Court File and Parties

CITATION: MacKinnon v. Volkswagen Group Canada Inc., 2022 ONSC 5501

DIVISIONAL COURT FILE NO.: 852/21

DATE: 2022-10-04

ONTARIO SUPERIOR COURT OF JUSTICE DIVISIONAL COURT

McWatt A.C.J.S.C., Molloy, and Morgan JJ

BETWEEN:

Stuart MacKinnon Appellant

– and –

Volkswagen Group Canada, Inc., Volkswagen Aktiengesellschaft, Volkswagen Group of America, Inc., Audi Canada, Inc., Audi Aktiengesellschaft, Audi of America Inc., and VW Credit Canada, Inc. Respondents

COUNSEL: David O’Connor and J. Adam Dewar, for the Appellant Cheryl Woodin, Ilan Ishai, and Peter Douglas, for the Respondents

HEARD: August 18, 2022

Reasons for Judgment

MORGAN J.

I. Introduction

[1] The Supreme Court of Canada has instructed that the expert methodology presented by a plaintiff seeking certification under the Class Proceedings Act, 1992, SO 1992, c 6 (“CPA”) must be “sufficiently credible or plausible to establish some basis in fact” to support the claim: Pro-Sys Consultants Ltd. v. Microsoft Corporation, 2013 SCC 57, [2013] 3 SCR 477, at para 118. In analyzing the expert evidence, it has been made clear that at this stage “it is not necessary that the methodology establish the actual loss sustained, just that a sufficiently credible or plausible methodology is capable of doing so”: Kalra v Mercedes Benz, 2017 ONSC 3795, at para 48.

[2] This appeal pushes the point one step further. It specifically asks: just how “credible or plausible” must the methodology actually be when it comes to presenting a claim for damages on a class-wide basis?

[3] The Appellant brings this appeal from the order of Justice Edward Belobaba dated September 23, 2021: Mackinnon v. Volkswagen, 2021 ONSC 5941. The motion judge dismissed the request for certification on the ground that the record contained no basis in fact for damages to be determined as a common issue.

II. Issues on this appeal

[4] The sole issue on appeal is whether there is a plausible methodology for measuring compensable loss on a class-wide basis.

III. Jurisdiction of the Court

[5] The appeal comes to Divisional Court as of right pursuant to section 30(1) of the CPA.

IV. Standard of Review

[6] The standard of review for judicial appeals, including those under the CPA, has been set out in Housen v. Nikolaisen, 2002 SCC 33, [2002] 2 SCR 235. On questions of law the standard is correctness, while on questions of fact the standard is palpable and overriding error. Swinton J. explained in Maginnis v. FCA Canada Inc., 2021 ONSC 3897, at para 17 (Div Ct), that on questions of mixed fact and law there is a spectrum; where the question is the application of the correct legal principles to the evidence, the standard is palpable and overriding error.

[7] As indicated at the outset, the case on appeal focused on the question of whether the Appellant’s expert witnesses presented a plausible methodology for measuring damages. That question, in turn, is an important part of the analysis of whether there is some basis in fact for the claim. Counsel for both parties agree that the issue is one of fact finding or the application of the relevant legal standard to the facts. The Appellant does not challenge the motion judge’s articulation of the law; rather, the appeal challenges the motion judge’s finding of fact – i.e. that there is no plausible methodology and thus no basis in fact for the damages claim.

[8] Findings of fact are generally entitled to deference on review or appeal: Schwartz v. Canada, [1996] 1 SCR 254, at para 32. Likewise, a finding of no facts, or ‘no basis in fact’, is also entitled to deference. Accordingly, the motions judge’s ruling can be interfered with only if there is a palpable and overriding error.

V. Background to the claim

[9] The claim was issued following revelations on September 18, 2015 that the Respondents had installed an unlawful “defeat device” in their diesel-powered vehicles which was designed to subvert automobile emissions regulations. The vehicles were marketed to purchasers under the misrepresentation that they were clean, emissions compliant, and environmentally friendly.

