The Sixth Circuit Steps Back in Time on Certification of Consumer Classes
The United States Court of Appeals for the Sixth Circuit recently upheld the certification of an Ohio consumer class action in In re: Whirlpool Corporation Front-Loading Washer Products Liability Litigation, 678 F.3d 409 (6th Cir. 2012) (“Glazer”). Glazer is the Sixth Circuit’s second opinion on class certification since the Supreme Court decided Wal-Mart Stores, Inc. v. Dukes, ___ US ___, 131 S.Ct. 2541, 180 L. Ed. 2d 374 (2011). The first opinion, Gooch v Life Investors Ins. Co. of America, 672 F.3d 402 (6th Cir. 2012), involved class certification under Fed. R. Civ. P. 23(b)(2). Glazer was the Sixth Circuit’s first post-Dukes opinion to address certification under Fed. R. Civ. P. 23(b)(3).
For a number of years, culminating in Dukes, federal courts have moved away from deciding class certification based on the allegations in the plaintiff’s complaint and have instead focused on critically analyzing and resolving factual disputes that bear on certification issues, even if those factual disputes involve the merits of the case. In Glazer, the Sixth Circuit returned to an earlier era of class action jurisprudence. The district court addressed only the plaintiffs’ allegations and theories when it certified a class, and the Sixth Circuit, straining to avoid remand, assumed without any real basis that the district court had conducted the “rigorous analysis” that Dukes requires. In doing so, the Sixth Circuit overlooked factual issues relevant to certification, resulting in an opinion upholding certification based simply on the plaintiffs’ allegations.
Glazer is one of several consumer product putative class actions pending against Whirlpool in Multi-District Litigation (“MDL”) in the Northern District of Ohio. The plaintiffs in those cases alleged that all of the high efficiency front-load washing machines Whirlpool sold since 2002 were defective because they were more likely to develop mold growth and resulting odors than regular top-load washing machines. The plaintiffs alleged that the machines fail to completely rinse residue from internal components and remove spent water from the machines after the completion of a wash cycle, which causes mold to grow. Although the district court and Sixth Circuit opinions refer to “various alleged design defects,” they did not identify any specific defect; they simply described what the machines failed to do. The mold growth phenomenon was well publicized by consumer groups, and, over time, Whirlpool undertook to reduce the likelihood of mold growth by redesigning the machines, revising use and care instructions, and selling new cleaning products. The plaintiffs alleged that those efforts were ineffective.
In Glazer, the plaintiffs sought certification of an Ohio class made up of people who purchased the washing machines for personal, family or household use, and not for resale. The district court certified a liability-only class with respect to claims for tortious breach of warranty, negligent design and negligent failure to warn.1 The district court refused to certify the class as to claims under the Ohio Consumer Sales Practices Act.
The District Court Opinion
At the class certification hearing, the district court acknowledged that there were several factual issues relevant to class certification, mostly having to do with “commonality” under Rule 23. In fact, the parties presented the district court with a mountain of depositions, expert reports and exhibits on those and other issues. However, the district court resolved none of the certification-related factual issues. Instead, citing Eisen, it took the position that any evidence that was contrary to the plaintiffs’ allegations on class certification, no matter how strong, went to the merits and could not be considered.
The district court granted certification based solely on the plaintiffs’ allegations and theories, and relied upon no facts, let alone did the court resolve any factual disputes. The district court wrote:
“Plaintiffs Allison and Glazer allege…”
“The defect, the plaintiffs contend…”
“The plaintiffs further allege….”
“The mold problem allegedly persisted …”
“The plaintiffs’ theory is…”
“The first element is common to the class, because, on the plaintiffs’ theory, …”
The district court’s class certification analysis reverted back to what Professor Nagareda has referred to as the “first-generation” of class certification decisions, based on “overreadings of Eisen” where the common questions were essentially defined by the allegations in the plaintiff’s complaint and the court was prohibited from probing the facts behind them, even if those facts were relevant to certification issues. Nagareda, “Class Certification in the Age of Aggregate Proof,” 84 N.Y.U.L. 97, 110 (2009).
Similarly, the district court also refused to analyze the defendant’s argument that the proposed class was overly broad because it included a significant number of class members who had manifested no injuries. Although the district court asked questions during argument about how many class members’ machines actually manifested the malodor problem, the opinion characterized that question as a merits issue, citing Eisen and Daffin.2
The Sixth Circuit Supported the Plaintiffs’ Theories with “Facts” in the Record
In affirming class certification, the Sixth Circuit observed:
Whirlpool contends that the district court improperly relied on Eisen to avoid consideration of the merits of plaintiffs' legal claims, failed to conduct the required "rigorous analysis" of the factual record, and failed to make specific findings to resolve factual disputes before certifying the liability class. We disagree. The district court [***11] closely examined the evidentiary record and conducted the necessary "rigorous analysis" to find that the prerequisites of Rule 23 were met. See Gooch, 672 F.3d at 418 (rejecting a similar argument and concluding that the district court "probed behind the pleadings, considering all of the relevant documents that were in evidence").
