COURT OF APPEAL FOR ONTARIO DATE: 20220209 DOCKET: C68590, C68649 & C68801

Lauwers, Harvison Young and Sossin JJ.A.

BETWEEN

Dara Fresco Plaintiff (Respondent)

and

Canadian Imperial Bank of Commerce Defendant (Appellant)

Counsel: Linda Plumpton, Sarah Whitmore, Ryan Lax, Lara Guest, Henry Federer, John Field, Lauri Reesor and Elisha Jamieson-Davies, for the appellant David F. O’Connor, J. Adam Dewar, Louis Sokolov, Steven Barrett, Peter Engelmann, Louis Century and Jody Brown, for the respondent

Heard: September 28-29, 2021

On appeal from the judgments of Justice Edward P. Belobaba of the Superior Court of Justice, dated March 30, 2020, with reasons reported at 2020 ONSC 75, 63 C.C.E.L. (4th) 60, dated August 10, 2020, with reasons reported at 2020 ONSC 4288, 66 C.C.E.L. (4th) 244, and dated October 21, 2020, with reasons reported at 2020 ONSC 6098.

Lauwers and Sossin JJ.A.:

A. Overview

[1] In 2007, Dara Fresco started a class action against the Canadian Imperial Bank of Commerce on behalf of 31,000 customer service employees who had worked for the Bank between 1993 and 2009. She claimed that two of the Bank’s policies enabled it to permit its employees to work overtime hours without appropriate compensation, contrary to the Canada Labour Code. [1]

[2] Two competing narratives set the scene. The Bank argued that its policies aimed to stop unnecessary overtime in order to control costs and to prevent overwork. To the contrary, Ms. Fresco argues that these policies resulted in the Bank getting the economic value of overtime work without compensating employees as required by the Code.

[3] The first overtime policy, which applied to employees in the retail branch network from February 1, 1993 to April 10, 2006, covered all class members (the “1993 Overtime Policy”). It provided for additional compensation where employees worked more than 8 hours a day or 37.5 hours a week, but required that employees get management approval before working overtime in order to receive payment (pre-approval). There was no provision for getting management approval after the overtime hours were worked (post-approval). The second policy, which applied to all of the Bank’s lines of business, was put in place starting on April 10, 2006 (the “2006 Overtime Policy”). This policy maintained the pre-approval requirement, but also allowed for post-approval in extenuating circumstances where approval was sought as soon as possible after the overtime work was done.

[4] This class action concerns the application of the Canada Labour Code to these two overtime policies. This court certified eight common issues in 2012. [2] The motion judge heard two summary judgment motions on the merits, one brought by each side.

[5] The motion judge released three decisions leading to these appeals. On liability, he granted summary judgment to Ms. Fresco as the representative plaintiff. [3] On damages, he certified aggregate damages as a common issue, leaving the merits of the proposed methodology for determining the class members’ individual damages entitlements to be assessed at a later stage. [4] On limitations, he dismissed the Bank’s summary judgment motion for a class-wide limitations order and left the Bank’s limitation defences to be addressed at the individual hearing stage. [5]

[6] For the reasons that follow, we would dismiss the Bank’s appeals.

B. The Issues

[7] Corresponding to the motion judge’s three decisions, the Bank brings three appeals. The Bank’s liability appeal raises two issues:

1. Did the motion judge misinterpret s. 174 of the Code?
2. In considering whether the Bank’s system-wide overtime policies and related practices contravened the requirements of the Code and the regulations under it, did the motion judge err in finding that the following were “institutional impediments” to the overtime claims of class members: (a) the Bank’s 1993 and 2006 Overtime Policies; and (b) the Bank’s record-keeping practices for tracking and compensating overtime hours?

[8] The Bank’s damages appeal raises one issue:

3. Did the motion judge err in certifying aggregate damages because this court had already determined that such damages are not available in this case?

[9] The Bank’s limitations appeal raises two issues:

4. Did the motion judge err by requiring the Bank to prove discoverability could be resolved on a class-wide basis, thereby reversing the onus of proof?
5. Did the motion judge err in refusing to answer the Bank’s constitutional question regarding the extra-territorial application of s. 28 of the Ontario Class Proceedings Act?

[10] We address each of these issues in turn.

[11] The Bank also submits that the motion judge did not directly address the evidence explaining its perspective but only the evidence supporting Ms. Fresco’s perspective as set out in her written submissions. The Bank argues that this was an error in itself and renders the reasons inadequate as a matter of law. We would not give effect to this submission. In our view, the motion judge took a broader view of the evidence than the Bank submits, as we point out from time to time.

C. Analysis

(1) The motion judge did not misinterpret s. 174 of the Code

[12] This issue relates to the first two certified common issues in the class action:

1. Are any parts of the Defendant's Overtime Policies (from February 1, 1993 to the present) unlawful, void or unenforceable for contravening the Canada Labour Code?
2. Did the Defendant have a duty (in contract or otherwise) to prevent Class Members from working, or a duty not to permit or not to encourage Class Members to work, overtime hours for which they were not properly compensated or for which the Defendant would not pay?

[13] To address these common issues, the motion judge had to interpret and apply s. 174 of the Canada Labour Code, which provides:

Overtime pay or time off

174 (1) Subject to any regulations made under section 175, when an employee is required or permitted to work overtime, they are entitled to (a) be paid for the overtime at a rate of wages not less than one and one-half times their regular rate of wages; or (b) be granted not less than one and one-half hours of time off with pay for each hour of overtime worked, subject to subsections (2) to (5).