[10] The revelations led to criminal prosecutions and, ultimately, guilty pleas by the Respondents. They also resulted in class actions in multiple jurisdictions brought by purchasers of the vehicles seeking to recover the post-disclosure drop in the value of their automobiles. Those actions, in turn, resulted in large settlements, including a $2.1 billion Canadian settlement: Quenneville v. Volkswagen, 2017 ONSC 2448.

[11] The current action has been filed on behalf of Volkswagen and Audi owners and lessees who were not included in the previous settlement. These claimants no longer owned vehicles that were fitted with the fraudulent defeat device at the time of its revelation. The class proposed by the Appellant is composed of purchasers of the Respondents’ diesel-powered vehicles containing the defeat device who sold their vehicles prior to the September 2015 announcement. The action seeks damages reflecting what are characterized as pre-disclosure losses sustained by these proposed class members.

[12] The motion judge found that the Appellant and his expert witnesses presented no plausible methodology to calculate damages. He also dismissed the request to certify the proposed common issues that were unrelated to the damages calculation – e.g. the liability issues that were contained in the Appellant’s proposed common issues – on the basis that if damages cannot be proved then the other issues do not move the action ahead in a meaningful way.

VI. Analysis

a) The expert evidence

[13] The factual basis of the claim is not in dispute in this appeal. That is, the Respondents’ use of the defeat device, the nature and purpose of that device, the misrepresentations about emissions compliance that accompanied the marketing of the subject vehicles, and the negative impact of all of this on the Respondents’ market reputation at large, have been well established in other cases and were taken as a given by the motion judge.

[14] The focus here is on what the motion judge called the “threshold issue” of whether damages can be calculated for this group of owners who sold their vehicles prior to the defeat device’s disclosure to the market. As the motion judge put it, at para 12 of his reasons for judgment, “The damages claim is straightforward: the putative class members allegedly paid for a clean diesel or low emissions benefit that they never received.” The question on which the certification turned is whether there is a plausible methodology for measuring this straightforward claim on a class-wide basis.

[15] It is important to note that the issue of controversy was not whether harm was suffered by the proposed class. That, too, was taken as a given under the circumstances of the case. Rather, the controversial question for the parties and for the court below was whether it could be determined how much this harm was worth. As the motion judge put it, at paras 18-19:

[18] There is no doubt that the pre-disclosure owners and lessees of the impugned VW and Audi vehicles paid a diesel-engine premium, in some cases, of several thousand dollars more than what they would have paid for the gasoline-engine equivalent. There is also no doubt that the pre-disclosure owners and lessees had received many of the diesel-related benefits when they sold or returned their vehicle — such as improved fuel economy, increased torque and longer engine life. However, it is also clear that they did not receive the promised low emission or clean diesel benefit.

[19] The question, however, is whether there is any evidence that the pre-disclosure plaintiff paid anything extra for the clean diesel feature. That is, can the low emissions or clean diesel feature that was not provided as promised be isolated and quantified in monetary terms for the purposes of this proposed class action? And, if so, has the plaintiff presented a plausible methodology for measuring this loss on a class-wide basis?

[16] In other words, it was accepted as a starting point that “class members acquired less value than what they expected to acquire – by purchasing a product which was represented to be ‘proven’ by science or in clinical trials to help provide the claimed effects, but which was not proven to do so”: Drynan v. Bausch Health Companies Inc., 2021 ONSC 7423, at para 257. Or, as an Australian court observed in another emissions-related class action, it was “glaringly obvious…that the presence of the Core Defect in the Relevant Vehicles reduced the value of the vehicles at the time they were supplied and each consumer therefore received a vehicle that was less valuable than the one they bargained for, resulting in an overpayment”: Williams v. Toyota Motor Corporation, 2022 FCA 344, at paras 173, 330.

[17] What was not glaringly obvious was whether, and how, the reduced value could ever be measured. The motion judge looked carefully at the expert evidence on damages; in fact, he gave the Appellant a second chance to improve the record and then looked again. But he ultimately found, at para 48, that the Appellant had failed “to provide…some evidence that would isolate the alleged overpayment for the clean diesel feature by the pre-disclosure plaintiff and provide the required methodology to measure this loss on a class-wide basis.”

