Class Certification in the Apple iPhone Antitrust Litigation Denied Due to Economist’s “Unjustifiable Assumptions”

May 2022

By Abiel Garcia  
Kesselman Brantly and Stockinger

On March 29, 2022, the Judge Rogers from the Northern District of California denied Consumer Plaintiffs’ motion for class certification without prejudice and then granted in part and denied in part Apple’s Daubert motion to exclude expert testimony.  As a refresher given the amount of on-going Apple litigation, this specific Apple litigation relates to Consumer Plaintiffs’ claims that Apple charges developers on the App Store supracompetitive commissions, which are then passed to consumers in the form of increased prices for app downloads or subscriptions.  Order Denying Plaintiffs’ Motion for Class Certification Without Prejudice, In re Apple iPhone Antitrust Litigation, case no. 11-cv-6714-YGR (March 29, 2022 N.D. Cal.). 

Daubert Motion

First, Apple challenged the overarching model used by the Consumer Plaintiffs’ economist, Professor McFadden.  Apple argued that Prof. McFadden’s model was not based on principles of hypothesis generation, scrupulous study design, meticulous data collection, and objective interpretation, but rather, the model was designed in order to achieve a desired, preconceived outcome.  The court disagreed.  In a brief paragraph, the court stated that while Apple’s experts disagreed with the assessment and quantification of an impact, they failed to dispute the fundamentals of the methodology.  The court denied the motion on said grounds.

Next, Apple challenged Prof. McFadden’s opinion on market definition, pointing to the fact that Prof. McFadden ignored the two-sidedness of the App Store.  In another brief paragraph, the court disagreed and found that the considerations undertaken were sound in rendering an expert opinion.

Turning to the actual damages model, the court undertook a three-step analysis to review Professor McFadden’s model.  In reviewing the benchmark analysis, which attempts to identify a “but-for” commission rate, the court disagreed with the expert’s claim that they used a benchmark analysis to find a “but-for” commission rate.  Rather, the court found examples of how the Prof. McFadden cherry-picked data in order to find the model’s commission rate.  The court reviewed various data points that were relevant that should have been considered in Prof. McFadden’s analysis but were conveniently omitted.  The court then granted Apple’s motion with respect to Prof. McFadden’s opinion on the “but-for” commission rate due to a lack of foundation.

For step two of the analysis, the court turned to the estimation of the app and in-app prices that consumers would have paid in the but-for world.  The court found multiple problems in the model, including changing definitions of what constitutes an “uninjured account,” computational errors throughout the model, and finally, Prof. McFadden’s concession that the individual damages calculation should have used a percentage calculation rather than a fixed-dollar calculation method.  Focusing on the calculation method, Apple offered evidence that showed Professor McFadden’s model led to absurd results, including some accounts having larger damages than actual spend and negative pricing. 

The court then walked through a variety of additional issues raised by Apple: Negative But-For Pricing, Focal Pricing and Pricing Tiers, Sample Size and Robustness, Free Apps Analysis, and the Identification of Unharmed Class Members.  In reviewing these topics, the court did find some additional problems, which led the court to ultimately find that the Professor McFadden’s current model was unreliable for determining class wide impact or damages.  The court  allowed Consumer Plaintiffs leave to resolve the issues.

Class Certification

After reviewing Prof. McFadden’s model and excluding it, the court turned to the Consumer Plaintiffs’ class certification motion under Fed. R. Civ. P. 23(a) and 23(b)(3).  In reviewing the typical requirements under Fed. R. Civ. P. 23(a), the court quickly found numerosity, four common questions of fact or law, and that the Consumer Plaintiffs satisfied the adequacy and typicality requirements under Fed. R. Civ. P. 23(a). 

Turning to the Fed. R. Civ. P. 23(b)(3), the Consumer Plaintiffs put forth, and Apple challenged, Prof. McFadden’s three-step methodology as a viable method for demonstrating class wide injury based on common proof.  Incorporating the previous Daubert analysis, the court dove deeper into the fact that the model found a substantial number of class members that have suffered no injury (about 14.6% of accounts).  Due to the model being flawed and ultimately excluded on Daubert, the court concluded that the Consumer Plaintiffs cannot meet their predominance burden because the model cannot reliably demonstrate which members, and how many, were injured, as common proof of class wide impact. This would lead to individual issues predominating with respect to injury.

On class wide damages, the Consumer Plaintiffs offered Prof. McFadden’s model to prove that damages are measurable.  Just as with the predominance analysis, the court found that the model failed to prove measurable damages.  Stepping outside of the flaws found in the Daubert analysis, the court discussed how the Consumer Plaintiffs indicated that they did not intend to run the model until after trial.  Rather, the Consumer Plaintiffs would tell the jury about a damages range between $7 billion and $10 billion and then use the model afterwards to determine distribution amounts.  While the court noted that this decision is novel, it believed it would result in damages that would be too speculative, without any adequate underlying rational for the decision.  Thus, again the court found the model to be unreliable in assessing class-wide damages and denied Consumer Plaintiffs’ motion for class certification.  Importantly, the court noted that while the economists’ models and assumptions are flawed, the court “anticipates that the deficiencies can be addressed” and thus, denied the motion without prejudice.