TRIAL BY SAMPLE: A POST-GAME, LOCKER ROOM CHAT EXPLORING THE MCADAMS V. MONIER TRIAL: A ROUNDTABLE With Trial Counsel Jeffrey Cereghino and William Stern

Moderated by Kimberly Kralowec; Edited by Asim Bhansali

The case of McAdams v. Monier went to trial in 2012 as one of the few class action cases under either the Unfair Competition Law or Consumers Legal Remedies Act to reach trial in the past few years. Following an eight-week trial, the jury found for plaintiffs, but returned a damages verdict of $7.4 million rather than the $250 million that the plaintiff sought. The trial court subsequently entered judgment for the defendant despite the verdict, finding the methodology used by plaintiff’s expert to be speculative.

This judgment came after a decade of litigation. Following an original filing in 2003, the case reached the Court of Appeal on certification issues. The Court of Appeal reversed the trial court’s decision not to certify the class, but the defendant appealed to the California Supreme Court. Following its decision in Tobacco II, the California Supreme Court remanded this case to the Court of Appeal to review in view of that decision. The Court of Appeal again decided that class certification was appropriate, and a trial followed.

What follows is a roundtable discussion among trial counsel for plaintiff, Jeffrey Cereghino of Ram, Olson, Cereghino & Kopczynski, and trial counsel for the defendant, William Stern, of Morrison & Foerster. This written roundtable forum has been edited by Asim Bhansali of Keker & Van Nest LLP. The original roundtable was presented at the 2013 Golden State Institute and moderated by Kimberly Kralowec of The Kralowec Law Group.

Below are more specifics about our illustrious panelists:

• Jeffrey B. Cereghino is a Partner at Ram, Olson, Cereghino & Kopczynski, LLP, located in San Francisco, California. His practice area primarily focuses upon complex litigation and class actions. He represents consumers in class action matters as well as property owners in complex construction and product liability actions. He has been or is a class counsel on numerous cases.
• William Stern is a partner in Morrison & Foerster’s Litigation Department, resident in the firm’s San Francisco office. He specializes in the defense of consumer class actions in a number of industries. Mr. Stern is the author of Business & Professions Code Section 17200 Practice (West Pub/TRG Press 2013) and the principal author of the 2004 voter initiative Proposition 64, which amended California’s unfair competition law (Cal. Bus. & Prof. Code § 17200 et seq.). William Stern is recommended as a leading lawyer by Legal 500 US 2013, Best Lawyers In America 2012 and Super Lawyers 2012.
• Kimberly A. Kralowec is the principal of The Kralowec Law Group in San Francisco, California. Ms. Kralowec handles class action litigation in the areas of antitrust, wage and hour, and consumer fraud. She served as lead appellate counsel for the class in Brinker Restaurant Corp. v. Superior Court, 53 Cal.4th 1004 (2012), and received a 2013 CLAY Award in recognition of her work. Since 2003, Ms. Kralowec has been the author of The UCL Practitioner (http://www.uclpractitioner.com), a weblog on the UCL, the CLRA, and class action practice in California and the Ninth Circuit.

[Page 60]

Moderator: Can you describe, at a high level, the key considerations in deciding whether or not to try a plaintiffs’ class action under the UCL or CLRA, or both?

Cereghino: I can actually answer that in the context of McAdams. From the plaintiff’s perspective, every case begins with a liability assessment. In McAdams, we knew the product was defective; we had clear evidence the defendant knew the product’s failure mechanism; and we had irrefutable evidence the defendant concealed material facts from everyone. Under those circumstances, we considered trial our only option, because we did not believe the class or the Court would accept any of the settlement offers. So the key consideration in trying the case was it was our only option.

That said, trial was where we wanted to be, even though we knew the defense team was very capable and committed. After nine years of litigation, two appeals, two summary judgment motions, a decertification motion, countless other motions and assorted skirmishes, we were finally in a courtroom. Two unanimous favorable appellate decisions helped frame our case, and gave us a solid roadmap to structure the trial. The trick was getting to court to follow that map.

In terms of whether to proceed at trial under the UCL or CLRA, choosing one or the other was really never on the table in McAdams. Both statutes offer unique remedies to the class, and electing one over the other would have deprived the class. Plus the evidence presentation was substantially identical between our UCL and CRLA case, so there was no efficiency benefit in choosing one or the other.