[18] The motion judge concluded that the failure to provide a methodology for proving damages was fatal to the certification effort, reasoning to the effect that tortious conduct “is wrongful only to the extent that it causes damage”: Atlantic Lottery Corp. Inc. v. Babstock, 2020 SCC 19, at para 37. He acknowledged that a certification motion is not the place to delve into the merits of an expert report or to resolve a conflict between the two sides’ experts, but he opined that the Appellant had failed in any way to “present a proposed methodology by a qualified person whose assumptions stand up to the lay reader”: Mackinnon, at para 51, quoting 038724 Ontario Ltd. v. Quizno’s Canada Restaurant Corporation, at para 102 (Div Ct).

[19] In an effort to provide a record that could satisfy the lay reader, the Appellant invested great deal of effort, and in this appeal has continued to invest substantial effort, in establishing that the clean emissions feature of the subject vehicles had value. Appellants’ counsel and expert witnesses contend that, as a corollary, the purchasers of these vehicles – including the class proposed here – must have been deprived of that value.

[20] Appellant’s counsel on appeal emphasize that this evidence comes from a primary source (the Appellant himself) as well as from the Appellants’ economic loss expert (Edward Stockton) and, indeed, from admissions by the Respondents’ economics expert (Dr. Lorin Hitt). As they explain it in their factum:

Mr. Stockton opined that Class Members had overpaid because the price or perceived value of the Vehicles was improperly inflated at acquisition for two rational reasons or factors: (1) Class Members paid for a Vehicle that was represented as having a certain valuable feature (so-called particularly clean/environmentally friendly emissions) which was in fact not available or provided; and (2) not only were the Vehicles not particularly “clean” and environmentally friendly (the positive feature) but they also came with an undisclosed Defeat Device…

Although the evidence from the Representative Plaintiff himself about paying more for the TDI Jetta (primarily because of the clean diesel representation) should have answered the question posed by the Motions Judge above, there was more evidence in the record of value.

Importantly, the fact that the clean diesel feature had value and was part of the entire TDI premium payment was admitted by VW’s own expert (Dr. Hitt) who confirmed or testified as follows:

a. Dr. Hitt explained that the TDI premium combines features including improved fuel economy, increased torque, engine longevity and greater range. Another element of the TDI premium is a low NOx emissions premium. …

[21] To a certain extent one can understand the motion judge’s frustration with this evidence. Everyone, including the Respondents and the motion judge, understood from the outset that the Respondents had advertised a clean emissions vehicle and that the purchasers had all desired to buy clean emissions vehicles. It was also well understood that the defeat device undermined that desire, that the clean emissions feature had subjective value to individual purchasers (who, after all, had paid for a package of features in which a clean emissions engine was included), and that the clean emissions feature had market value (or it would not have been part of the Respondents’ advertising).

[22] The question that needed an answer, however, was not whether there was value in this feature. Rather, it was whether that value could ever be measured.

[23] Much of the Appellants’ case, both below and in appeal, is dedicated to the former, easy-to-answer question. From a commonsense point of view, the motion judge perceived the case as something akin to a scrambled egg, with the crucial question being how to measure the yolk. Everyone understands that a scrambled egg contains yolk together with the rest of the egg; that is the easy-to-answer question. The difficult question, and the one for which the motion judge needed to see a plausible methodology, is whether and how the amount of yolk in any one forkful can be measured. Pointing out, as the Appellant and his experts do, that there was yolk in the egg to begin with and so the yolk must still be there now, does not address that question.

[24] Having said that, Appellants’ counsel points out that Mr. Stockton did identify at least two methodologies for distinguishing the value of the clean emissions feature from the other features of the subject vehicles: the “Diminished Value Proxy Method” and the “Direct Cost Based Bundle Overpayment Method”. Either of those approaches to measuring the economic loss caused by the defeat device could potentially satisfy the methodology criterion built into the some-basis-in-fact requirement, if they are shown to be plausible. As indicated, the motion judge found that neither methodology was plausible, as neither could “stand up to a lay reader”.

[25] The diminished value approach takes into account the post-revelation drop in market price for the subject vehicles. From there it works backward to determine the amount of extra premium each purchaser would have paid prior to the revelation.