Glazer, 678 F.3d at 418. In light of the argument in the district court and the express language of the district court’s opinion, the Sixth Circuit’s conclusory analysis comes up short.
The oral argument in the district court makes clear that it failed to conduct a rigorous analysis. During one exchange, the district court pounced on defendant’s counsel after he simply acknowledged the plaintiff’s allegations on commonality, indicating that the court treated that mere acknowledgement as defendant’s concession that there were common questions:
THE COURT: You seem to just have completely undercut your argument. You had earlier argued that all these changes [in the designs of the machines] were significant, and now you are saying that the Plaintiffs' real claim is, irrespective of any changes made in the machines, that there is one common defect that continues to cause oder [sic] problems without regard to whether you changed it in '05, '06, '07.
So you have characterized their claim as being a common one; that there is something inherent in this design that causes the oder [sic] irrespective of what the consumer uses and irrespective of what changes you have made in the machine.
Trans. at 36. In fact, defense counsel had merely identified some of the individualized questions that would have to be decided in response to Plaintiffs’ claim:
“They will have to prove at trial that each of the design changes Whirlpool made failed to bring the level of malodor down. They will have to prove that for each of the machines, each of the plaintiffs, each of the members of the class, that they complied with the instructions in the owners’ manual.”
Trans. at 35.
Moreover, the district court specifically stated in the hearing that the defendant’s arguments about the lack of common questions based on the numerous design changes went to the merits (Trans. at 51) and, in the court’s written opinion, reaffirmed that it could not make any preliminary inquiry into the merits. Glazer, 2010 U.S. Dist. LEXIS 69254 at *2. Finally, as described supra, in analyzing commonality the court referred only to the plaintiffs’ theories and allegations.
The district court quite clearly treated any factual rebuttal to the plaintiffs’ allegations relating to class certification issues as merits issues that it could not, and did not, analyze. Nonetheless, the Sixth Circuit concluded that it did so by referring to the Sixth Circuit’s prior decision in Gooch, suggesting that its treatment of the district court’s certification analysis reviewed there applied equally to the analysis being reviewed in Glazer.
However, comparison to or reliance on Gooch is not appropriate for characterizing the district court’s analysis in Glazer. In Gooch, the district court, in reciting generic class certification law, included the following statement:
Albeit in a different type of legal claim, in Coleman v. General Motors Acceptance Corp., 196 F.R.D. 315, 318 (M.D. Tenn. 2000),vacated on other grounds, 296 F.3d 443 (6th Cir. 2002), the Court stated that: “The Court takes the allegations of the plaintiff as true and any doubts as to certifying the class should be resolved in the plaintiff’s favor.”
Gooch v. Life Investors Insur. Co. of America, 246 F.R.D. 340, 347 (MD Tenn. 2009). However, the district court went on to say that although “sometimes the issues are plain enough from the pleadings to determine whether the interests of the absent parties are fairly encompassed within the named plaintiff’s claim…sometimes it may be necessary for the court to probe behind the pleadings before coming to rest on the certification question.” Id. The court then said that “the parties had submitted affidavits and exhibits that the court deems sufficient for the limited factual analysis required.” Id. Thus, the district court in Gooch did not take plaintiff’s allegations as true. The judge conducted the “limited factual inquiry” where it was necessary. In fact, the district court described the plaintiff’s deposition testimony that it considered on the issue of the plaintiff’s credibility when evaluating adequacy of representation. As to the issue of commonality, the district court in Goochindicated that the central issue in the case was the interpretation of an insurance policy issued and administered by the defendant which, under Alabama law, was an issue of law for the judge to decide. Thus, no factual analysis was required on the issue of commonality.
The district court in Gooch recited a proposition of law that it did not follow, and the principle claim was largely legal and required no factual analysis. The Sixth Circuit in Gooch emphasized “that the issues at the core of the certification dispute are legal and not factual.” Gooch, 672 F.3d at 417.
In stark contrast, in Glazer the issues at the core of the certification dispute were not legal. There were many factual issues presented to the district court, along with a raft of related evidence and testimony. Given the statements and characterizations in the district court’s opinion, it did not rely on or critically evaluate any of that material. The district court did not conduct a rigorous analysis, and the Sixth Circuit took it upon itself to identify facts that it determined might support the allegations and theories referenced by the district court, as well as other positions the plaintiffs assert to support certification. The Sixth Circuit went well beyond identifying allegations and theories, like the district court did in a single “Background” paragraph. In comparison, the Background portion of the Sixth Circuit opinion is comprised of 15 paragraphs it distilled from the record on appeal.