Moderator: What are the circumstances under which you would advise a defendant to take a class to trial in California court, assuming a class is certified, in some form?

Stern: A defendant does not get to decide. A defendant either settles or goes to trial.

In this case, we could not settle. We faced a determined and veteran group of trial lawyers. They were confident, they had made a sizable commitment and personal investment, and they were not afraid to "roll the dice." But their settlement demand was aggressive, and was pretty much our "worst case scenario" after trial. That made the decision to go to trial easy.

The overwhelming majority of defendants settle once a class is certified, many of them at a premium above their true value. Economists and risk managers will tell you that all litigation entails risks. But what sets class actions apart from other litigation is the asymmetric nature of the risk. Any litigant can lose, but losing for a class action defendant is far worse than for the plaintiff. Why? The whole purpose of class actions is to aggregate claims. That means that a mistake gets amplified. For example, Mr. McAdams’ individual claim of $4,000 was amplified by 125,000 similarly situated claimants—to a half billion dollars.

So, once a class is certified, "the defendant will be under great pressure to settle even if the merits of the case are slight." In re Rhone-Poulenc Rorer, Inc., 51 F.3d 1293, 1298-99 (7 th Cir.1995). This is especially true if the company is publicly traded. It also explains why there are very few jury trials in class action cases and few lawyers, even among experienced counsel, who have tried a class action to verdict.

[Page 61]

Moderator: Can you describe the McAdams case from a plaintiff’s perspective, including the "proviso" about exposure that the Court of Appeal required?

Cereghino: McAdams began with a homeowner who noticed his red roof tiles were eroding to bare concrete. He had purchased the product directly from the defendant, and had relied upon their advertising that included language such as "permanent color," lifetime roof and the like. He contacted the defendant and inquired about what was happening to the roof tiles. He was essentially told that he could pay for a roof re-coat. He ultimately made contact with a large group of homeowners with similar problems, who received the same response.

The case was filed in 2003. Discovery revealed the defendant, Monier, knew the tiles would erode, and never told anyone, not even internally, except for two senior executives. So there was a common issue, a common omission to the entire class. The trial court was not persuaded and denied class certification.

We appealed and, the Court of Appeal reversed the trial court by unanimous decision. The class was certified. Among the issues argued by Monier was the absence of a common representation and reliance. The Court of Appeal—the Third District—recognized that building products are typically marketed indirectly, through roofers, contractors and builders. That marketing paradigm helped fashion their opinion.

Monier appealed to state Supreme Court, which swept this case into the Tobacco II net, such that it was remanded after the decision in Tobacco II. After vigorous oral argument on remand from the Supreme Court, the Court of Appeal again unanimously ordered the class certified.

However, this time, the Court added the "proviso." The proviso set out class membership criteria. Class members had to be exposed to a statement "along the lines of, permanent color, life-time, or maintenance-free." McAdams v. Monier, Inc. , 182 Cal. App. 4th 174 (2010). From our perspective, the Court again recognized the indirect way building products are marketed and thus articulated a broad definition of the class.

Interpreting the "proviso" was the primary issue that framed this case, and how we collectively approached the trial.

Moderator: Was losing class certification a win in disguise for the defendant?

Stern: It did not seem that way when the decision came out. We moved twice to decertify. Prior to trial we moved to decertify and lost. So, at that time, we certainly didn’t think it was a blessing in disguise. We had filed a simultaneous motion for summary judgment but we felt that decertification was the more likely result, if we were going to notch a win short of trial. However, we faced a Court of Appeal decision that had reversed the earlier denial of class certification, and ordered the class certified on remand. It would have taken a brave (or defiant) judge to order the class decertified in light of that published appellate ruling.

After plaintiff rested, we moved again to decertify. And yes, in hindsight, I suppose that is a blessing in disguise. It now means that the Third District could affirm on the separate ground that, quite apart from the merits ruling, and in light of the evidence presented by plaintiff, the case should have been decertified and can be affirmed on that basis.

[Page 62]

Moderator: Can you provide additional background key issues from the defensive perspective?

