May 31, 2019

Confusion over “some basis in fact” rolls on in British Columbia Court of Appeal’s RoRo decision

Certification is a vital step in every class action.  In order for a class action to be certified, the proposed representative plaintiff must show “some basis in fact” to believe that the certification requirements are met. These requirements include that there are common issues of fact or law and that a class action would be the preferable procedure for resolving those common issues. The Supreme Court of Canada was clear in its decision in Pro-Sys Consultants Ltd v Microsoft Corporation that the some basis in fact standard is less onerous than a balance of probabilities standard. However, how that standard is to be applied remains a source of great difficulty for courts.

In cases where the plaintiff’s certification motion relies heavily on expert evidence to establish the existence of common issues, the some basis in fact standard has proved particularly challenging to apply. While the Supreme Court of Canada has cautioned that certification should not become a battle of the experts, there remains significant litigation over what the quality of expert evidence must be in order for the some basis in fact standard to be met. The recent decision of the British Columbia Court of Appeal in Ewert v Nippon Yusen Kabushiki Kaisha highlights how courts have been grappling with this issue.

At issue in that case was a proposed price-fixing class action against marine shippers who transported cars and other vehicles overseas to Canada using roll-on/roll-off (RoRo) vessels. The plaintiff alleged that there was a conspiracy between marine shippers using RoRo vessels to raise prices, which had the effect of increasing the price of vehicles that were later purchased in Canada. The plaintiff brought an application in British Columbia to certify the proceeding as a class action on behalf of all British Columbia residents who during the class period of February 1, 1997 to December 31, 2012 purchased Vehicle Carrier Services from the defendants, or purchased or leased a new vehicle in British Columbia transported by RoRo.

As in many price-fixing class actions, a central issue on the certification motion was whether there was a methodology the plaintiff could show that loss had been suffered on a class-wide basis by all purchasers. This is frequently the central issue at the certification motion in price-fixing class actions. Where a loss cannot be established on a class-wide basis, courts often hold that either there are not sufficient common issues for the class proceeding to be certified and that a class proceeding would not be the preferable procedure.

This question of class-wide loss in turn routinely depends on expert opinion evidence, typically from economists. It is routine for plaintiffs to deliver expert economic evidence setting out models explaining how loss can be established on a class-wide basis and provide a means of estimating that loss. Defendants in turn deliver opinions from experts who opine that loss cannot be established on a class-wide basis and the plaintiffs’ experts’ methodologies are not workable.

In the present case, at first instance, the British Columbia Supreme Court accepted the regression model put forward by the plaintiff’s experts to establish class-wide loss as a plausible methodology. However, the motions judge refused to certify the proceeding as a class action, holding that there was no evidence that the data necessary to implement the plaintiff’s expert’s model was available. In so holding, the motions judge relied on the statement of Justice Rothstein in the Supreme Court of Canada’s decision in Pro-Sys where he held that, “[t]here must be some evidence of the availability of the data to which the methodology is to be applied.” Here, the motions judge held there was no such evidence and dismissed the certification motion. 

The plaintiff appealed the decision. In a decision released on May 29, 2019, the British Columbia Court of Appeal allowed the appeal in part and certified the proceeding as a class action.

The British Columbia Court of Appeal reiterated that the some basis in fact standard is a low bar. The Court of Appeal expressly contrasted the “some basis in fact” with there being no basis in fact, suggesting that certification will only be denied where there is absolutely no evidence on a particular element of certification.

As it pertains to expert evidence of class-wide loss, at paragraph 104 of its decision, the British Columbia Court of Appeal set the bar very low for what the plaintiff must show:

It is required that a plaintiff lead some evidence that there is a plausible and realistic methodology to establish loss on a class-wide basis, but where the methodology consists of an econometric model, it is not necessary to build the model or identify with precision what information will be used to populate the model, as long as there is some evidence that information will be available to do so.


On the facts of this case, the Court of Appeal held that that standard was met. The plaintiff’s expert had opined that “It should be possible to obtain a number of documents as well as a significant amount of data on the pricing and costing of Vehicle Carrier Services from the defendants themselves (contracts or records of sales), at the time of pre-trial discovery.” The Court held that this evidence was sufficient to meet the some basis in fact standard. It was not necessary for the expert to identify the precise data that would be used to populate the model or to specify exactly what the source of that data would be. Rather, it was sufficient for the expert to describe in general terms what type of data would be used, and to suggest that it may be available from a particular source.

The British Columbia Court of Appeal’s decision confirms that plaintiffs face a low bar as to the quality of the evidence they must lead at a certification hearing. However, there are good reasons that the evidentiary bar for certification should not be too low. 

While courts have confirmed that the certification process is not meant to address the substantive merits of the case, the certification motion is clearly meant to ensure that the case can meaningfully move forward as a class action without becoming unmanageable.  The central requirement for a case proceeding as a class action is that there are meaningful common issues whose resolution will impact all or nearly all members of the class. This should mean requiring plaintiffs to establish on a balance of probabilities that there are common issues. However, that has not been the standard applied by Canadian courts: to the contrary, the Supreme Court of Canada specifically confirmed as recently as 2014 in Pro-Sys that the “some basis in fact” standard is lower than the balance of probabilities. This is unfortunate.

If a court cannot confidently say that it is more likely than not that class members have issues that can be resolved on a common basis and that will substantially advance the litigation, then a class proceeding has a high risk of degrading into an expensive morass of individual issues. Where a certification judge cannot conclude that common issues exist on a balance of probabilities, how can that judge expect a trial judge to later adjudicate on those issues on a common basis at trial? And how can the certification judge expect the parties to lead evidence on those common issues, when the court cannot even confidently say that such issues are more likely common across the class than not? 

In price-fixing cases in particular, courts should be able to conclude on a balance of probabilities that the loss as a result of the alleged conspiracy is in fact a common issue across all class members in order for the case to be certified. If the court cannot be satisfied of this, a class action is not an appropriate mechanism by which the case can proceed, as the case will break down into individualized inquiries as to whether particular purchasers suffered losses. This should mean that plaintiffs should be able to persuade the court that the expect evidence actually shows class-wide loss, not that it merely could do so at a later time. That requirement does not necessarily mean estimating the amount of the loss suffered by class members, but it might mean populating a model sufficiently to show that all or nearly all class members have in fact suffered a loss. Unfortunately, this has not been the standard that courts have set in these cases.

Courts have been resistant to placing too high a burden on plaintiffs in establishing the certification requirements, noting that class certification is intended to take place at an early stage in the proceedings when broad discovery has not yet occurred. Consequently, courts have been concerned about whether it would be unfair to require plaintiffs to prove their case at such an early stage. This ignores two important points. 

First, while it is true that there is no automatic right pre-certification discovery of all relevant documents, plaintiffs still have significant tools to seek access to information relevant to the certification motion. Tools available include resorting to cross-examinations on affidavits, conducting examinations witnesses on pending motions, and reliance on courts’ broad powers to control the conduct of class proceedings to order the production of relevant documents.

Second, as described above, absolving plaintiffs of the obligation to lead evidence that makes it more likely than not that loss can be established on a class-wide basis does not ultimately make anything easier for courts or plaintiffs. Rather, it just kicks the commonality question down the line for the parties and the common issues trial judge to grapple with, resulting in many cases being certified as class actions where in fact there are no meaningful common issues. This is not desirable: only in cases where it is more likely than not that a common issues judge will be able confidently give a single answer to a question on a class-wide should be certified.