In re ConAgra Foods, Inc., No. CV 11–05379, 2015 WL 1062756 (C.D. Cal. Feb. 23, 2015)
I’ve tried to limit discussion of many issues in this huge opinion (40,000 words without footnotes). Plaintiffs, consumers from eleven different states, alleged that they bought Wesson Oils because ConAgra deceptively marketed them as “100% Natural” on the front label, even though they were made with GMOs. They alleged claims for violation of state consumer protection laws, breach of express warranty, breach of the implied warranty of merchantability, and unjust enrichment.
Given Dukes, the court applied Daubert to the experts at the class stage. Plaintiffs’ expert Colin Weir was Vice President of Economics and Technology, Inc. (“ETI”), a research and consulting firm specializing in economics, statistics, regulation, and public policy. He opined that it would be possible to determine damages “with a reasonable degree of specificity, certainty, and accuracy, attributable to ConAgra’s conduct of placing the ‘100% Natural’ claim on the label of every bottle of Wesson Oil by applying the scientifically valid economic methodology of hedonic regression to common, class-wide, aggregate historical retail price and attribute data for Wesson Oil and competing cooking oils to calculate a class wide Price Premium, and then multiplying that Price Premium by the total retail amounts all Class Members paid for Wesson Oil to yield total class-wide damages.” This would require conjoint analysis. The court previously found that he failed to provide a reliable damages model including the relevant variables or data, even if the methodologies he described were capable of calculating damges in theory. ConAgra argued that he still failed to identify data in his possesssion to which hedonic regression could be applied, or relevant variables. However, the court found he’d remedied the earlier deficiencies by preparing a preliminary regression model to measure price/price premium. He analyzed twenty product attributes, using data from various spreadsheets and reports reflecting historical price, cost, profit and attribute information for Wesson Oils and competitor brands. He acknowledged that more geographically and temporally specific data could be obtained and used for more refined regressions.
ConAgra argued that the data he had wasn’t useful in performing hedonic regression analysis, and that the data were incomplete and didn’t control for geography, sales channels, or retailers. Its expert opined that Weir’s proposed methodology was flawed. The court found that Weir’s testimony was sufficiently reliable; more geographic information was available and while Weir might not obtain the same price premium results once it was included, that just meant that his testimony might not be helpful to plaintiffs, not that it was unreliable.
ConAgra also sought to exclude the declaration of Elizabeth Howlett, who opined on a consumer survey about GMOs. The survey described GMOs in a way that “alarmed and confused survey respondents” (including “bacteria,” “virus,” or “toxic to certain insects”), and it had a small sample size and high nonresponse rate. While most challenges to a survey go to weight rather than admissibility, the issues here were so severe that this survey didn’t satisfy Daubert. Given that 51% of respondents responded incorrectly to a question designed to ensure they understood the definition of the GMO process, and that Howlett didn’t participate in designing and administering the survey, her opinion that it properly defined the GMO process was insufficient. The sample size was a problem because the court didn’t know the sampling method used and because the sample didn’t approximate the relevant characteristics of the customer population. “[G]iven Howlett’s inability to validate that the survey was reliably designed and administered, such concerns reasonably suggest that the survey’s methodology may be flawed.”
However, Howlett’s academic training and practical experience qualified her to testify to the calculation of damages using a conjoint analysis. Conjoint analysis would be used to assess the percentage of the “100% Natural” claim that was attributable to the absence of GMOs as opposed to other “non-natural” aspects of the Wesson Oils. Howlett sufficiently explained why she chose the attributes she did, and had sufficient experience with conjoint analysis for her testimony to be deemed reliable.
ConAgra also objected to the testimony of nine named plaintiffs, eight of whom said that they’d “very interested” in buying Wesson Oils labeled “100% Natural” if they did not contain GMOs, and that they “might consider” or “will consider” purchasing Wesson Oils in the future if they continue to contain GMOs and ConAgra stops labeling them “100% Natural.” ConAgra argued that these were shams rather than true witness testimony. The fact that attorneys may have prepared the declarations was standard; declarants are under penalty of perjury and can refuse to sign any document that is incorrect or inconsistent with their recollections or beliefs. Their statements that they’d consider buying Wesson oil again weren’t shams; the declarations clarified earlier claims that plaintiffs had stopped buying Wesson oil. “While the declarations may negatively affect plaintiffs’ ability to prove materiality, causation, and/or reliance, this does not compel the conclusion that they are false or directly contradictory of prior testimony.”
Turning to the motion for class certification, plaintiffs sought classes for California, Colorado, Florida, Illinois, Indiana, Nebraska, New York, Ohio, Oregon, South Dakota, and Texas consumers.
ConAgra argued that the named plaintiffs lacked standing for want of injury, given that after they sued, they continued to purchase cooking oils and other products that were labeled “natural” but contained non-organic GMO ingredients, and they didn’t know what price they paid for Wesson despite alleging a “premium.” The court disagreed. Though their subsequent purchases, and their willingness to buy Wesson oil again even if it still had GMOs, might seriously undercut their claim that “100% Natural” was material, that didn’t deprive them of standing. Nor did the fact that plaintiffs couldn’t recall the exact price they paid deprived them of standing, given the potential for a workable damages methodology.