Stern: The key to unlocking the case for the defendant, and probably for both sides, was the "proviso." It was perhaps the only "silver lining" in what was otherwise a defiantly pro-plaintiff class certification decision by the appellate court. The problem is, what did the Court of Appeal mean? The parties and the judge wrestled with that enigma every day, through every witness and every argument.

On the one hand, if it meant (as defendant contended) that every class member had to prove he or she was "exposed" to an actionable representation and was damaged "as a result," that is incompatible with a class action. Indeed, plaintiff argued exactly that. Plaintiff argued that this cannot have been what the Court of Appeal meant: why would the appellate court reverse and order a class to be certified on remand, only to erect a standard that is impossible to satisfy?

We felt this was a false dilemma. The "proviso" wasn’t inherently impossible to prove. In fact, the appellate court gave examples of how it could be proven, e.g., a uniform brochure or advertisement that every homeowner would have seen. The problem for plaintiff was that there was no such evidence. Therein lay the problem. Plaintiff had to find another way to prove "exposure" and classwide reliance and injury.

Moderator: Can you briefly explain what happened in the trial phase, from the plaintiff’s perspective, including the evidence used to prove exposure as required by the class certification proviso?

Cereghino: We initially proposed a bifurcated trial plan, which the trial court declined to adopt. The case was in Placer County. The assigned judge elected to bring in a retired judge from Southern California, because of the limited civil trial capacity in Placer County and the defense trial estimate of six months. The trial lasted about seven weeks. The defense’s original four month estimate for their case in chief ultimately became four days.

Plaintiff’s witnesses consisted of the class representative, the sample group of sixteen, several other class members, four former defendant employees, and five experts. The primary evidence to prove exposure came from the sample group. Each witness was randomly selected. Their home was inspected to ensure they had the product. Then their deposition was taken. The statistician was only used to validate the mathematics. It was up to the jury to decide, for each sample witness, whether that witness met the proviso. Once that determination was made, then it was statistics 101 to determine the number of class members represented statistically by that witness.

Moderator: Can you describe how class certification, including the proviso, was presented to the jury?

Cereghino: There is truly a finite number of ways to determine class membership. The defense contended that each potential class member needed to testify at trial. Given the potential class size of over 120,000 (and trial counsel’s life expectancy), we didn’t see ourselves trying this case for the next twenty years. The trial court agreed. So that leaves either a bifurcated trial, with a liability phase and a damages phase or process, or using statistical methodologies. Because the trial court rejected bifurcation, we were left with using statistical methodologies.

[Page 63]

The first step in any statistical sampling is determining the potential class size. In this case, no senior Monier executive knew how much product was sold during the class period, and records were very limited. The next problem was identifying the product on potential class members’ roofs. Monier sold two different products during the class period, and another manufacturer had a similar tile profile. So any random survey, without actually confirming the product was on a participant’s home was fraught with error. We ultimately randomly sampled 22 potential class members, inspecting their homes to ensure they had the product, and then deposing them so the defense had full access to cross-examination, unlike blind survey methodologies. Then the 16 who we believed were class members testified, and the jury made the final call.

Moderator: What was the key to the defendant being able to undo the jury verdict?

Stern: The plaintiff’s entire case, stem to stern, was built on statistics. At trial, we wound up fighting a "phantom" adversary, a fictional plaintiff who was a compilation of the elements of many—exposure (liability) drawn from one homeowner, reliance from another, damages from a third—but who resembled no actual, flesh-and-blood homeowner. Plaintiff hired a formidable expert, a retired professor of statistics at CalTech. But this expert admitted his assignment was merely to "size" the class, i.e., he was retained to investigate "exposure of homeowners to statements about Monier tile roofs of a certain kind."

Class membership is one thing; proof of liability is something else altogether. There was no evidence of a classwide statement for purposes of establishing liability—that was not even on the expert’s radar. Nor could he mass-produce through statistics a classwide case of reliance, materiality, or absence of the statute of limitations.

Ultimately, the trial court in its post-verdict ruling accepted the arguments we made at the Kelly/Evidence Code § 402 hearing: (i) The process used to recruit sampled homeowners was slanted, and raised concerns about bias; (ii) there was an extremely small sample, which meant that errors were distorted; and (iii) "The particular methodology Plaintiff designed for this case is inconsistent with long-established principles regarding the use of surveys and statistics and is without any legal authority."