[14] Then Code provisions make clear that the standard hours of work cannot exceed eight hours per day and forty hours per week. An employee who is “required or permitted” to work more than the standard hours of work must be paid time and a half. Additionally, under s. 24(2) of the Canada Labour Standards Regulations, [6] every employer is required to record the hours worked each day by every employee and keep this information on file for at least three years.

(a) The motion judge’s interpretation

[15] Against this legislative backdrop, what does the expression “required or permitted” in s. 174 of the Code mean? In his liability reasons, the motion judge focused on the interpretation of the word “permitted” and noted that “[t]he policy question is whether ‘permitted’ should be interpreted narrowly favouring the employer (and meaning ‘impliedly required’) or more broadly favouring the employee (and meaning ‘allowed’ or ‘not prevented’)”. [7] He cited Machtinger v. HOJ Industries, in which the Supreme Court answered this interpretive question in favour of employees, taking into account the power dynamics in the modern workplace and the importance of employment standards legislation:

The harm which the Act seeks to remedy is that individual employees, and, in particular non-unionized employees, are often in an unequal bargaining position in relation to their employers.… Accordingly, an interpretation of the Act which encourages employers to comply with the minimum requirements of the Act, and so extends its protections to as many employees as possible, is to be favoured over one that does not. [8]

[16] The motion judge explained that the labour arbitration cases interpreting s. 174 of the Code have, with only one exception, viewed the section as a worker protection provision in which “permitted” means “allow” or “fail to prevent”. He listed the main takeaways from the case law interpreting “permitted” in the context of overtime work as follows:

▪ The Code imposes liability for overtime “whenever it is permitted, even if it is not required or authorized. The intent of the Code is to protect employees who are simply allowed to work overtime without pay.” ▪ An employer cannot “simply look the other way when an employee is working beyond the standard hours” then claim the work was not required or permitted. ▪ An employer cannot “avoid these statutory obligations by knowingly permitting employees to work overtime and then later taking the position the overtime was not authorized. This is in fact the mischief sought to be avoided by the use of the word ‘permitted’ in Section 174.” ▪ In other words, an employer is liable for permitting overtime if it “acquiesce[s] by its failure to prevent.” [9]

[17] The motion judge accordingly restated the standard under s. 174 as: “When an employee is required or allowed to work or is not prevented from working in excess of the standard hours of work, the employee shall be paid for the overtime at a rate of wages not less than one and one-half times his regular rate of wages.” [10] We accept the motion judge’s interpretation of s. 174, which is well-supported by the case law.

(b) The Bank’s overtime policies breached s. 174 of the Code as interpreted by the motion judge

[18] We outline the parties’ arguments. The Bank makes two interrelated arguments about its overtime policies, one focussing on the purposes of its policies and the other focussing on the scope of its duties under the Code.

[19] First, the Bank highlights that the Code gives the employer the right to determine when overtime hours will be worked. The Bank argues that, respectful of the Code’s direction that an employer must pay overtime that it requires or permits, its overtime policies were developed for the purpose of discouraging overtime work. They were never intended to permit overtime work that was not compensated.

[20] The Bank argues that in finding the policies to be inconsistent with the Code, the motion judge ignored evidence in the policies themselves, which formed part of the employment contracts, that the Bank intended to comply with the Code. The Bank points to the 1993 Overtime Policy, which expressly stated that overtime work in excess of 8 hours in a day or 37.5 hours over a week would be paid. It also reminded employees to obtain the “approval of their supervisor or manager” before working overtime and instructs: “It is against the law to not pay overtime.” The text of the 2006 Overtime Policy, under the “Intent” heading, reads: “CIBC has developed this Employee Overtime Policy (Canada) to help management align our resources appropriately and in accordance with the legal and regulatory framework governing overtime.” The Bank argues that the clear intent of the policies was to comply with the Code. Accordingly, the Bank submits that the motion judge erred by interpreting the policies contrary to “the objective intention” of the Bank and a “common sense” reading of the policies’ text. [11]

[21] In response, the respondent argues that the motivation behind a policy is not relevant to whether it breaches the Code. Because the Bank’s policies caused uncompensated overtime on a systemic basis, the Bank breached s. 174 of the Code.

[22] Second, the Bank submits that although authorization was a pre-condition for working overtime hours, it was not a pre-condition for compensation. The 1993 Overtime Policy set out that employees would be compensated for overtime that a manager pre-approved. The 2006 Overtime Policy provided for both pre- and post-approval, each of which would result in compensation. The 2006 Overtime Policy read, in part:

In order for employees to be compensated for overtime hours worked, the hours must be pre-approved by a manager in advance. Overtime, for which prior management approval was not obtained, will not be compensated unless there are extenuating circumstances and approval is obtained as soon as possible afterwards.

[23] The Bank argues that pre-approval allowed it to control its employees’ hours of work and allocate resources in a cost-effective manner while compensating employees for overtime hours worked, as it was permitted to do, according to the authorities. The Bank relies on the principle that an employer must have knowledge of the overtime worked and that the employer must expressly or impliedly indicate that the work could be undertaken. In other words, “[e]mployees cannot unilaterally elect to perform work outside of their scheduled hours of work and claim compensation for such work at overtime rates or even at straight time rates”. [12]

[24] The Bank submits that the motion judge’s interpretation of the Code imposes a positive duty on an employer to prevent overtime hours it does not want to be worked, and requires the employer to prevent employees from working overtime hours that the employer is not aware are being worked. The Bank submits that imposing this positive duty is contrary to the Act and the approach taken in Lafarge, to the effect that overtime hours cannot be considered “required or permitted” where the employer has no knowledge of the work.