The district court opinion in Glazer is also in stark contrast to the district court opinion discussed in Young v. Nationwide Mutual Insur. Co., 2012 U.S. App. LEXIS 18625, 2012 Fed. App. 0302P (6th Cir. 2012). In Young, the Sixth Circuit held for the third time that it is harmless error for a lower court to state that it must accept the substantive allegations in the complaint as true when evaluating class certification. However, the district court’s written opinion in Young showed that it conducted a rigorous analysis by analyzing fact issues, and it even conducted a Daubert hearing relating to the plaintiff’s experts’ opinions on class proofs. The lower court’s written opinion in Glazershowed just the opposite.
The Sixth Circuit’s de novo fleshing out of the district court’s less-than-rigorous analysis had two effects. First, it prevented the defendant from challenging the appellate court’s fact finding, except through its subsequent unsuccessful motion for rehearing. Second, the appellate court was able to side-step important factual issues that bore on certification.
Proof of Defect Through Common Evidence
The Sixth Circuit’s overly narrow focus on facts supporting the plaintiffs’ allegations and theories caused it to fail to appropriately address a key issue for class certification, which is “predominance.” The district court found that the plaintiffs satisfied the commonality requirement based solely on the plaintiffs’ allegations and by “Plaintiffs’ theor[ies].” The Sixth Circuit upheld those “findings” based on facts it identified from the record on appeal that supported the plaintiffs’ theories:
Based on the evidentiary record, the district court properly concluded that whether design defects in the [front loading washers] proximately caused mold or mildew to grow and whether Whirlpool adequately warned consumers about the propensity for mold growth are liability issues common to the class…[t]hey will generate common answers likely to drive the resolution of the lawsuit.
Glazer, 678 F.3d at 419. Neither the district court nor the Sixth Circuit meaningfully analyzed the predominance requirement. Both simply referred back to what they stated on the issue of commonality.
The Sixth Circuit failed to address whether defects in the washers that allegedly resulted in mold growth could be established through common evidence. Instead of identifying and then rigorously analyzing product specific and manifestation-related facts to determine whether defects could be established through common evidence, the court ignored them in favor of a conclusory opinion by plaintiff’s expert that in effect presupposed commonality, i.e., differences don’t matter. While this may be a step beyond simply relying on the allegations in the complaint, it is not a very big step.3
Whirlpool presented unrebutted testimony that it manufactured the front-load washers over a long period of time (9 years), and that it sold over 3,000,000 units nation-wide (more than 150,000 in Ohio) consisting of 21 different models whose designs changed over time, in some cases to specifically address the mold issue. Maintenance and care instructions also evolved during this period. The plaintiffs’ response to those facts, which the district court accepted at face value, was the conclusory statement that none of it mattered. Unrebutted testimony also established that all washing machines, regardless of whether they were top or front-load, and regardless of whether the machines were high efficiency, can and do develop mold. Finally, very strong evidence established that less than 5% of consumers experienced a mold issue across all Whirlpool front-load models and designs. The only conflicting evidence on manifestation rates was a single survey that the plaintiffs represented showed that 35-50% of users reported a problem, and two pieces of correspondence, one obvious hearsay, misinterpreting the survey’s results. In response, Whirlpool established that the survey included both top and front-load washers, without breaking the results down between the two, and included machines manufactured by all manufacturers again, without breaking the results down by manufacturer. The district court did not address those “facts,” and did not resolve any factual issue. The Sixth Circuit acknowledged some of the positions Whirlpool advanced based on those facts, but made no effort to quantify the manifestation rate and relied only on plaintiffs’ expert’s opinion that changes did not matter.
A real question existed whether the plaintiffs could establish, on a class wide basis, that all products, models, and designs over 9 years had the same “various alleged design defects,” Glazer, 678 F.3d at 417, when 95% of purchasers never experienced a mold issue and, for the 5% that experienced the problem, it was a problem that could occur in non-defective machines. How will proof that one purchaser’s particular model of washing machine developed a mold problem more readily than it otherwise should have, due to any particular set of circumstances and causes, answer the question whether purchasers of all the other different models with different designs and who have not experienced any mold issue have defective machines, or that those defects caused malodor, under different circumstances? The common questions the Sixth Circuit identified essentially assumed a common defect. Despite the facts Whirlpool presented regarding multiple models and designs and low manifestation rates, the court essentially ignored them in favor of the plaintiffs’ expert’s conclusory opinion that variation doesn’t matter, thus allowing it to treat the case like Daffin where there was a single defect claim in one specific model of a vehicle manufactured during two years based solely on plaintiffs’ allegations. Glazer and Daffin could not be more different from one another.
