data. Mr. Boedeker has reviewed this data as it applies to five employees over a one-month period - as set out in the Deloitte Report – and this limited review helped to inform the formulation of his proposed methodology. But the class consists of more than 30,000 employees over a multi-year time period. The bulk of the electronic data remains to be extracted from the bank’s computer systems.

[48] The plaintiff says that over the last six years she has requested access to this data repeatedly and was met with continuing refusals. The defendant bank disputes this narrative. Both sides have provided detailed submissions to support their respective positions about who “refused” or “did not request.” I am inclined to find, based on the documentation, that the plaintiff’s narrative is the more credible. However, no such finding need be made because at root two things are clear: (i) the great bulk of the time-stamped data has not even been accessed by the bank, let alone provided to the plaintiff’s expert; and (ii) the “fair and expeditious” [32] determination of the aggregate damages question requires that this time-stamped data be provided to the plaintiff’s expert for his review and report.

[49] Although I have jurisdiction under s. 12 of the CPA to make this production order, I refrain at this point from doing so because this very production obligation was accepted by counsel for the defendant bank in her letter to the court on April 15, 2020 - about two months before the hearing. The letter assumed that the time-stamped data was indeed relevant but explained that it could not be readily produced for two reasons: (i) because of the Covid-19 pandemic, it was physically difficult to access the relevant computer systems and extract the data; and (ii) even after the data was accessed, “a substantial period” of time would still be needed to convert the data into usable form. The letter also made clear that the bank was of the view that the court should at this stage simply focus on (step one) whether aggregate damages should be certified as a common issue:

[T]he disruption caused by the COVID-19 pandemic has had a significant impact on the ability of CIBC to address any steps in the proceeding requiring access to class members’ time-stamped computer data. Among other things, key employees are working from home and are dedicated entirely to urgent business continuity issues. Access to physical premises and resources, including the actual data, is also severely restricted for the immediate future. CIBC is determining whether and to what extent relevant data can be accessed from third-party storage providers during the current workplace shutdown but that issue also remains unresolved

In any event, even after the data can be accessed, we anticipate that it will take a substantial period to put it into a usable form. As noted in the evidence filed on the summary judgment motion, much of the data is stored on media that will require considerable time and resources to convert to usable form.

In the interim, the Court could proceed to consider … whether the issue of aggregate damages should be certified …

[50] It follows from this that there is nothing unfair or unreasonable in proceeding exactly as counsel for the defendant bank has suggested: decide if aggregate damages should be added as a common issue (which I have now done) and then allow the defendant bank the time needed to extract the time-stamped data and convert it into useable form so that the plaintiff’s expert can complete his aggregate damages report.

[51] A final comment about sampling. We won’t know until the plaintiff’s proposed damages report is completed and submitted whether there are any evidentiary gaps and whether statistical sampling will actually be used to fill in these gaps. If such turns out to be the case, it will only be at this point that the suggested judicial error in Fulawka (and Fresco), that sampling can’t be used to help determine aggregate damages, will have to be addressed.

[52] Having certified the aggregate damages question as a common issue, I adjourn the balance of the damages hearing sine die to await the plaintiff’s aggregate damages report and the defendant bank’s response.

Disposition

[53] Common Issues 6, 7 and 8 are answered as set out above.

[54] The question of aggregate damages is added as a new Common Issue 9 as set out above in paragraph 45.

[55] The availability of aggregate damages and the final quantum, if any, will be determined at a further hearing after the plaintiff’s expert submits his proposed damages report and the defendant bank submits its response. If any specific court orders or directions are needed to expedite the data access and review process, counsel should advise.

[56] Costs are deferred to the completion of this matter in its entirety.

Signed: Justice Edward P. Belobaba

Notwithstanding Rule 59.05, this Judgment [Order] is effective from the date it is made, and is enforceable without any need for entry and filing. In accordance with Rules 77.07(6) and 1.04, no formal Judgment [Order] need be entered and filed unless an appeal or a motion for leave to appeal is brought to an appellate court. Any party to this Judgment [Order] may nonetheless submit a formal Judgment [Order] for original signing, entry and filing when the Court returns to regular operations.