Ascertainability: ConAgra argued that the classes were not ascertainable because there was no way to determine the identity of consumers who purchased its products. The court disagreed that self-certification was impossible; ConAgra’s position would “effectively prohibit class actions involving low priced consumer goods—the very type of claims that would not be filed individually—thereby upending ‘[t]he policy at the very core of the class action mechanism.’” Because every putative class member has been exposed to the alleged misrepresentation, the fact that some class members may have not been injured by the “100% Natural” claim does not render the class unascertainable.
Numerosity: of course. Commonality: all the consumers were exposed to the challenged label. Typicality: ConAgra argued that the named plaintiffs’ claims weren’t typical because the “100% Natural” label was not a significant factor driving purchases of Wesson Oil. Its survey showed no statistically significant difference between the purchasing decisions of survey respondents shown a “100% Natural” label and those who saw a label without the phrase. Only 5–6 percent of respondents who saw the “100% Natural” label mentioned “natural” ingredients when describing why they would or would not buy a Wesson Oil product, and identifying the factors that were important to them when purchasing cooking oil. But typicality asks whether the named plaintiffs’ claims arise from the same course of conduct as the class members’ claims and whether they’re subject to unique defenses; ConAgra’s arguments didn’t show atypicality.
Adequacy: yep.
Plaintiffs sought to certify proposed classes separately for injunctive relief and damages. On injunctive relief, plaintiffs sought to show that they could represent the class because they’d consider buying Wesson again. They argued that their evidence wasn’t speculative, merely conditional: if ConAgra changed its conduct, they’d buy Wesson again. The court didn’t agree. Considering a future purchase wasn’t concrete enough to show a sufficient likelihood of being harmed in the future
Damages: for 23(b)(3) certification, plaintiffs had to show that common issues predominated. The key issues here are reliance and causation. First, the court had to determine whether a classwide inference of reliance and causation was available upon a showing of falsity and materiality under each state’s laws. California: yes, for consumer protection claims and express warranty claims (privity was required for breach of implied warranty, and that couldn’t be shown here on a class basis).
Colorado: A classwide inference of reliance and causation could be made for material misrepresentations under Colorado consumer protection law and warranty claims, but not unjust enrichment, which would require individual proof of causation.
Florida: Similar: state consumer protection claims could be adjudicated on a classwide basis, but not unjust enrichment. “Even if plaintiffs can prove that the ‘100% Natural’ was false, it does not necessarily follow that ConAgra’s retention of the full purchase price would be inequitable with respect to a consumer who did not notice or did not rely on the ‘100% Natural’ claim.”
Illinois: While reliance wasn’t required, proximate causation was a requirement under Illinois consumer protection law; individual issues would almost always be present, but wouldn’t necessarily predominate. Where the representation being challenged was made to all putative class members, “Illinois courts have concluded that causation is susceptible of classwide proof and that individualized inquiries concerning causation do not predominate if plaintiffs are able to adduce sufficient evidence that the representation was material.” Unjust enrichment would stand or fall with the statutory claim, so was subject to the same analysis.
Indiana: Here there were only warranty and unjust enrichment claims. Uniform misrepresentations allowed common issues to predominate for unjust enrichment purposes, whereas the express warranty claims required either privity or a showing of reliance, and there was no showing that Indiana accepted reliance as capable of classwide proof. By contrast, breach of implied warranty of merchantability required proximate cause, and plaintiffs argued that their price premium theory showed proximate cause of harm for every class member. The court agreed.
Nebraska: These were also warranty and unjust enrichment claims, and the court reached the same results.
New York: Plaintiffs alleged violation of the New York GBL, breach of express warranty, and unjust enrichment. Proof of reliance and scienter are not elements of a GBL claim; likelihood of misleading a reasonable consumer was the key issue and could be established on a classwide basis. Materiality could be proved classwide for breach of express warranty. But an unjust enrichment class couldn’t be certified because individualized inquiries as to whether “equity and good conscience require restitution” are not susceptible of classwide proof under New York law.
Ohio: Plaintiffs alleged only state consumer protection law violations. In Ohio, a classwide inference of reliance is permitted where defendant’s fraudulent or deceptive conduct is common to all consumers, so common issues could predominate.
Oregon: Plaintiffs alleged violations of state consumer protection law and unjust enrichment. Here again, causation and reliance were susceptible of classwide proof for consumer protection law, and also unjust enrichment could be given class treatment because the putative class members were subjected to “uniform treatment” by the defendant.
South Dakota: Same asserted violations as Oregon, but no South Dakota precedent either way on whether reliance or causation could be proved on a classwide basis for consumer protection law violations. The court predicted that they could be upon a showing of materiality, guided by the “broad, remedial purpose” of South Dakota’s consumer protection law and by the South Dakota Supreme Court’s suggestion that “class certification ‘is favored by courts in questionable cases.’” Also common issues predominated with unjust enrichment because of the defendant’s uniform conduct.