Following Dukes, courts have conducted thorough factual analyses of whether defects can be proven through common evidence. Those courts deny certification where the cases involve multiple models and product designs, multiple sources or root causes of the manifestation or occurrence that allegedly results, and/or there is an infrequent likelihood of the manifestation or occurrence.
In Cholakyan v. Mercedes-Benz USA, 2012 U.S. Dist. LEXIS 44073 (C.D. Cal. March 28, 2012), the plaintiff consumer complained of interior water damage caused by leaks. The plaintiff sought to certify a California class of current or former owners or lessees of all 2003-2009 E-Class vehicles. The plaintiff alleged that the vehicles had defective “water drainage systems,” and further alleged the defects were unreasonably dangerous because the leaks could cause electrical failures, and that the defendant concealed the water leak defect.
On the issue of common questions, the plaintiff’s “central point” was that the class vehicles have a unitary “water management system” that is uniformly defective. The court did not accept that allegation on its face, but instead it probed the facts behind the contention. As a result, the court rejected the plaintiff’s argument.
Assuming arguendo that class vehicles experience water leaks, and that the leaks have a propensity to cause electrical malfunctions, the crucial question that must be answered is why each class member's vehicle experienced water leaks. Cholakyan's attempt to demonstrate that this question has a common answer for all class vehicles fails for several reasons. First, despite his efforts to identify a "water management system" in the class vehicles, the evidence that has been adduced shows that this so-called "system" is in fact an amalgamation of many different vehicle parts. There is no evidence that these disparate parts are conceptually part of a single system or physically connected to one another in any material way.
Id. at *56. (emphasis added)
Putting aside whether any of these "systems" clog to the extent that they cause electrical malfunctions in the vehicles, Cholakyan has failed to establish that they are connected with one another or that they are designed to work in concert. There is also no evidence that a single design flaw that is common across all of the drains in question is responsible for the alleged water leak defect.
Id. at *58.
The court concluded:
Beyond Cholakyan's inability to identify a single system causing the water leaks, the parties adduce evidence that there is substantial design variation among the class vehicles. Cholakyan's proposed class definition includes owners and lessees of Model 211 vehicles manufactured and sold over six years. As one example of this variation, Cholakyan relies on defects in the purported "sunroof drainage system." Not all the class vehicles even have a sunroof, however, much less a drainage system connected to a sunroof. [footnote omitted] There is also variation in vehicles that have sunroofs, specifically in whether they have drain tubes [*65] that run down the A-pillars and exit through the areas identified in the DTBs. 98 Additionally, the design of the vehicles varied from year to year, further undercutting Cholakyan's evidence of commonality. [footnote omitted] While it is not entirely clear that these variations are material to the functioning of the vehicles' drains, it is Cholakyan's burden to demonstrate commonality, and he has failed convincingly to rebut defendant's evidence regarding the lack of an integrated water management system and regarding design variations among class vehicles.
Id. at *64-5. In addition, the court observed in a footnote that “[c]ausation problems also stand in the way of identifying common issues here because environmental circumstances, use factors, and a vehicle owner’s maintenance habits all contribute to whether or not the vehicle's drains clog.”
Relying in part on Cholakyan, a similar result was reached in Edwards v. Ford Motor Company, 2012 U.S. Dist. LEXIS 81330 (S.D. Cal. June 12, 2012). In Edwards, the plaintiff sought certification of a putative class of current and former California residents who owned 2005 - 2007 Ford Freestyle vehicles with allegedly defective electronic throttle control ("ETC") systems. The plaintiff claimed that this common defect caused sudden, unintended "surging." She claimed that Ford knew about the defective ETC system but failed to disclose the defect to consumers.
Initially, the court noted that because all of the plaintiff’s claims depended on the a defect that causes surging, the fundamental question in the case was “whether the 2005 through 2007 Ford Freestyle models contain a defect that causes the cars to accelerate without corresponding driver input, and, if so, how to define the defect.” Id. at *12.
On that issue, Ford presented unrebutted expert testimony indicating that numerous systems and components in the vehicles besides the ETC system could cause the problem plaintiff described as surging. Although there may have been a common manifestation of a defect – surging -- there was not a common cause of the manifestation. Because the cause of the surging would have to be analyzed for each individual plaintiff, common questions did not predominate.
In Johnson v. Harley-Davidson Motor Company Group LLC, 2012 U.S. Dist. LEXIS 72048 (E.D. Cal. May 23, 2012), the plaintiffs claimed that certain motor cycle engines generated too much heat which resulted in premature engine wear, alleging that all of the engines suffered from a common design defect. The plaintiffs identified the “defect” as a combination of the size of the engines and the fact they were designed to operate on lean fuel mixtures to satisfy emissions regulations. In addition, the plaintiffs alleged that the excessive heat could distract the rider, which raised safety concerns. The plaintiffs sought to certify a class of “[a]ll persons who between October 1, 2006 and the date the class may be certified, purchased in California a motorcycle manufactured by Defendants with an air-cooled 88 or 96 cubic inch V-Twin Cam engine.” Id. at *6.
