Liu v. Uber Technologies Inc.

Background

Thomas Liu alleged that Uber’s practice of using customer ratings in driver termination decisions had a racially disparate impact—meaning it harmed a protected racial group more than other groups even without requiring proof of intentional discrimination. The court explained that such a claim requires allegations of: (1) a significant employment disparity affecting the protected group, (2) a specific employment practice, and (3) a causal connection between that practice and the disparity.

Liu alleged that racial bias could affect customer ratings in the rideshare industry and that Uber relied on those ratings when terminating drivers. The court previously found that these allegations plausibly addressed the challenged practice and a possible causal connection, but did not adequately show that the concern about racial bias actually resulted in racially unequal terminations at Uber.

Survey allegations

To address that deficiency, Liu added a survey of Uber drivers. The survey asked drivers to identify their race and, if they had been deactivated, whether low ratings caused the deactivation. Liu alleged that minority drivers were terminated for low ratings at a higher rate than white drivers.

The court had previously found the survey “essentially meaningless” because it focused on drivers who had already been deactivated. That design could show a difference among deactivated drivers without showing a difference across Uber’s entire driver population. The court also found the survey flawed because its use of the term “Latinx” likely caused some Latino respondents to select “other.”

In the third amended complaint, Liu added that counsel sent a follow-up email to respondents who had answered “no” to the initial survey. Of those who responded to the follow-up, 51.7% said they had not been deactivated. The court found that this information made the allegations weaker because it showed that some drivers misunderstood the original survey, making the survey even less useful for supporting an inference of racial disparity.

Court’s reasoning

The court clarified that statistical workplace evidence is not always required at the pleading stage. If an employment policy plainly has a disparate impact on a protected group, the impact may be inferred without additional workplace statistics. But the court found that Liu’s theory was different: although racial bias could affect customer ratings, the allegations did not show that using those ratings would necessarily cause a racial disparity in every case. Liu therefore needed meaningful, non-conclusory facts about the effect of the practice at Uber itself.

The court said the problem was not that Liu’s theory was implausible. Rather, he had not developed adequate factual allegations showing a significant racial disparity in Uber’s employment decisions.

Disposition

The court granted Uber’s motion to dismiss the third amended complaint. Because the court was not convinced that Liu could never state a disparate-impact claim, it allowed him one final opportunity to file an amended complaint within 28 days. The court stated that if Liu did not file an amended complaint by that deadline, dismissal of the third amended complaint would be with prejudice. It also stated that dismissal would be with prejudice if the next complaint failed to state a claim.