While working as a business banking officer for U.S. Bank National Association, Chad Penza made $5 million to $10 million in sales every quarter, making him one of the most successful salespeople in the company. But it wasn’t easy to get there. On weekdays, Penza regularly arrived at the office by 6 a.m. and didn’t leave until 6:30 or 7 p.m. He also worked most Saturdays and Sundays, calling prospects from 7 a.m. until noon. Penza routinely clocked more than 60 hours of work a week.

In late 2001, Penza was identified as a member of a class of 260 business banking officers (BBOs) whom U.S. Bank classified as outside sales personnel exempt from overtime laws. In Duran v. U.S. Bank, the plaintiffs allege that Penza and other BBOs should have been classified as nonexempt employees who were entitled to earn overtime pay.

What ensued was a hard-fought war of attrition that lasted the better part of the next decade, with both sides spending hundreds of thousands of dollars in attorneys’ fees and expert witness costs. At each stage of the litigation, U.S. Bank protested trial procedures that hampered its defense, including the trial judge’s decision to use a sample of 21 class members to determine the damages sustained on a class-wide basis.

Despite U.S. Bank’s protests, a 2009 bench trial resulted in a $15 million verdict for the plaintiffs. But it turned out to be a pyrrhic victory. The California appellate court overturned not only the verdict, but also the decision to certify a class in the first place, determining that the procedures the trial court used were “fatally flawed” and deprived U.S. Bank of due process.

In doing so, the court delivered a stern rebuke to trial judges for the increasing use of statistical evidence to resolve class actions. To hear the plaintiffs’ attorney tell it, the decision sounds the death knell for class action overtime misclassification claims.

“Employees will be deprived of their right to pursue back wages, and employers will be immunized for their illegal conduct,” says Edward Wynne, class counsel in Duran v. U.S. Bank.

Margin of Error

Misclassification class actions are a veritable cottage industry in California. In the past year, plaintiffs have obtained multimillion-dollar settlements or judgments against some companies for allegedly misclassifying workers as exempt (see “Misclassification Judgments”).

In many such cases, each plaintiff’s lost wages are too small to attract plaintiffs’ lawyers. But in the aggregate, the claims add up. The plaintiffs bar is using increasingly sophisticated tactics to shape these cases to their advantage, including an aggressive campaign to use statistical evidence. The California Supreme Court aided these efforts by encouraging trial judges to adopt “innovative” procedures to manage class actions in the 2004 case Sav-On v. Superior Court.

“The use of statistical evidence was not common until the past few years,” says H. Scott Leviant, a partner at Spiro Moss. “In overtime misclassification cases, it’s becoming a common method to analyze the time employees spend on exempt and nonexempt tasks.”

It was often labor intensive to determine whether a class of workers should have been exempt. Typically, both sides would submit declarations from employees explaining how they spend their time. Relying on Sav-On, courts had begun to make determinations based on a randomly selected subset of class members.

For plaintiffs, such procedures are advantageous because they translate into shorter trials and allow them to avoid problems of proof with individual class members’ claims. From the employers’ side, statistical evidence is problematic because it deprives them of the opportunity to present defenses to individual workers’ claims.

In Duran, for example, the court considered the testimony of a randomly selected sample of BBOs to determine whether people with that job title spent more than half of their work time on outside sales. After determining that, on average, workers should have been nonexempt, the court again used a sample to decide how much the group was owed in back pay.

There was variation in the testimony, with some workers testifying that they were required to be in the office and others stating that they spent most of their time on outside sales. There also was disparity in the testimony about hours of work. Some rarely worked more than 40 hours a week and others logged overtime hours daily. Relying on expert testimony from two statisticians, the trial judge decided that the evidence supported a substantial verdict, despite the fact that the sample had, by one estimate, a 43.3 percent margin of error.

Certification Stage

The appellate court flatly rejected the idea that such evidence was sufficiently reliable and chided the court for failing to allow U.S. Bank to present defenses to individual claims in a case where there was wide variation among class members.

“The trial court exceeded acceptable due process parameters by limiting the presentation of evidence of liability to the testifying BBOs only,” the court wrote. “Fundamental due process issues are implicated not only by the unprecedented and inconsistent use of statistical procedures in the liability and damages phases, but also by the manner in which USB was hobbled in its ability to prove its affirmative defense.”

Many employment litigators say Duran will put an end to the use of statistical evidence in the vast majority of class action trials. Perhaps more significantly, it also will put a damper on the use of statistics at the class-certification stage.

“If judges take this seriously, plaintiffs will have to come up with a way to try these cases that does not involve sampling,” says Thomas Kaufman, a partner at Sheppard Mullin Richter & Hampton. “This allows defendants to bring up these issues with the judge sooner. If there is not a practicable way to try the case without sampling, then the class won’t be certified.”

In many cases, the certification stage is the key for the employer—once a class is certified, the risks of going forward become too great, and plaintiffs have significant leverage to force a settlement.

The likely effect is that fewer plaintiffs will bring misclassification cases. While the Fair Labor Standards Act and California law provide for the recovery of attorneys fees for a successful plaintiff, Wynne, for his part, describes it as “pure fantasy” that plaintiffs lawyers will be willing to pursue individual claims. “It is not economically viable,” he says.

End of the Road

However, Duran does not completely put a stop to the use of statistics in employment cases. If plaintiffs are challenging a policy that is uniformly applied to a class of workers, sampling may still be appropriate. For example, if an employer has a policy of rounding time on employee punch cards to the nearest quarter hour, a random sample of all punch cards could be used to determine whether that policy deprived workers of wages.

“What you won’t see is certification of classes that involve any individual analysis of the employees’ duties,” Kaufman says.

Duran petitioned the state Supreme Court for review, and Wynne believes that the decision should be overturned. “The court shouldn’t permit the wholesale destruction of the class device,” he says.