For many years, it was next to impossible to certify a price-fixing class action in Canada. Today, certification is so common that refusals to certify are the rare exception. Ewert v Nippon Yusen Kabushiki Kaisha, 2017 BCSC 2357 [Ewert], is such an exception. While the threshold for certification in Canada is low, the court emphasized that certification was not a “‘file, smile and certify’ exercise.” It refused to certify the case owing to the plaintiff’s failure to demonstrate the availability of key data inputs for its econometric methodology.

The plaintiff’s allegations in Ewert followed a typical pattern for price-fixing cases. The defendants operated roll-on/roll-off vessels used to transport vehicles and heavy equipment to Canada. The plaintiff claimed that the defendants had fixed shipping prices, resulting in higher costs for the Canadian class members who purchased the imported vehicles. Several defendants had pleaded guilty to price-fixing charges in the United States.

At the certification hearing, the plaintiff’s economist set out a methodology by which he proposed to prove that harm was common to all class members. The defendants’ economist responded by arguing that the plaintiff’s model was fundamentally flawed in several respects. In particular, he stated that the proposed methodology was not credible or plausible given: (i) the model lacks the required structural stability, as the alleged conspiracy spanned 16 years, with different market characteristics; (ii) prices are privately negotiated, rendering it impossible to estimate a single standard price; and (iii) since only a quarter of cars sold in British Columbia during the proposed class period were imported from overseas, anything less than a perfectly competitive market would require individual analysis to establish a pass-through.

The plaintiff’s expert replied that: (i) the model can be adjusted for structural instability; (ii) the model relies on prices actually paid by manufacturers, not common pricing; and (iii) the model assumes the market is “close to perfectly competitive” (this point was reiterated in the plaintiff’s reply evidence)—a critical assumption the plaintiff’s expert believed to be reasonable.

These replies proved sufficient for the court to conclude that the plaintiff’s methodology had a realistic prospect of demonstrating harm on a class-wide basis at trial. While the court was skeptical that the plaintiff’s methodology would withstand the defendants’ criticisms at trial, it held that further inquiry into the relative merits of each side’s arguments at certification would descend into an impermissible “battle of the experts.”

The plaintiff was less fortunate when the court examined the data required by his expert’s methodology. His expert proposed to use (i) public information; and (ii) private information that the defendants would make available during discovery. Although the expert testified that public information was a necessary input for his methodology, he did not adequately explain how he would obtain that information or what that information would reveal.

The court noted that plaintiffs must lead evidence as to the availability of the data required by their methodology. To the extent the plaintiff’s expert relies on publically available data, the expert must specifically identify that data so that the court can ascertain that the data exists, that it will likely be available, and how it functions in the proposed method. In this case, the court held that even though the plaintiff’s expert identified some data sources, it was not clear from his evidence that the identified sources alone were sufficient for the proposed model. The court stopped short of holding that the plaintiff must also have obtained the data prior to certification, as that determination was unnecessary for its reasoning.

In so holding, the court emphasized it was not requiring that the model must be developed for certification, that documents must be obtained prior to certification, or weighing the competing expert opinions.

Although not strictly necessary, given its refusal to certify the case generally, the court also refused to certify sub-classes based on a lack of evidence. In each instance, the plaintiff baldly asserted in argument or in evidence that the general methodology could apply to the specific sub-classes. The court held that this was inadequate. For example, in his reply evidence, the plaintiff’s expert asserted that his methodology developed for the automobile market could also apply to the “high and heavy” market (being trucks, buses, agricultural and construction equipment). The court held that such evidence is not the proper subject of reply evidence (“it is something he ought to have addressed in his original report”), and in any event that the plaintiff had not shown that the methodology was workable for this market given its different characteristics.

Defendants should take heart that courts may more carefully scrutinize plaintiffs’ evidence at certification. Plaintiffs in future cases may be held to account for better explanations of the availability and utility of their methodology’s data inputs, as well as how the methodology can work for all proposed class members. We can expect more argument at certification on just how much explanation plaintiffs must now provide.