[25] The respondent relies on a countervailing line of arbitral decisions, which holds “that ‘permitted’ does not put the onus on the employee to seek permission; instead overtime is ‘required or permitted’ if an employer knows or ought to know that an employee is working overtime but fails to take reasonable steps to prevent the employee from working”. [13] The respondent also argues that the motion judge’s interpretation is consistent with the remedial purpose of the Code and case law which requires a liberal and generous interpretation of the protections afforded to employees.

[26] The motion judge held that the onus falls on the employer to show how the policy ensured all overtime hours were compensated. Relying on the T-Line Services line of cases, the motion judge’s focus was on whether the Bank breached its duty to the class by permitting or failing to prevent overtime hours, yet creating a system that all but prohibited overtime compensation. He concluded:

In my view, the plaintiff has established that both the 1993 and 2006 overtime policies contravened the requirements set out in s. 174 of the Code. The parties filed expert reports to support their respective submissions. There is no need for me to rely [on] or even refer to these reports. I find that the defendant bank breached its statutory and contractual duty to the class member employees. I can make this finding by simply contrasting the language of the defendant’s system-wide policies with s. 174 of the Code. [14]

[27] With respect to the 1993 Overtime Policy, the motion judge found that the imposition of a pre-approval requirement as a precondition for overtime compensation was more restrictive than the “required or permitted” language in s. 174 of the Code. With respect to the 2006 Overtime Policy, he found that the addition of possible post-approval in extenuating circumstances did not cure the deficiencies because the Code required that overtime be paid whenever such hours were required or permitted, without exception. However, the motion judge concluded that pre-authorization or post-approval requirements did not, in and of themselves, violate the Code; rather, it was the effect of making one or the other a pre-condition for payment that constituted the violation.

[28] As the motion judge pointed out, most arbitral decisions agree that s. 174 does impose a positive duty on employers to actively prevent employees from working overtime hours. This duty does not conflict with the employer’s right to manage its workforce. This duty does not subject an employer to indeterminate liability because the employer will not be found to have permitted overtime work unless the employer has actual or constructive knowledge that its employees are working beyond the hours permitted under the Code. The risk of silence in the face of actual or constructive knowledge falls on the employer. The motion judge did not impose a duty on employers to compensate employees for overtime hours of which it was not aware because an employer cannot be said to permit what it does not know.

[29] Moreover, as we set out below, the motion judge did not simply rely on the wording of the policies in finding liability. He also made findings of fact on the evidence that support his conclusion that the effect of the Bank’s overtime policies was that it failed to prevent overtime from being worked without compensation.

[30] We do not accept the Bank’s submission that the motion judge erred in his interpretation of the Code or its application to the Bank’s overtime policies. We find no legal error in the motion judge’s finding that, read as a whole, and together with the surrounding evidence of manuals, circulars and guidelines, the Bank’s overtime policies required pre-authorization (or post-authorization in extenuating circumstances in the case of the 2006 Overtime Policy). Overtime hours that were permitted but not authorized under the policies would not be paid, contrary to the Code.

[31] We turn now to the second issue.

(2) The motion judge did not err in finding that the Bank’s system-wide overtime policies and record-keeping practices breached its duties under the Code and its regulations

[32] In considering whether the Bank’s system-wide overtime policies and related practices contravened the requirements of the Code and the regulations under it, the motion judge did not err in finding that (a) the Bank’s 1993 and 2006 Overtime Policies; and (b) the Bank’s record-keeping practices for tracking and compensating overtime hours were “institutional impediments” to the overtime claims of class members.

(a) What is an “institutional impediment”?

[33] The language of “institutional impediment” draws on the comments of Strathy J. (as he then was) in Fulawka v. Bank of Nova Scotia. [15] He stated: “The understandable need for managers to control overtime costs and the pre-approval requirement in the policy create institutional impediments to claims for overtime pay.” [16] The motion judge echoed this language in the decision below, noting with respect to this court’s certification decision:

Chief Justice Winkler, writing for a unanimous Court, made clear that in order to prevail at the common issues trial, the plaintiff would have to prove that CIBC’s system-wide overtime policy and related practices were “institutional impediments” to class member overtime claims that were otherwise compensable under the Code. [17]

[34] In the barest terms, an impediment is “institutional” and therefore systemic if it is a characteristic of the operation of the employment system. When considering whether an employer’s policy or practice serves as an institutional impediment, the operative question is how employees were harmed by it. If the policy creates a systemic hurdle to appropriate compensation, then it operates as an institutional impediment. This is the case even if there were some employees who were not, in practice, denied compensation as a result of the policy.

[35] In this court’s certification decision, Fresco (ONCA), Winkler C.J.O. concluded that the lower courts’ view of the pre-approval policies “ignore[d] the factual assertions in the pleadings about the alleged reality of the workplace in CIBC retail branches.” [18] He noted that the “claim does not turn exclusively or even primarily on the per se legality of the [policies]”. [19] Rather, the alleged breach resulted from the interaction between the policies and the actual work assigned and recorded.