Moderator: Was this just a case of an inappropriate methodology, or is the proof of the kind required by the Court of Appeal’s proviso inherently impossible?

Stern: As I said, the "proviso" was not inherently impossible to prove. The Court of Appeal suggested that it could be proven, for example, by a uniform brochure or advertisement that every homeowner would have seen. But when discovery showed that this didn’t happen, plaintiff had to find another method. He had to invent evidence. And to do that, he turned to statistical extrapolations. It was bad evidence—really just hocus pocus.

Moderator: Did Sargon Enterprises, Inc. v. University of Southern California, 55 Cal.4th 747 (2012) have an effect on how this case ultimately turned out?

Stern: The result would have been the same without Sargon. Sargon was handed down while we were still in trial. But while the court cited Sargon in its post-trial ruling, Sargon broke no new ground as far as our issues were concerned. It merely confirmed that the court has a "gatekeeper" function when it comes to scientific evidence.

[Page 64]

Cereghino: I agree with Will, that it really did not impact this case.

Moderator: Regardless of whether you agree with the trial court’s decision, is there anything plaintiffs might have done with the survey that would have improved its chances on review?

Cereghino: As I indicated above, we really had no viable choice but to proceed with statistical sampling, in view of the decision not to bifurcate liability and damages. One can always argue that more putative class members should have been sampled, but frankly doubling the size would not have foreclosed defense arguments, and merely moved the confidence intervals slightly higher.

Moderator: Do you think the proviso could have been avoided altogether by presenting class certification arguments in a different way?

Cereghino: Our case was essentially an omission case, with a single material omission made to the entire class. Regardless of that predominating common question, we faced the typical difficulties in trying any product class case: class membership and liability, class damages, and defenses such as statute of limitations or product misuse to name a few.

That said, this case began as a clear omissions case as I indicated. Thus, given the ruling in Tobacco II, I don’t think any certification argument could have avoided the proviso. The appellate court made the proviso very broad, which in our view was a clear recognition of trying to apply the CRLA and UCL as legislatively intended, in a scenario where class members had the product but exposure was naturally limited to mostly third party statements (even where those third-party statements repeated the defendant’s marketing statements).

Moderator: How would this case have proceeded differently in federal court?

Stern: In all likelihood, it probably would never have gone to trial. Plaintiff’s statistical evidence in this case went far beyond what any federal court had ever allowed, even before Wal-Mart Stores, Inc. v. Dukes, 131 S. Ct. 2541 (2011). Here, plaintiff was planning to prove everything—class size, class membership, liability, reliance, causation, materiality, damages, and absence of a limitations bar—through statistical sampling. Most federal courts would have denied class certification, or entered summary judgment long before trial, or would have weeded this out in the pretrial conference or a Daubert hearing.

Moderator: Do you have any predictions for appeal?

Stern: The losing party following a lengthy jury trial always has an uphill battle arguing for reversal. So, plaintiff has his work cut out, especially given that the Court of Appeal could affirm on the ground that the trial court should have decertified the class. A wild card could be how the California Supreme Court decides Duran v. U.S. Bank N.A., Case No. S200923, a case that tests the extent to which statistical sampling may be used to prove liability in a wage-and-hour class action. The California Supreme Court is slated to decide Duran by June 2014.

Cereghino: I think the Duran decision will have a significant impact. I also believe the Court of Appeal will give the jury verdict substantial weight, and it will recognize a liability verdict is not dependent on whether a statistician testifies.

Moderator: Moving away from McAdams for the last question, what are the top two UCL, CLRA and class action cases or issues to watch for 2014?

[Page 65]

Stern: That is easy. The first is arbitration, and the extent to which the California Supreme Court will follow the United State Supreme Court’s decision in Concepcion and start enforcing consumer arbitration agreements. The case that will test this is Sanchez v. Valencia Holding Co., No. S199119 (review granted, March 12, 2012). In the employment area, the comparable case is Iskanian v. CLS Transportation of Los Angeles, LLC, No. S204032 (review granted, Sept. 19, 2012).

The second, of course, is Duran, which we discussed above, the state-court counterpart to Wal-Mart v. Dukes. Duran will test the extent to which statistical evidence can be used to prove liability in a class action case in California state court.

Cereghino: I agree with Will that Duran is a key case to watch. My second case to watch would be this one.

[Page 66]