The latest development in the long-running case of Brazil v. Dole Packaged Foods, LLC, No. 12-CV-01831-LHK (N.D. Cal. Nov. 6, 2014), came when Judge Lucy Koh largely granted the defendant’s motion to decertify a class that she had just certified months earlier. In a blog posting on June 26, 2014, I discussed that earlier certification decision and noted several weaknesses in the plaintiff’s expert’s analysis. The defendant used those weaknesses (and others) to decertify the Rule 23(b)(3) class. And Brazil points to a problem with these class actions that many in the defense bar have been raising for quite some time—how can plaintiffs “prove” an injury from purchasing a low-cost product in a dynamic and competitive marketplace for food?

As with many food labeling cases, this plaintiff alleges that “all natural” statements on 10 Dole products are misleading. Those products contain ascorbic acid (i.e., vitamin C) and citric acid, which the plaintiff contends are synthetic ingredients, making the products improperly labeled.

On May 30, 2014, Judge Koh certified classes under Rule 23(b)(2) and (b)(3). As to the (b)(3) damages class, the court originally relied on a damages expert’s assertions that he would use a regression analysis to accurately calculate how much the class members overpaid for the products based on the “all natural” labeling. That expert, Dr. Oral Capps, has been a favorite of plaintiffs’ counsel in these class actions. Dr. Capps had not finished his regression analysis at the time of that earlier decision, contending that Dole had failed to provide some necessary information in discovery.

After Dr. Capps provided his final analysis and Dole deposed him, however, Dole moved to decertify the class. Discovery showed that labels on 9 of the 10 products did change during the class period. That was important because a regression analysis typically would analyze a product’s price with the challenge statement compared to the price without the challenge statement. In theory, an expert can isolate movements in the price attributable to the challenged labeling. Because the products’ labels did not change, however, Dr. Capps had to use a hedonic regression analysis. In that type of analysis, the expert compares the challenged products to what he or she concludes are “comparable products” to try to isolate the effect of the challenged label statement on the price. Dole identified six flaws that it contended made Dr. Capps’ analysis inadequate. Judge Koh rejected several of those, so I will focus on the flaws she agreed rendered his analysis unacceptable.

First, Dr. Capps failed to show how his regression model controlled for other variables affecting price, such as advertising. In fact, Dr. Capps admitted that he did not attempt to control for advertising. That was a fatal flaw. “How else could the Court know whether any price premium on the challenged products was due to Dole’s ‘All Natural Fruit’ labeling claim rather than to its advertising expenditures?”

Dr. Capps also relied on false or untested assumptions about his comparable products, such as failing to verify whether those products made an “all natural” labeling claim. In fact, he never explained why he did not verify those labels or why the plaintiff did not. Thus, he couldn’t accurately explain whether he was comparing Dole prices to competing products with or without “all natural” label statements. “So if the model is unsure whether the non-Dole products actually made an ‘all natural’ labeling claim, then how can the Court know whether the price premium the model generates is based on Dole’s labeling claim rather than on some other factor? Put simply, it cannot.”

Another problem was Dr. Capps’ failure to account for the fact that some products have multiple labeling statements. For example, a product may state it is “all natural” and also that it has “no sugar added.” Dr. Capps never explained how he could isolate the effect of one labeling statement from the other, but this case only alleged injury the “all natural” statements.

Things got worse for Dr. Capps when he failed to account for packaging differences among products. For example, he would treat a competing product consisting of one 16 oz. can the same as a Dole product containing four 4 oz. cups. Consumers may very well pay a premium for that type of packaging convenience (e.g., for putting in children’s school lunches), but Dr. Capps could not control for that type of pricing variable.

All of these shortcomings with Dr. Capps’ analysis meant that his opinions could not survive scrutiny under Comcast Corp. v. Behrend, 133 S. Ct. 1426, 1433 (2013). And that focus on Comcast was central to Judge Koh’s conclusions. Judge Koh also was unimpressed that Dr. Capps said in a declaration in a different food labeling matter that it was impossible to provide a regression analysis because the labels in that case (as in this matter) had not changed during the class period.

As she has done in other food labeling cases, however, Judge Koh refused to apply a stringent ascertainability requirement for class certification. It was sufficient that the plaintiff defined the class based on the objective criterion of purchasing the identified products within the class period. She distinguished this case from others that lacked ascertainability because those cases involved dozens of different products with changing labels. Moreover, only certain types of those food products in other cases qualified based on their flavor. Here, in contrast, potential class members would only need to recall that they purchased one of the challenged Dole products, not the flavor, the label contents, etc.

Important take-away points for practitioners include:

    • Regression analysis is not a silver bullet for class plaintiffs. Work with your defense expert to identify the variables the plaintiff’s expert cannot control for.
    • As with any expert, test the opposing expert’s methodology. Is he relying on unfounded or false assumptions the way Dr. Capps incorrectly assumed that competing products did not have “all natural” labels? Did he fail to verify easily-verifiable facts?
    • Always check the opposing expert’s testimonial history in declarations or depositions. As we saw in this case, it was insurmountable when Dr. Capps opined in an earlier case that the regression model was impossible but took the opposite stance here.
    • Not stated in this decision but always important is to examine your expert’s testimonial history. You do not want to be on the opposite side of having to explain why your expert is taking inconsistent positions in similar cases.