On the issue of common proof of defect, the plaintiffs’ position was that both parties could simply measure the temperature of the skin of riders to prove or disprove whether the air-cooled V-Twin design causes riders to experience excessive and unsafe skin temperatures. However, the defendant produced evidence that the task would not be so easy.
The defendant noted that the motorcycles at issue were comprised of over 130 different configurations and designs, with varying engine sizes and output and various exhaust system designs, that could affect both engine temperature and the rider’s sensation of the heat. In addition, the defendant identified numerous other factors, such as the rider's sitting position, posture and leg position that could affect the heat experienced by the rider. Based on those facts, the court held that proof of defect could not be established through common evidence.
Johnson relied on In re Hitachi Television Optical Block Cases, 2011 U.S. Dist. LEXIS 109995, 2011 W.L. 4499036 (S.D. Cal. Sept. 27, 2011). Johnson described In re Hitachi as follows:
In Hitachi, the Plaintiffs identified forty-three different model televisions alleged to have the same design defect. 2011 U.S. Dist. LEXIS 109995, 2011 WL 4499036, at *3. These models all used one of seven different Optical Blocks. Id. The Hitachi Plaintiffs' expert stated, similar to Darnell's testimony, that although there are minor differences in the Optical Blocks, there is no material difference amongst them in the common design scheme. 2011 U.S. Dist. LEXIS 109995, [WL] at *3-4. The Hitachi Defendants, like the Defendants here, pointed out that there were numerous and significant distinctions in specific parts of the product. 2011 U.S. Dist. LEXIS 109995, [WL] at *4. The Hitachi Plaintiffs, like the instant Plaintiffs, did not appear to dispute that these differences exist, instead they argued the differences were immaterial. 2011 U.S. Dist. LEXIS 109995, [WL] at *5. The Hitachi court held that:
There is no dispute that the Optical Block is made up of numerous component parts. It is not a single piece of equipment, nor was its design uniform across the different models at issue in this case. Under these circumstances, common issues do not predominate over individual issues on the issue of a design defect.
Johnson at *12-13. Based on the facts and the decision in In re Hitachi, the court in Johnson concluded that there was no common method of proof to show whether there is a design defect. The court also noted that it was telling that there were literally no customer complaints about allegedly excessive heat.
Other courts have made similar rulings in cases involving multiple models or alleged defects, emphasizing the low rates of manifestation of the injury resulting from the alleged defects. In Maloney v. Microsoft Corporation, 2012 U.S. Dist. LEXIS 28676 (D. N.J. March 5, 2012), the court denied certification of a class in a matter involving multiple alleged design defects and low manifestation rates. The plaintiff sought to represent a class of purchasers and owners of certain MP3 players whose LCD inner screens cracked without cracking of the outer skin. The cracks allegedly resulted from 5 alleged design defects, and the manifestation rate of cracking was 1.3% of the MP3 players the defendant sold. The court found that common questions did not predominate on either causation or defect. As to causation the court stated:
To begin, any factual dispute concerning whether causation is capable of proof at trial through common evidence must be determined by the Court. See In re Hydrogen Peroxide, 552 F.3d at 307. This requires the "weighing [of] conflicting expert testimony," and the Court must "[r]esolve expert disputes in order to determine whether [the predominance] requirement has been met." Id. at 323-24. The Court must engage in this analysis even if it overlaps with the merits. Id. at 317-18. However, "[t]he Court's role is not to determine which side will prevail on the merits, but [only] to determine if the putative class could prevail on the merits using common proofs." Marcus v. BMW of N.A., LLC, No. 08-5859, 2010 U.S. Dist. LEXIS 122908, at *30-31 (D.N.J. Nov. 19, 2010).
Id. at *7. The court also stated that:
Plaintiff then points to three pieces of evidence that are purportedly common proof that these alleged design defects caused each proposed class member's injury: (1) proposed class members' LCD screens fractured without external damage to the outer lens covering the LCD screen; (2) these LCD screen fractures were "disproportionately clustered around [*7] four identified internal design defects"; and (3) 30gb model Zunes "were 20 times more likely to crack without external damage than were LCD screens on the later-model 80GB Zune." (Id. 2). These purported proofs, however, fail to establish that any of the five alleged design defects caused class members' injuries because this evidence suffers from what the United States Supreme Court has termed a "failure of inference." Wal-Mart Stores, Inc. v. Dukes, U.S. , 131 S. Ct. 2541, 2555, 180 L. Ed. 2d 374 (2011).
Id. at *6-7.