[26] The Respondents argued, and the motion judge accepted, that the diminished value method was inherently faulty as it did not distinguish between a price reduction caused by the defeat device itself and the overall “brand reduction” caused by the revelation of the defeat device and the market’s general loss of confidence in the Respondents’ automobiles. After all, it stands to reason that as owners who sold their vehicles prior to the revelation, the post-revelation brand reduction would not apply to them.

[27] In countering this point, Appellant counsel draw from expert evidence presented in U.S. litigation by economics expert Timothy Breshnahan, as follows:

If consumers place an economic value on the level of NOx emissions of the at-issue engines, the market value of at-issue TDI vehicles will decline after the market becomes aware of the Defeat Software and the associated excess emissions. This decline in value can be used to assess the proportion of the purchase price or lease payments that was related to the emissions and, therefore, the proportion of the purchase price or lease payment that can be claimed as economic damages. I measure the excess diminution in value of at-issue TDI vehicles using a difference-in-difference regression based on actual transaction prices of vehicles before and after the disclosure of the Defeat Software but before the announcement of the Fix.

[28] The “difference in difference” regression, as Dr. Breshnahan put it, represents an analysis of the difference between the market price of the subject diesel-powered vehicles and that of other vehicles manufactured by the Respondents which were not marketed as clean emission vehicles and thus were not fitted with a defeat device. This latter group would have suffered the brand reduction in market price subsequent to the revelation of the defeat device, but would not have suffered the loss of any premium initially paid for the clean emissions feature. According to Dr. Breshnahan, the difference in the two different price reductions is a proxy for the diminished value to those purchasers who bought vehicles in which a defeat device was installed.

[29] Turning to the direct cost bundle methodology, this approach is described by Appellant’s expert Mr. Stockton as one that separates out the component attributes of the subject vehicles’ market value. It then applies hedonic regression analysis to estimate the amount that vehicle purchasers pay for each of those individual components. As explained in Mr. Stockton’s evidence, this method evaluates the portion of the vehicle price specifically attributable to the clean emissions feature.

[30] In Mr. Stockton’s affidavit, he explains that the hedonic regression examines the vehicle price changes over time – comparing periods when the Respondents marketed their vehicles as clean diesel with periods when they did not. The methodology would then have to account for other features added to the vehicles over time (e.g., added torque, improved fuel economy, etc.), after which any remaining unexplained premium for the subject vehicles would represent a measure of the clean emissions premium.

[31] Both of these methodologies are geared toward determining the premium which purchasers would have paid for the clean emissions feature in the first place. But then, the proposed class members face the additional hurdle of having sold their vehicles prior to the revelation of the defeat device to the market. The post-revelation diminution in value was thus not accounted for in the price at which they disposed of their automobiles. Intuitively, one would surmise that if these purchasers paid a premium for clean emissions at the point of purchase, they recouped that premium at the point of sale.

[32] Respondents’ counsel submits that this does not present a problem in calculating damages for the proposed class. In fact, the motion judge in his reasons appears to have accepted that Mr. Stockton’s evidence provided a methodologically sound way of resolving this issue based on the depreciation of the value of the clean emissions feature of the vehicles over time. According to the Appellant’s expert evidence, this depreciation would result in the proposed class members non-recovery on resale or trade-in all of the original overpayment relating to this feature.

b) The request for more evidence

[33] The motion judge opined, at para 57 of his reasons, that while this explanation might be acceptable, it left unresolved the starting point – i.e. quantification of the initial premium paid on purchase:

Had the plaintiff been able to establish the required starting point — some evidence that would isolate the value of the clean diesel feature — then Mr. Stockton’s “excess depreciation” and related analyses, and in particular, his “off the lot” depreciation analysis, if based on this starting point, could arguably have established the existence of some economic loss. And, once this was established, one could then go on to add ‘indirect damages’ such as the time value of money, financing costs, insurance and taxes. [Emphasis added.]

[34] This, then, circles back to the question of whether one or both of the approaches to quantification contained in the expert evidence – the diminished value approach or the direct cost bundle approach – provides the requisite plausible methodology. In considering this question, it is to be kept in mind, once again, that, “The requirement that there be an evidentiary foundation -- or some basis in fact -- to support the certification criteria does not include a preliminary merits test and should not involve an assessment of the merits”: 2038724 Ontario Ltd. v. Quizno’s Canada Restaurant Corp. (2009), 96 OR (3d) 252, at para 74 (Div Ct).