Texas: Same claims, but here only the consumer protection law claims survived; for unjust enrichment, the Texas Supreme Court had held that, even in situations where the price paid by class members to the defendant, is uniform, “individual differences between each class member’s experience” will necessitate individualized inquiries to “determine in whose favor the equities weigh in resolving [class members’] claims.”
Whew. All of this depended on whether materiality could be proved on a classwide basis; the court therefore turned to that issue. Plaintiffs offered various third party surveys to show materiality of “100% Natural,” such as a 2014 report by the Consumer Reports National Research Center, which surveyed a nationally representative sample of consumers and found that 59% look for a “natural” claim when shopping for packaged or processed foods such as Wesson Oils. A 2010 survey also found that 65% of respondents were “somewhat interested” or “very interested” in purchasing natural products and a substantial majority of consumers attested that it was worth paying more for “natural” products. Plaintiffs also cited some of ConAgra’s market research, which purportedly showed that consumers exposed to a “100% Natural” or “Natural” claim on ConAgra product labels generally consider the representation a significant factor in their purchasing decisions, as well as ConAgra internal strategy documents identifying the claim as a favorable one. None of the surveys specifically linked consumers’ understanding of “100% Natural” to whether they thought that a product with that label had no GMO ingredients, but they did tend to show materiality of the claim.
Plaintiffs also cited surveys to show that consumers believe that “Natural” means without GMOs, such as the Consumer Reports survey, in which 64% of respondents believed that a “natural” claim on food products meant that the product contained “no GMOs” or “genetically modified ingredients.” Two other studies by the Hartman Group found that a majority of consumers understood “natural” to mean an “absence of genetically modified foods,” and that “[c]onsumers perceive [GMO] foods as inherently unnatural and worry about adverse health effects” from such foods. A HealthFocus International study also concluded that a substantial majority of consumers associate a “natural” claim with the absence of GMOs. Moreover, plaintiffs submitted evidence that the company received consumer complaints about the “100% Natural” label on Wesson Oils after consumers discovered that they contained GMOs.
This was enough of a showing of materiality for purposes of class certification. Plaintiffs didn’t need to show that every customer would find the claim material or believe it meant no GMOs. Instead, they only needed to show that a reasonable consumer would understand it that way and find it material. Courts have accepted materiality claims from significantly smaller percentages.
ConAgra argued that the claim couldn’t be material because the FDA has refused to designate genetically engineered foods and food ingredients non-natural and has concluded that the presence of GMOs is not a “material fact” that must be disclosed under FDA regulations. But the FDA’s view of genetically engineered foods wasn’t the relevant question, which was what a reasonable consumer would have thought. Plaintiffs’ possible interest in purchasing Wesson again could support an inference of immateriality, but that didn’t show immateriality to a reasonable consumer, especially in light of the survey evidence.
Comcast says that Rule 23(b)(3) is satisfied only if plaintiffs can show that “damages are capable of measurement on a classwide basis,” using a method of proof tied to plaintiffs’ theory of liability. Previously, the court found Weir’s calculations insufficiently specific, since consumers might attribute multiple possible characteristics to a “natural” label, and Weir didn’t isolate the price premium from misleading consumers about GMOs. (Hmm. I don’t see why that matters, since if it’s deceptive to use “natural” when a product has GMOs, then the product shouldn’t use the term and those other meanings would also be unavailable to consumers and unable to bring their share of the price premium. I doubt “GMOs, but otherwise natural” is really an available message for ConAgra.) Here, Weir’s methodology plus Howlett’s conjoint analysis were sufficient in combination. Howlett proposed to use consumer surveys to segregate the percentage of the price premium specifically attributable to a customer’s belief that “100% Natural” means “no GMOs,” and multiplying that times the premium “would necessarily produce a damage figure attributable solely to ConAgra’s alleged misconduct—i.e., misleading consumers to believe that Wesson Oils contain no GMOs by placing a ‘100% Natural’ label on the products.” Criticisms of Howlett’s methods were not dispositive at this stage.
ConAgra next argued that individualized inquiries would be required to figure out how many bottles, what sizes, etc. individual consumers bought. The damages inquiry could account for that. Thus, predominance was present.
Superiority: yep. ConAgra argued that eleven state classes would be unmanageable. The court agreed with plaintiffs that separate classes avoided choice of law concerns; that the laws all fell into consistent patterns; and that the warranty claims were all based on the same statutory test. “Under the various consumer protection statutes, plaintiffs must show, for example, that ConAgra’s conduct is deceptive and misleads reasonable consumers and/or class members.” The surviving unjust enrichment claims all involved “substantially the same question—whether ConAgra received some benefit from plaintiffs that it would be inequitable to allow it to keep in light of its conduct.” So too with the breach of warranty claims.
Rebecca Tushnet 