[36] Ultimately, this court certified the action because “CIBC’s overtime policies governing overtime compensation and the accompanying standard forms that class members submit when requesting such compensation, apply to all class members”. [20] The issue, according to Winkler C.J.O., was “whether CIBC had a duty to implement an overtime system that satisfies its obligations under the Code, and whether its actual system met these obligations”. [21]

[37] Similarly, in Cavanaugh v. Grenville Christian College, this court upheld a trial decision that found the defendant school to be systemically negligent because it caused harm through its “operational” characteristics. [22] The trial judge had rejected Grenville’s argument that the inflictions of harm were “one-offs” concerning individual students and that the harm was systemic because it flowed from Greenville’s character as an institution. In upholding the trial judge’s treatment of the systemic breach issue, van Rensburg J.A. stated: “The trial judge recognized that systemic negligence involved an assessment of how the school was run – its practices and the extent to which the practices created a risk of harm.” [23]

(b) The Bank’s overtime policies were institutional impediments

[38] The motion judge summarized the basis for his finding of liability in the following terms: “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.” [24]

[39] In other words, the motion judge found that the policies link the class members and their claims and create the class. The Bank’s breach was not systemic because it prevented all employees from receiving overtime compensation; rather, the breach was systemic because the policies acted as an institutional impediment for any employee that earned overtime compensation. The policies imposed additional hurdles on employees seeking overtime compensation systemically across the institution.

[40] The Bank argues that “the plaintiff’s systemic claim for breach of the class members’ employment contracts could only succeed on the merits if a causal link existed between class members’ claims for uncompensated overtime and CIBC’s policies and practices”. The Bank points out that the overtime policies were applied flexibly and that pre-approval was granted regularly. The Bank referred to evidence of 80 internal audits conducted between 2002 and 2009, 77 of which found no failure to compensate for overtime. The three audits that did find problems led to remedial action.

[41] The Bank asserts that the motion judge did not have sufficient evidence to conclude that its policies caused any employees to work overtime hours without compensation.

[42] However, the Bank’s argument misunderstands the motion judge’s reasoning. His approach flows from the systemic nature of the breaches. The Bank’s breaches were systemic because the regular denial of overtime pay resulted from the interaction between the Bank’s overtime policies, the Code, and its workforce. The analysis does not boil down to “an issue of numbers” because it does not depend on the interaction between individual managers and employees. [25] The motion judge expressed the test for liability under the third certified common issue, as: “One, has the plaintiff established that at least some of the class members worked uncompensated overtime? And two, has the plaintiff established that it is more likely than not that these hours of uncompensated overtime work were permitted or not prevented by the defendant bank?” [26]

[43] We agree with the motion judge’s approach. The class cannot establish that the Bank deprived class members of overtime compensation without first showing that such compensation was due: “What is a breach is failing to pay overtime that is actually owed”. [27]

[44] To succeed on this aspect of the claim, the respondent did not need to show that every class member was owed overtime compensation, but only that some class members were owed compensation because they were not paid as a result of the operation of the Bank’s overtime policies.

[45] It was therefore necessary for the motion judge to find that the policies in fact deprived some employees of overtime compensation. He found that some of the class members did work uncompensated overtime. There was ample evidence in the record to support this finding, including an open forum survey, a workplace effectiveness project, and theme reports. This evidence, produced by the Bank, included specific references to Bank employees working overtime hours that were not compensated. For example, the motion judge points to “hundreds of comments” relating to overtime in employee survey evidence produced by the Bank. [28]

[46] The motion judge’s line of analysis is consistent with other cases analyzing systemic breaches. For example, in Insurance Corp. of British Columbia, the arbitrator found irrelevant the employer’s argument that unpaid overtime was not a “pervasive” issue. The arbitrator wrote that the claim “is not an issue of numbers.” [29] The plaintiff in that case needed only to show that some employees worked beyond their regular hours because the issue was whether the employer “failed to put mechanisms in place to… prevent employees working beyond their regularly scheduled shifts.” [30]

[47] As we stated above, it is more appropriate to ask how employees were denied overtime compensation than to ask how many employees were denied compensation. Here, the respondent had to show that employees were denied overtime compensation because of the operation of the policies. The motion judge found that she did so. Since the policies impact the class in its entirety, this finding establishes the Bank’s liability to the class as a whole.

(c) The Bank’s record-keeping practices were institutional impediments

[48] This issue on appeal relates to the third certified question in the class action:

3. Did the Defendant have a duty (in contract or otherwise) to accurately record and maintain a record of all hours worked by Class Members to ensure that Class Members were appropriately compensated for same?

[49] Earlier we noted that under s. 24(2) of the Regulations, every employer is required to record the hours worked each day by every employee and keep this information on file for at least three years.

[50] In the Bank’s submission, the Regulations require employers to keep records of the time worked and the compensation paid (including for overtime hours) for a period of three years. It does not specify the manner of such recording.

[51] The Bank delegated the task of recording compensated hours to individual branch managers. After 2003, the Bank relied on a human resources software program, PeopleSoft, to record hours. The Bank’s evidence was that this software met the industry standard at the time.

[52] The Bank’s record keeping consisted of timesheets. Managers were responsible for inputting overtime hours into the payroll system and ensuring payment, including time off in lieu of payment if elected by the employee. The Bank produced evidence from various affiants that employees completed timesheets daily or when their hours of work differed from their standard hours. According to the Bank, there was no “direct evidence” that only “approved” hours were recorded on these timesheets.

[53] The Bank challenges the motion judge’s assessment that “[i]n the vast majority of cases, the only hours recorded were the regular hours and the approved overtime hours”, as opposed to all overtime hours worked. The Bank argues that the motion judge “ignored” its evidence on this point, and that there was no evidence in the record capable of supporting his conclusion.

[54] The motion judge relied on the systemic nature of the respondent’s claims to explain his finding of a breach on this basis:

Despite the defendant’s submissions to the contrary, I have no difficulty finding on the evidence before me that actual hours of work were not recorded. This was a system-wide, indeed systemic deficiency, that contravened the Code.