….. As framed by the Plaintiff, the LCD cracks covered under the proposed class result from a muddled mix of causes and effects. There is no indication that each purported cause led to a uniform result (e.g., an origination point in the same location), which would permit the Court or a jury to draw an inference of a design defect. See Ford Ignition, 194 F.R.D. at 490, n.4 Thus, there is no way to determine which of these [*13] five purported causes or which grouping of these causes led to which individual LCD crack or group of LCD cracks.
Id. at *13-14.
With respect to proof of a defect by common evidence, the court similarly found that common questions would not predominate:
…..Plaintiff' s claim, however, is premised on a design defect in all 30gb Zunes-- the vast majority of which showed no signs of a defect. This "lack of evidence of a common defect amongst the majority of users [is] instructive as to the plaintiff's failure to demonstrate that common factual issues predominate over individual ones." Payne, 2010 U.S. Dist. LEXIS 52808, at *14 (citing Arabian v. Sony Elecs. Inc., No 05-1741, 2007 U.S. Dist. LEXIS 12715, at *39).
Id. at *22. See also, In re Canon Cameras Litigation, 237 F.R.D. 357 (S.D.N.Y. 2010) (Plaintiff could not establish predominance in its claims that purchasers of 13 specific Canon cameras were defective due to various defects. Canon pointed out that less than 0.2% of the cameras in issue malfunctioned for any reason and those few that did, malfunctioned so due to a variety of factors, including customer misuse.) Payne v. Fujifilm USA, 2010 U.S. Dist. LEXIS 52808 (D. N.J. May 28, 2010) (class not certified where manifestation rate for all problems was 4% and less than 2% for specific problem at issue).
Glazer has many of the same attributes that led to denial of certification in these decisions - multiple defects, multiple models and designs, changing use and care instructions, and low rates of manifestation. Thus, the issue of whether the plaintiffs could establish a defect and causation through common evidence, and whether common issues predominated, should have been more thoroughly evaluated in this case. Instead, the district court and Sixth Circuit effectively mischaracterized the facts and claims as “Daffin-like” and misdirected its attention elsewhere.
Certification of a Class of Ohio Purchasers of Front-Load Washers Where a Majority of the Class Has Not Sustained Any Injury
Instead of properly analyzing the fact issues, including lack of manifestation, presented to the court on whether defect and causation could be proven with common evidence, or whether like proof of damages, individual questions would predominate, both the district court and the Sixth Circuit in Glazer instead looked at the issue of manifestation only in the context of whether the class as defined was overly broad because it included class members whose washers had not developed mold or malodor. That focus resulted largely from Whirlpool’s briefing.
In the district court, Whirlpool argued that, as defined, the proposed class was overly broad because it included purchasers whose washers had not manifested any mold problems. The district court rejected that argument, stating that “whether any particular plaintiff has suffered harm is a merits issue not relevant to class certification.” The court cited Eisen and Daffin for the general proposition that the court should not make a merits inquiry in evaluating certification. The district court’s reliance on Daffin was misplaced because Daffin contains no such holding. Daffin simply held that whether the express warranty at issue required manifestation was a merits issue that could not decide in determining whether to certify. Because it could not address that predicate issue, the lack of manifestation could not be a factor in evaluating typicality. The court went on to say that if the district court subsequently determined that the warranty required manifestation, the class could be decertified.
On appeal in Glazer, Whirlpool made the same argument, albeit in a limited way. Its basic point was that “a class comprised primarily of uninjured persons who would not have standing to bring a claim on their own behalf cannot be certified.” Phrased differently, no class may be certified that contains members lacking standing. Several circuit courts have held this to be the law. Arvitt v. Reliastar Ins. Co., 615 F.3d 1023(8th Cir. 2010); Denny v. Deutsche Bank AG, 443 F.3d 253 (2nd Cir. 2006); Mazza. v. American Honda Motor Co., Inc., 666 F.3d 581 (9th Cir. 2011).
The Sixth Circuit rejected that argument referencing an eclectic collection of precedent. First, the court referenced (b)(2) class decisions:
"What is necessary is that the challenged conduct or lack of conduct be premised on a ground that is applicable to the entire class." Gooch, 672 F.3d at 428 (internal quotation marks omitted). Class certification is appropriate "if class members complain of a pattern or practice that is generally applicable to the class as a whole. Even if some class members have not been injured by the challenged practice, a class may nevertheless be appropriate." Id. (quoting Walters v. Reno, 145 F.3d 1032, 1047 (9th Cir. 1998)) (internal quotation marks omitted).
Glazer, 678 F.3d at 420. Walters was a civil rights case challenging, on procedural due process grounds, administrative procedures used by the INS. The class certified below was a Rule 23(b)(2) class. The language quoted in Glazer actually appears in a statement distinguishing a (b)(2) from a (b)(3) class action.
