Justice Carol Corrigan, California Supreme Court
Justice Carol Corrigan, California Supreme Court (Jason Doiy / The Recorder)

SAN FRANCISCO — California’s highest court sounded ready Tuesday to throw out a $15 million overtime class action that employers describe as a poster child for the misuse of statistical evidence.

But the court seemed reluctant to devise broad new rules limiting the use of statistics and sampling, which should come as a relief to plaintiffs lawyers anxious about another blow to employment class actions.

“You aren’t taking the position that statistical analysis can never be admitted, are you?” Justice Carol Corrigan asked Carothers DiSante & Freudenberger partner Timothy Freudenberger.

“Do we need to answer the abstract question” about statistical evidence and due process, Chief Justice Tani Cantil-Sakauye asked.

Freudenberger, who represents employer U.S. Bank, told them both the answer is no, but Justice Goodwin Liu did not sound reassured. He suggested that even an ostensibly narrow ruling tied to the unusual circumstances in Duran v. U.S. Bank could knock out a broad swath of cases.

“Your position is that class certification is never appropriate for this kind of case,” he told Freudenberger. “You’re trying to tiptoe around that by taking a narrow position, but I don’t see it, counsel.”

The Duran dispute centers on 260 business bank officers—agents who sell loans and other products to small businesses—who weren’t paid overtime. Rather than hold numerous individual trials, Alameda County Superior Court Judge Robert Freedman certified a class and took testimony from 20 representative officers. All testified that they did most of their work from inside a bank branch, which would render them eligible for extra pay.

U.S. Bank says the sample was skewed, and that Freedman ignored declarations from some 75 other officers who said they did most of their work outside the bank. The First District Court of Appeal ruled in 2012 that Freedman improperly relied on the sampled testimony to find liability, then compounded the problem by extrapolating damages that had a margin of error of 43 percent.

Tuesday’s hearing drew a large crowd of employment lawyers, and even Freedman’s successor in Alameda County’s complex litigation department, Judge Wynne Carvill. The zealous advocacy that seemed to give Freedman headaches was on display, with the justices having to caution both Freudenberger and class counsel Edward Wynne from interrupting the court at various points of the hour-long hearing.

The justices quickly latched onto the issue of sample size, wanting to know how Freedman arrived at the number 20.

“It was completely arbitrary,” Freudenberger asserted. “It was for convenience and case management.”

“Let’s assume the court misstepped there,” Justice Kathryn Mickle Werdegar said. Would a broader sample size solve the problem?

No, Freudenberger argued, because the trial proved that individualized issues predominate over common issues. “The court still hasn’t heard the testimony of over 90 percent of the class,” he said. “We were prevented from doing that. And yet, according to this judgment, they’re entitled to about $14 million.”

Corrigan sounded sympathetic. “At some point during the trial of this case, it became apparent that the individual questions were swamping the common questions,” she posited, later describing Freedman’s approach as, “Don’t confuse me with details now that it looks a bit messy midtrial.”

Class counsel Wynne, meanwhile, argued that Freedman chose the sample size only after receiving expert testimony that assured him of its statistical validity. “The court did not pick it out of thin air,” he said.

Justice Marvin Baxter, usually one of the quieter justices, fired the first question Tuesday and remained active throughout the hearing. Much of his focus was on class treatment of a case that turns on the location of individual employees. “The problem I see is how do you establish common proof of that?” he asked.

Wynne said the officers testified not only about their own personal experience, but the bank’s policy of encouraging them to stay in-house and work the phones. Freedman also heard testimony from bank managers, and properly found the bank’s 75 declarations lacking credibility because of coercion. “The defense is entitled to due process, not unlimited process,” he said.

But most of the justices sounded inclined to send the case back to superior court, probably on the basis that the sample in the case was faulty. “We don’t have to resolve the question of whether always and everywhere, a case of this nature can’t be handled as a class action,” Corrigan said. “We don’t need to make a sweeping statement about statistical sampling.”

Liu seemed concerned, though, that even a “narrow” ruling would touch a wide range of misclassification cases. “Taken to its logical extreme, your position is it’s always going to come down to individualized proof,” he told Freudenberger, “no matter what the company’s policies were.”

However sampling might be handled, there will always be some margin of error, Liu said. And live testimony is hardly infallible. “Do they keep records?” Liu asked Freudenberger about his 75 declarants. “Do they tabulate all the time they spend inside and outside the office?”

“It’s the best evidence we’ve got,” Freudenberger replied.

Rudy, Exelrod, Zieff & Lowe partner Steven Zieff, one of the pioneers of statistical evidence and cocounsel to amicus curiae California Employment Lawyers Association, said after the hearing it was clear the justices had concerns about the rigor of proceedings in Duran. But he said he did not foresee the court overturning recent precedents on sampling such as Sav-On Drug Stores v. Superior Court or his own Bell v. Farmers Insurance. In many misclassification cases, he said, “the class action remedy is the only way to vindicate the rights of workers.”

Contact the reporter at sgraham@alm.com.