[35] In other words, the issue is merely whether the experts present a plausible – indeed, even a barely plausible – methodology; not whether it will ultimately work. It is by now well established that “a certification motion is not the time for finely calibrated assessments of the expert opinions”: Darmar Farms Inc. v. Syngenta Canada Inc., 2021 ONSC 6411, at para 90.

[36] Counsel for the Appellant submit that the motion judge did more than seek a methodology from the expert opinions in evidence; he assessed it and found it wanting. That, they contend, was beyond what is permitted at this stage of proceedings. In fact, it is precisely what previous cases have advised certification courts not to do: “courts including the highest courts in the land have frowned upon motions judges determining at a certification motion the reliability and utility of an expert’s methodology as part of a certification motion”: Mancinelli v. Royal Bank of Canada, 2020 ONSC 1646, at para 95.

[37] Having reviewed the expert evidence, the motion judge concluded, at para 48, that the record lacked “some evidence (consisting of evidence grounded in facts) that the affected owners or lessees paid extra for a clean diesel or low emissions benefit that was promised but not provided.” He reached this conclusion after studying the methodological descriptions provided by the experts and reviewed above.

[38] At that point, the motion judge clearly was dubious about the methodology proposed by the Appellant’s experts; he did not think it would work. He was particularly dubious about the Appellant’s expert’s focus on the post-revelation price drop for the Respondents’ vehicles, which was done, as the motion judge stated at para 45 of his reasons, “without regard for brand effects or regulatory uncertainty or any other post-disclosure market factors.”

[39] As the motion judge pointed out at para 43, this theory was put forward by Appellant’s expert even after the motion was adjourned and the Appellant was asked “to file an additional Expert Report that actually isolates and quantifies the alleged price premium of the Clean Diesel feature…” What the motion judge wanted was some demonstration that the proposed methodology could really do what it purported to do – i.e. provide a measurement of loss. He went so far as to ask Appellant’s counsel if they wanted yet another adjournment to work up the evidence that he sought, but Appellant’s counsel declined the invitation. As a result, no convincing demonstration was forthcoming.

[40] In seeking this additional information, the motion judge was not illogical; one might well want to see a concrete example of how the methodology will work before concluding that a counter-intuitive economic approach is plausible. But this level of probing went one step over the line drawn around certification motions.

[41] This Court has made it clear that while a party moving for certification must show that some harm has been suffered by the putative class, that harm need not be quantified at this stage. The Appellant’s burden was to establish “some basis in fact for finding that any compensable loss at all had been suffered by [them…[T]hey]…were not required to prove actual loss”: Maginnis, supra, at para 30 (Div Ct).

[42] Accordingly, no tangible demonstration of the proposed methodology was required. “It is not necessary at the certification stage that the methodology establish the actual loss to the class…”: Pro-Sys, supra, at para 115. The methodology need only meet a bare level of plausibility; its actual merit is to be tested at trial.

[43] Again, one can comprehend the motion judge’s frustration with the evidence before him. There were in the record a couple of alternative methodologies proposed by the Appellant’s experts, and so there was, literally, some basis in fact. But it was doubtful to the motion judge that the methodology would work. His error was in translating that skepticism into a conclusion, at para 54, that the Appellant had “provided no evidence of economic loss”.

[44] Given the low threshold reflected in the “some basis in fact” standard for certification, this error was palpable and overriding.

VII. Disposition

[45] The appeal is allowed. There is a plausible methodology for measuring compensable loss on a class-wide basis.

[46] Under the circumstances, and considering that the motion judge has already given the parties two chances to convince him of their case on damages, it is preferable for a new judge to hear the balance of the motion afresh. The matter is therefore remitted to the Superior Court of Justice for a new judge to be assigned to re-hear the balance of the certification issues.

[47] The parties have agreed on costs to the successful party – i.e. to the Appellant – in the all-inclusive amount of $30,000.

Morgan J.

I agree McWatt A.C.J.S.C.J.

I agree Molloy J.

Released: October 4, 2022