The defendant bank expected and directed class members to write down their actual hours only on an “exceptional basis… when they sought to be paid for hours worked beyond their regularly scheduled hours.” The timesheet that the bank says was used for seeking post-approval expressly repeats the pre-approval requirement. To reiterate, system-wide policies told class members that overtime work would not be compensated unless it was pre-approved (or post-approved in extenuating circumstances after 2006). Hours worked that were otherwise permitted (not prevented) were not recorded and not compensated. [31]

[55] Consequently, the motion judge accepted the respondent’s position that the Bank had breached its duty to the class to ensure that all their hours of work were recorded, and that all required or permitted overtime was compensated.

[56] We reiterate that resolving the issue of liability in a systemic class action is not a question of numbers. It was unnecessary for the motion judge to quantify the number of cases where uncompensated overtime hours were not captured due to the Bank’s system of record-keeping. To support a finding of a breach of this common issue, the motion judge had to find instances in which some employees’ real hours of work were not recorded due to the record-keeping system. Having done so, it was not material whether this occurred in a majority of cases.

[57] Based on his evaluation of the evidence, the motion judge accepted that the actual hours of work might have been recorded for some employees at some branches on some occasions. However, he found the Bank was deficient in having no system to ensure this was done consistently across all branches. There was sufficient evidence in the record to support such a conclusion.

[58] The motion judge relied, for example, on admissions contained in bank documents entitled “Overtime Policy Canada - Compliance Monitoring” [32] and Overtime Monitoring Reports”, which indicated that PeopleSoft was not used to track actual hours worked and that the Bank was unable to determine whether an employee worked overtime due to that lack of tracking.

[59] In our view, based on the evidence before the motion judge, it would have been preferable if he had refrained from quantifying the occasions when the Bank’s system of record-keeping led to overtime hours not being recorded. That said, we see no basis for interfering with the motion judge’s finding on this aspect of the liability analysis.

(3) The motion judge did not err in certifying the aggregate damages issue

[60] A plaintiff’s access to damages determined in the aggregate is governed by s. 24 of the Class Proceedings Act, which provides:

24 (1) The court may determine the aggregate or a part of a defendant’s liability to class members and give judgment accordingly where, (a) monetary relief is claimed on behalf of some or all class members; (b) no questions of fact or law other than those relating to the assessment of monetary relief remain to be determined in order to establish the amount of the defendant’s monetary liability; (c) the aggregate or a part of the defendant's liability to some or all class members can reasonably be determined without proof by individual class members.

[61] We address the background to this issue, the test for certifying an aggregate damages issue, and whether this court’s refusal to certify an aggregate damages common issue is res judicata.

(a) Background

[62] On the appeal of the certification motion, this court refused to certify the proposed common issue concerning the aggregate assessment of damages. Winkler C.J.O. stated: “For the reasons given in Fulawka, at paras. 110-39, the preconditions in s. 24(1) of the CPA for ordering an aggregate assessment of monetary relief cannot be satisfied in this case.” [33]

[63] The particular stumbling block in Fulawka concerned the language of s. 24 (1)(c), which requires that “the aggregate or a part of the defendant’s liability to some or all class members can reasonably be determined without proof by individual class members.” In Fulawka, this court found that because the “proposed method is based on proof from a limited subsection of the class, it… impermissibly requires proof from individual class members in order to arrive at an aggregate damages figure.” [34] Winkler C.J.O. summarized: “[A]n aggregate assessment of monetary relief may only be certified as a common issue where resolving the other certifiable common issues could be determinative of monetary liability and where the quantum of damages could ‘reasonably’ be calculated without proof by individual class members.” [35]

[64] Despite this court’s refusal to certify the aggregate damages question, the plaintiff sought an order directing an assessment of aggregate damages (or certifying aggregate damages as a new common issue) and directing the Bank to produce paper and electronic records relevant to the aggregate assessment. The motion judge heard further argument and certified a ninth common issue:

9. Can the defendant’s monetary liability be determined on an aggregate basis? If so, in what amount? [36]

The motion judge adjourned the balance of the damages hearing “to await the plaintiff's aggregate damages report and the defendant bank’s response”. [37]

[65] The Bank challenges the certification of the aggregate damages issue on two grounds: (i) the proposed methodology does not meet the legal test for certifying the aggregate damages issue, and (ii) this court finally determined that aggregate damages were not available in this case, so the matter is res judicata. We address these arguments, both of which we reject, in turn.

(b) The test for certifying the aggregate damages issue is met

[66] Apart from the requirements under s. 24(1) of the Act, there are also jurisprudential governing principles.

(i) The governing principles for certifying an aggregate damages common issue

[67] The test for certifying aggregate damages as a common question is whether there is “a ‘reasonable likelihood’ that the conditions required in s. 24 of the Class Proceedings Act for determining aggregate damages would be satisfied if the [plaintiff is] otherwise successful at the common issues trial”. [38]

[68] The Supreme Court considered the standard for evaluating a plaintiff’s proposed methodology in Pro-Sys Consultants Ltd. v. Microsoft Corporation. [39] Rothstein J. said: “[T]he expert methodology must be sufficiently credible or plausible to establish some basis in fact for the commonality requirement”. [40] As Karakatsanis J. explained in her partly dissenting reasons in Atlantic Lottery Corp. Inc. v. Babstock, this means that the methodology must offer “a realistic prospect of assessing class-wide monetary relief in the aggregate”. [41]

[69] Rothstein J. added in Pro-Sys that the methodology “cannot be purely theoretical or hypothetical, but must be grounded in the facts” and there must be some evidence that data is available. [42] He noted that “resolving conflicts between the experts is an issue for the trial judge and not one that should be engaged in at certification”. [43] Finally, Rothstein J. found that the common issues trial judge has the “ultimate” responsibility for deciding whether aggregate damages are available. [44]