Although common issues must predominate for class certification under Rule 23(b)(3), no such requirement exists under 23(b)(2). It is sufficient if class members complain of a pattern or practice that is generally applicable to the class as a whole. Even if some class members have not been injured by the challenged practice, a class may nevertheless be appropriate.”
Walters, 145 F.3d at 1047. Thus, the applicability of this language to a (b)(3) class requiring that common questions predominate is questionable.
Next, the Sixth Circuit properly did not rely on Daffin for the proposition that “proof of manifestation is not a prerequisite to class certification.” Instead it cited a Ninth Circuit decision, Wolin v. Jaguar Land Rover North America, LLC 617 F.3d 1168 (9th Cir. 2010).4 Wolinindeed contains the quoted language, but it too cites to an earlier decision, Blackie v Barrack, 524 F.2d 891 (9th Cir. 1975), for the proposition.
Blackie was a securities fraud case, and was an appeal from a “conditional” grant of class certification. The defendants contended that the district court engaged in improper speculation when it determined whether common questions existed. The Ninth Circuit’s response was as follows:
Defendants misconceive the showing required to establish a class under Hotel Telephone Charges. We indicated there that the judge may not conditionally certify an improper class on the basis of a speculative possibility that it may later meet the requirements. 500 F.2d at 90. However, neither the possibility that a plaintiff will be unable to prove his allegations, nor the possibility that the later course of the suit might unforeseeably prove the original [**28] decision to certify the class wrong, is a basis for declining to certify a class which apparently satisfies the Rule. The district judge is required by Fed. R. Civ. P. 23 (c)(1) to determine "as soon as practicable after the commencement of an action brought as a class action . . . whether it is to be so maintained." The Court made clear in Eisen IV that that determination does not permit or require a preliminary inquiry into the merits, 417 U.S. at 177-178; thus, the district judge is necessarily bound to some degree of speculation by the uncertain state of the record on which he must rule. An extensive evidentiary showing of the sort requested by defendants is not required. So long as he has sufficient material before him to determine the nature of the allegations, and rule on compliance with the Rule's requirements, and he bases his ruling on that material, his approach cannot be faulted because plaintiffs' proof may fail at trial…..
524 F.2d at 901.
It made some sense in the age of overreadings of Eisen and in the context of addressing class certification on a conditional basis very early in the case, when little discovery has been done, that speculative manifestation rates would not preclude class certification. The court would likely not know whether 90% or 2% of the putative class had actually suffered the complained of failure. However, it makes much less sense today when conditional certification is no longer allowed and where thorough discovery precedes consideration of class certification. Manifestation rates can be thoroughly tested and analyzed in this context, especially when the class period spans a decade or more during which manifestation rates can be more easily verified. Today, the blanket proposition upon which Wolinrelied is an anachronism. The proposition should not be used as a basis for excluding consideration of the lack of manifestation where it bears on certification requirements like predominance or typicality. It makes little sense to certify a huge class in which all but a few of the putative class have not been harmed through the manifestation of a defect.
Next, the court cited opinions from the Ninth Circuit for the proposition that Article III standing in California state Unfair Competition Law (UCL) claims can be established by alleging point-of-sale injury (i.e., that the plaintiff paid more for the product than it was worth). The court presents the proposition without discussion of the facts and Ohio tort law causes of action in Glazer.
Finally, Glazer cited Sullivan v. DB Investments, Inc., 667 F.3d 273 (3rd Cir. 2011), for the proposition that “’Rule 23(b)(3) does not…require individual class members to state a valid claim for relief’ and that the ‘question is not what valid claim can plaintiffs assert; rather, it is simply whether common questions of common fact or law predominate.’” Glazer 678 F3d at 421. Glazer’s citation to Sullivan is interesting and shows how far a court can go in glossing over certification requirements in certifying a settlement class, something Glazer suggested be done in this case.5
Setting aside Glazer’s reliance on a mixed collection of inapposite authority, there is a practical problem with the result. Certifying a broad class comprised of all owners or purchasers of many different models of a product where the manifestation rate for the alleged defect is very small creates a pretty big cudgel to brandish against a defendant. That problem has been recognized in the context of overly broad class definitions.
In Kohen v. Pacific Investment Management Company LLC, 571 F.3d 672 (7th Cir. 2009), the plaintiff accused the defendants of violating the Commodity Exchange Act by cornering a futures market. The defendant claimed that the class should not be certified because not all class members lost money while speculating in the futures market. Some made money. The court responded in part as follows:
A related point is that a class should not be certified if it is apparent that it contains a great many persons who have suffered no injury [**13] at the hands of the defendant, [citations omitted] if only because of the [*678] in terrorem character of a class action. [citations omitted] When the potential liability created by the lawsuit is very great, even though the probability that the plaintiff will succeed in establishing liability is slight, the defendant will be under pressure to settle rather than to bet the company, even if the betting odds are good. [citations omitted] For by aggregating a large number of claims, a class action can impose a huge contingent liability on a defendant…..This suit does not jeopardize [PIMCO’s] existence. But it has good reason not to want to be hit with a multi-hundred-million-dollar claim that will embroil it in protracted and costly litigation – the class has more than a thousand members, and determining the value of their claims, were liability established, might thus require more than a thousand separate hearings.