Date: August 10, 2020

Footnotes

[1] Fresco v. Canadian Imperial Bank of Commerce, 2020 ONSC 75. [2] Ibid., at paras. 91-94. [3] Ibid., at paras. 63 and 77. [4] Baroch v. Canada Cartage, 2015 ONSC 40. [5] Ibid., at paras. 49-50. [6] McInnes, The Canadian Law of Unjust Enrichment and Restitution. (2014) at 645 and cases discussed therein. In Windisman v. Toronto College Park Ltd., [1996] O.J. No. 552, (Div. Ct.), Sharpe J., as he then was, noted at 536 that the “law's general policy favouring security of transactions means that the courts will not grant restitutionary recovery where the parties have entered into a contract”. See also Morrison-Knudsen Co. Inc. v. British Columbia Hydro & Power Authority, [1978] B.C.J. No. 1218, at 224-35 (B.C.C.A.). Most recently, see Organigram Holdings Inc. v. Downton, 2020 NSCA 38 at para. 46. [7] Whiten v. Pilot Insurance Co., 2002 SCC 18, [2002] 1 S.C.R. 595 at para. 36; Boucher v. Wal-Mart Canada Corp., 2014 ONCA 419 at para. 79. [8] Boucher, supra, note 7, at para. 79. [9] Whiten, supra, note 7, at paras. 78-83; Boucher, supra, note 7, at para. 80. [10] Liability Decision, supra, note 1, at para. 90. [11] Supra, note 7. [12] Class Proceedings Act, 1992, S.O. 1992, c. 6. [13] Liability Decision, supra, note 1 at para. 54. [14] Fresco v. Canadian Imperial Bank of Commerce, 2012 ONCA 444, at para. 109, referring to the reasons released the same day in Fulawka v. Bank of Nova Scotia, 2012 ONCA 443, at paras. 110-39. [15] Ibid. [16] Nolevaux v. King and John Festival Corporation, 2013 ONSC 5451, at paras. 13-19. [17] Markson v. MBNA Canada Bank, 2007 ONCA 334, at paras. 41-58. [18] Cassano v. Toronto Dominion Bank, 2007 ONCA 781, at paras. 39-43. [19] Cloud v. Canada (Attorney General) (2004), 73 O.R. (3d) 401 (C.A.) at para. 70, where the Court of Appeal interpreted s. 24(1)(c) as requiring proof of loss by “each” or every class member before aggregate damages are precluded, not proof of loss by “any” class member as was done in Fulawka. This is a key to understanding why Winkler C.J.O.’s interpretation of s. 24(1)(c) and his rejection of sampling to determine aggregate damages is, with respect, in error. A more detailed analysis of this error can be found in my discussion in Nolevaux, supra, note 16, at paras. 13-19. [20] Markson, supra, note 17, at para. 45. [21] 2038724 Ontario Ltd. v. Quizno’s Canada Restaurant Corporation, at paras. 122-123 and 142, affirmed 2010 ONCA 466, at paras. 54-58. [22] See, for example, Wilkins v. Rogers Communications Inc. at paras. 55-57; Bernstein v Peoples Trust Company, 2017 ONSC 752 at paras. 111-114; Bernstein v. Peoples Trust Company, 2019 ONSC 2867 at paras. 294-314; and Lee Valley Tools Ltd. v. Canada Post Corporation at para. 41. [23] Pro-Sys Consultants Ltd. v. Microsoft Corporation, 2013 SCC 57, at para. 134. [24] Markson, supra, note 17, at para. 44. [25] Vezina v. Loblaw Companies Ltd., [2005] O.J. No. 1974 (S.C.J.) [26] Ibid., at para. 25, where Cullity J. refers to “the possibility of such an assessment.” Also, if one Googles the meaning of “reasonable likelihood” one finds that courts and tribunals in other common law countries understand “reasonable likelihood” as meaning something more than “possible” but not much more – that is, the meaning given is “not fanciful or remote and more than merely plausible.” Whatever the nuance, “reasonable likelihood” is more akin to “reasonable possibility” and thus a relatively low standard. [27] Ramdath v. George Brown College, 2014 ONSC 3066. [28] Ibid., at para. 46. [29] Ramdath v. George Brown College of Applied Arts and Technology, 2015 ONCA 921. [30] Ibid., at para. 76. (Internal case citations omitted.) [31] Bozsik v. Livingston, 2019 ONSC 5340. [32] Section 12 of the CPA provides that “the court … may make any order it considers appropriate respecting the conduct of a class proceeding to ensure its fair and expeditious determination …”