(ii) The motion judge’s reasoning

[70] The motion judge saw his responsibility as determining whether the methodology proposed by the plaintiff’s expert offered a “reasonable possibility” of assessing damages in the aggregate without proof by individual class members, and that the methodology would result “in a fair and sufficiently reliable determination of the defendant’s monetary liability”. [45]

[71] The motion judge described the proposed methodology. It 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.” [46] He explained: “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.” [47] The motion judge was confident in the methodology because it had been used successfully in “scores of [American] unpaid overtime cases.” [48] He added that the respondent’s expert had reviewed the limited available data respecting five employees, which helped inform the formulation of his proposed methodology. [49]

[72] This led the motion judge to conclude “that it is reasonably possible that [the expert’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.” [50] He pointed out that “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.” [51]

[73] As we discuss in more detail below, the motion judge took the position that Pro-Sys, which was decided after this court’s refusal to certify the aggregate damages question, allowed him to reconsider the aggregate damages certification issue.

(iii) The governing principles applied

[74] The motion judge properly expressed the standard for certifying aggregate damages as being “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”. [52] He cited Markson and observed that the “reasonable likelihood” standard originated in Cullity J.’s comments in Vezina v. Loblaw Companies Ltd. [53]

[75] The motion judge added this footnote, with which the Bank takes issue:

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. [54]

The Bank argues that the motion judge erred in adopting a “reasonable possibility” test instead of the “reasonable likelihood” one.

[76] We would not give effect to this argument because, on the facts of the case, the “reasonable likelihood” standard has been met, for all the reasons the motion judge provided about the methodology. To paraphrase Pro-Sys, the proposed methodology is sufficiently credible or plausible to establish some basis in fact for the commonality requirement, and it offers a realistic prospect of assessing class-wide monetary relief in the aggregate that is grounded in the facts and the available data. It is the task of the trial judge, not the certification motion judge, to resolve any conflicts between the experts.

[77] We now turn to the Bank’s other assertion, which is that because this court had previously finally determined that aggregate damages were not available in this case, the matter is res judicata.

(c) This court’s decision refusing to certify aggregate damages does not render the issue res judicata

[78] We address the doctrine of res judicata in three steps: the governing principles; the motion judge’s reasoning; and the principles applied.

(i) The governing principles for res judicata

[79] The governing principles for the issue estoppel branch of res judicata were prescribed by the Supreme Court in Danyluk v. Ainsworth Technologies Inc. [55] A party is prohibited from re-litigating an issue where (1) the same issue has been previously decided; (2) that judicial decision was final; and, (3) the parties are the same. Binnie J. stated the purpose of the doctrine:

The law rightly seeks a finality to litigation. To advance that objective, it requires litigants to put their best foot forward to establish the truth of their allegations when first called upon to do so. A litigant, to use the vernacular, is only entitled to one bite at the cherry. [56]

He added: “The underlying purpose [of issue estoppel] is to balance the public interest in the finality of litigation with the public interest in ensuring that justice is done on the facts of a particular case.” [57]

[80] Tulloch J.A. considered the rationale for preventing re-litigation in the context of the abuse of process doctrine in The Catalyst Capital Group Inc. v. VimpelCom Ltd. [58] He noted: “The law… seeks to avoid re-litigation primarily for two reasons: first, to prevent overlap and wasting judicial resources; and second, to avoid the risk of inconsistent findings.” [59]

[81] Even where the three requirements for issue estoppel are met, the courts retain a residual discretion to refuse to apply the doctrine. Finch J.A. stated in British Columbia (Minister of Forests) v. Bugbusters Pest Management Inc.:

The doctrine of issue estoppel is designed as an implement of justice, and a protection against injustice. It inevitably calls upon the exercise of a judicial discretion to achieve fairness according to the circumstances of each case. [60]

Binnie J. adopted this statement in Danyluk. [61] He commented that “[t]he objective is to ensure that the operation of issue estoppel promotes the orderly administration of justice but not at the cost of real injustice in the particular case.” [62]

[82] This court commented on the judicial discretion to refuse to apply issue estoppel in Schweneke v. Ontario. [63] Doherty and Feldman JJ.A. stated: “In exercising the discretion the court must ask – is there something in the circumstances of this case such that the usual operation of the doctrine of issue estoppel would work an injustice?” [64] More recently, this court discussed the role of discretion in declining to apply res judicata and the related doctrine of abuse of process in Dosen v. Meloche Monnex Financial Services Inc. (Security National Insurance Company). [65]

[83] Because declining to give effect to issue estoppel is a matter of discretion, this court owes deference to a motion judge’s decision and should only intervene “if the motions judge misdirected himself, came to a decision that is so clearly wrong as to be an injustice, or gave no or insufficient weight to relevant considerations.” [66]

(ii) The motion judge’s decision

[84] In his damages decision, the motion judge took the position that he was free to add aggregate damages as a ninth common issue, despite this court’s earlier refusal, on two bases. First, he determined that the doctrine of res judicata did not strictly apply because the proposed methodology was different than the initial sampling methodology this court rejected. Second, he noted that in Pro-Sys, which was released after this court’s certification decision, the Supreme Court established that the trial judge has ultimate authority “to add an aggregate damages question even where this very question was rejected at certification.” [67]

(iii) The governing principles applied

[85] In Pro-Sys, Rothstein J. held that the trial judge has ultimate responsibility for deciding whether aggregate damages are available:

The question of whether damages assessed in the aggregate are an appropriate remedy can be certified as a common issue. However, this common issue is only determined at the common issues trial after a finding of liability has been made. The ultimate decision as to whether the aggregate damages provisions of the CPA should be available is one that should be left to the common issues trial judge. Further, the failure to propose or certify aggregate damages, or another remedy, as a common issue does not preclude a trial judge from invoking the provisions if considered appropriate once liability is found. [68]

[86] We make two observations about these words in Pro-Sys. First, the Supreme Court does not discourage the identification of aggregate damages as an issue to be certified at the outset. This makes sense because early identification of an issue is always a good thing. Second, the Supreme Court leaves the final decision about the availability of aggregate damages to the trial judge, even where the issue was not previously proposed or certified. This too makes sense because the trial judge becomes deeply familiar with the case as it crystallizes, which makes the trial judge uniquely able to make the appropriate call.

[87] Does the law’s clarification or change in Pro-Sys displace this court’s refusal to certify aggregate damages as a common issue? In our view, it does. This court made the “ultimate decision” that the Supreme Court later stipulated “should be left to the common issues trial judge”. In other words, neither the certification judge’s refusal nor this court’s refusal on appeal to certify aggregate damages as a common issue should be the final disposition. We are obliged to give effect to Pro-Sys. [69] Accordingly, the motion judge was correct in concluding on the basis of Pro-Sys that as the common issues judge, he had the ultimate authority to certify the aggregate damages common issue.

[88] However, we add that Pro-Sys does not displace this court’s earlier legal ruling on sampling as a methodology for determining aggregate damages. The motion judge alluded quite fairly to the implication of potential gaps in the evidence: “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.” [70] This raises the prospect that this court’s legal finding that random sampling of the class members is not an acceptable method for determining aggregate damages might need to be revisited.

[89] The motion judge has taken the strong position that Winkler C.J.O.’s analysis of the sampling methodology was “probably wrong”, but he explained that the question was premature in this case: “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.” [71]

[90] Time will tell if statistical sampling will be needed to fill evidentiary gaps. If it is used, then the Bank could challenge the result based on this court’s ruling on the sampling methodology. It will then be open to the respondent to argue, based on a full evidentiary record, that this court’s decision was wrong and should be set aside.

(4) In determining whether the Bank had a class-wide limitations defence, the motion judge did not err in requiring the Bank to prove discoverability could be resolved on a class-wide basis

[91] This court did not certify the effect of limitation periods as a common issue. Winkler C.J.O. said: “The issue of limitation periods is not an ingredient of the class members’ claims, but instead may be relied upon by CIBC in its defence.” [72] He continued: “The question of how individual issues are best resolved is a procedural matter that would follow after the common issues trial.” Despite these words, the Bank cross-moved for summary judgment on the limitation issue.

[92] An effective class-wide limitation defence would greatly assist the Bank, which is faced by a class period that begins on February 1, 1993, when the 1993 Overtime Policy took effect, and ends on June 18, 2009, the certification date approved by this court. The class action is national in scope, with a 16-year class period and about 31,000 class members.

[93] The Bank focused on two arguments. The first concerns the application of the “discoverability” test in Ontario’s Limitations Act, 2002 [73] and its analogs in Saskatchewan and Alberta’s respective limitations statutes. The second concerns the non-application of the “appropriate means” aspect of discoverability to claimants residing in parts of Canada in which the relevant limitations legislation does not include a statutory discoverability test.

(a) The legislated discoverability test

[94] The Bank’s argument hinges on the issue of discoverability. Section 5 of Ontario’s Limitations Act, 2002 provides:

5 (1) A claim is discovered on the earlier of, (a) the day on which the person with the claim first knew, (i) that the injury, loss or damage had occurred, (ii) that the injury, loss or damage was caused by or contributed to by an act or omission, (iii) that the act or omission was that of the person against whom the claim is made, and (iv) that, having regard to the nature of the injury, loss or damage, a proceeding would be an appropriate means to seek to remedy it; and (b) the day on which a reasonable person with the abilities and in the circumstances of the person with the claim first ought to have known of the matters referred to in clause (a). (2) A person with a claim shall be presumed to have known of the matters referred to in clause (1) (a) on the day the act or omission on which the claim is based took place, unless the contrary is proved.

[95] The motion judge discussed the test set by s. 5: “[L]imitation periods begin to run as soon as the claimant reasonably discovers that she has sustained a loss, that the loss was caused by the defendant and that taking legal action was appropriate.” [74] The motion judge noted that: “Every time a class member received their bi-weekly pay, they would have known if they had been paid for overtime, and if not, that this loss was caused by their defendant employer.” [75] Accordingly, the first two branches of the test were met.

[96] The discoverability issue rested, for the motion judge, on the third branch: whether class members knew taking legal action was appropriate. This turns on the interpretation of ss. 5(1)(a)(iv) and 5(1)(b).

[97] The motion judge found that the “appropriate means” requirement applied so that the limitations period would “not begin to run if taking legal action was not reasonably appropriate given the plaintiff’s circumstances.” [76] He gave two main reasons for concluding that the “appropriate means” test was not met. First, “some (and perhaps many) of the class members feared reprisal and were afraid that they might lose their job if they sued the bank for unpaid overtime”. [77] Second, “some (and perhaps many) of the class members reasonably relied on the bank’s repeated misrepresentations throughout the 16-year class period that the bank’s overtime policies complied with federal labour law.” [78]

[98] The motion judge found that these reasons combined to require individual assessments of when discoverability was met for an individual claimant, consistent with the general rule that “the viability of a limitations defence is best determined on an individual basis with individual assessments – hence its usual relegation to the individual hearings phase.” [79] The motion judge concluded:

The defendant bank has not established on the evidence that the limitation period that applies to every class member’s claim (outside the limitation periods noted in its Schedule) can be determined in common on a class-wide basis and that individual discoverability is not needed. In my view, the evidence strongly suggests that individual discovery will be needed in at least some cases to fairly determine whether the class member delayed in taking legal action because they were in reasonable fear of losing their job; because they reasonably relied on the bank’s misrepresentations about the legality of its overtime policy; or because they were otherwise impeded by the bank’s systemic policies and practices. [80]

[99] We are not persuaded that the first factor, that “some (and perhaps many) of the class members feared reprisal and were afraid that they might lose their job if they sued the bank for unpaid overtime” is a valid basis on which the limitations period can be suspended. However, there is merit in the second factor of reasonable reliance on misrepresentation. The applicable law is set out in this court’s decision in Presley v. Van Dusen. [81] Sharpe J.A. discussed the governing principles, and then referred to one of the “guiding principles” expressed by Pardu J.A. in Presidential MSH Corp. v. Marr Foster & Co. LLP: “Resort to legal action may be ‘inappropriate’ in cases where the plaintiff is relying on the superior knowledge and expertise of the defendant, which often, although not exclusively, occurs in a professional relationship.” [82]

[100] Sharpe J.A. added:

Moreover, reliance on superior knowledge and expertise sufficient to delay commencing proceedings is not restricted to strictly professional relationships. I acknowledge that the previous cases where this court has made a finding that it was reasonable for the plaintiff to rely on the defendant’s superior knowledge and expertise have concerned defendants belonging to traditional expert professions.... However, recent Superior Court decisions have applied the superior knowledge and expertise prong of Presidential MSH to persons who are members of non-traditional professions or who are not professionals at all. [83]

He pointed to a case involving a franchisor-franchisee relationship, and another involving portfolio managers and investors. The categories are not closed. [84]

[101] On the facts of this case, it is quite plausible, as the motion judge found, that some class members reasonably relied on the Bank’s misrepresentations that its overtime policies complied with federal labour law. The influence of this factor on individual class members is really a matter best left to individual assessment, as this court noted in the earlier certification decision.

(b) Common law discoverability

[102] The Bank’s second argument is that the question of whether a class member knew that a proceeding was an “appropriate means” to remedy unpaid overtime is only relevant to class members’ claims governed by statutes that include such discoverability language, that is, claims in Ontario, Saskatchewan, and Alberta. On that basis, a class-wide limitations order would be appropriate for all other claims. The Bank adds that this argument might also extend to claims in Ontario, Saskatchewan, and Alberta that predate the amendments adding discoverability language into the statutory text.

[103] The Bank argues that the “appropriate means” criterion in s. 5(1)(a)(iv) of the Limitations Act, 2002 is not an element of the common law discoverability rules, relying on 407 ETR Concession Company Limited v. Day and other cases. [85] We do not agree that common law discoverability rules could not be found to function in an equivalent manner. The ordinary development of the common law means that the categories are not closed. Whether this argument has traction is a matter to be decided on the individual assessments and not on a fact-free, class-wide basis.

(c) Reversing the onus

[104] The Bank also argues that the motion judge effectively reversed the burden of proof applicable to discoverability. Once it had proven that the claimants were aware of their claims, the Bank argues that “the burden was on the plaintiff to establish that there was a basis to delay the running of the limitations period.”

[105] We would not give effect to this argument. Having moved for summary judgment, the onus was on the Bank to establish that it was so entitled. In any event, as we explain above, the respondent has established a sufficient basis to require the application of the limitations defence to be worked out on an individual basis.

(5) The motion judge did not err in refusing to address the purported extra-provincial reach of s. 28 of the Class Proceedings Act

[106] This class proceeding has a national reach, with class members across the provinces. The Bank argues that because limitation periods affect the substantive rights of plaintiffs and defendants, they fall squarely within provincial power over “property and civil rights” under s. 92(13) of the Constitution Act, 1867. Accordingly, because Ontario may not legislate extra-territorially with respect to substantive rights, s. 28(1) of the Class Proceedings Act – which suspends the running of limitation periods in favour of class members – should not serve to suspend the limitation periods applicable under local legislation to claims of class members who reside outside of Ontario.

[107] The motion judge declined to consider this constitutional argument, on the basis that ruling on the extra-territorial applicability of s. 28 would be premature. He stated:

The defendant bank has asked that I rule on the s. 28(1) extra-provincial question even if I dismiss its request for a class-wide limitations order because this constitutional question may arise again at the individual hearings stage. I decline to do so. This litigation may never reach an individual hearings stage. The constitutional question is premature. [86]

[108] The Bank asks this court to consider this constitutional question on the basis that it must be resolved for all claims governed by non-Ontario law. It asserts that rendering a decision on a class-wide basis would preserve judicial economy, efficiency, and consistency in results.

[109] We decline to decide this issue in the absence of a lower court decision and in the absence of a better evidentiary landscape. We agree with the motion judge that the issue is premature, and we therefore also decline to remit the matter back to him for resolution. Courts should not decide constitutional questions unnecessarily. [87]

D. Disposition

[110] We would dismiss the appeal with costs payable to the respondent. If the parties are unable to agree on costs, then the respondent may file a written submission no more than three pages in length within ten days of the date of the release of these reasons; the appellant may file a written submission no more than three pages in length within ten days of the date the respondent’s submission is due; and the respondent may file a reply submission no more than one page in length within five days of the date the appellant’s submission is due.

Released: February 9, 2022 “P.L.” “P. Lauwers J.A.” “L. Sossin J.A.” “I agree. Harvison Young J.A.”