So if the class definition clearly were overbroad, this would be a compelling reason to require that it be narrowed. [citations omitted]
Id. at 677-8. The court concluded that PIMCO had not yet established the magnitude of the class members that had not been injured and hence the extent to which the class was overly broad.
The Sixth Circuit’s complete rejection of the argument that a class definition should be narrowed where it includes mostly uninjured members while at the same time avoiding the issue of lack of manifestation in the context of predominance does nothing to alleviate the impact certification can have on the defendant. Although Glazer only involved an Ohio class, at least 13 other cases involve other state-wide classes pending in the MDL. The impact of class certification on the defendant is not at all small. Given the evolution of class action law away from determining certification based on the plaintiff’s allegations and requiring a more meaningful and rigorous analysis of the underlying facts, factual disputes and expert testimony relating to certification after full discovery of class issues, courts should consider lack of manifestation in determining whether to certify the class and not simply reject it out of hand. Put another way, if the certified class, like the class in Glazer, contains mostly members whose machines have not and likely will not develop mold and odors, then the court should take a harder look at certification issues of predominance and typicality.6
Glazer is a step back in time to the days of the “first generation” class certification decisions. Its outcome was dictated by complete reliance on the plaintiffs’ allegations, theories and causes of action, with little attention given to analyzing underlying factual disputes bearing on certification. Although the Sixth Circuit acknowledged the admonitions in Dukes concerning the proper interpretation of Eisen, it did not follow those principles. Instead, the court back-tracked to the pre-Dukes era where the plaintiffs’ allegations guided the decision, while largely ignoring the underlying facts and issues bearing on certification. As a result, the Sixth Circuit upheld certification of a consumer product class without the necessary rigorous analysis of all of the class certification requirements.
1 Plaintiffs conceded that damages required individualized proofs. In re: Whirlpool Corporation Front-Loading Washer Products Liability Litigation, 2010 US Dist LEXIS 69254 (ND Ohio July 12, 2010)*9 (hereinafter also referred to as “Glazer”).
2 Daffin v. Ford Motor Company, 458 F.3d 549 (6th Cir. 2006). The Sixth Circuit in Daffin said that whether Ford’s express warranty requires that a defect manifest itself in order to make the warranty applicable is a question of contract interpretation, which is a merits issue the court could not decide in a class certification motion. Ford’s express warranty said it would “repair or replace…any parts on your vehicle that are defective in material or workmanship….” [as opposed to parts that fail due to a defect]. Ford argued the warranty required manifestation, but the warranty did not expressly say so. Therefore that the defect had not manifested itself in some class members’ vehicles was not a basis to deny certification at that stage of the litigation. A single defect was alleged in a single vehicle model, so whether there was a defect in that vehicle was a common question. If the district court later agreed with Ford’s interpretation of its warranty, the court could de-certify the class.
3 Professor Nagareda made a similar observation in the context of reliance on expert opinions of aggregate proof of sophisticated statistical or economic analysis “that presumes a view of the proposed class in the aggregate.”
“Certification based simply on the assertions in the complaint or an admissible expert submission exhibits a troubling circularity. The legitimacy of aggregation as a procedural matter would stem from the [strategic] shaping of proof [for purposes of class certification] that presupposes the very aggregate unit whose propriety the court is to assess.”
Nagareda, 84 N.Y.U.L. at 125-26.
4 The appellate court in Wolin reversed a lower court ruling that certification is not appropriate unless the plaintiffs proved that the defect manifested itself in a majority of the class’s vehicles.
5 In the “Background” section of its opinion, the Circuit Court in Glazer included reference to the fact that Whirlpool had settled an earlier class action involving different washing machines that allegedly experienced some degree of mold growth. Later, it also referenced the possibility of settlement is this action as well. One can’t help but wonder whether the direction of the Circuit Court’s opinion was influenced by a belief that settlement was the appropriate resolution for this case as well and it was, in effect, certifying a settlement class.
6 The Sixth Circuit implicitly acknowledged concern about including members in the class who had no manifest injury. It suggested that the district court consider certifying a Rule 23(b)(2) subclass of members who have not had any mold issues which could obtain declaratory and injunctive relief to protect their interests should they experience the problem in the future. Glazer, 678 F3d at 421. Effectively, the court would rewrite Whirlpool’s warranty for this subclass to waive time limits. This of course does not resolve the issue of whether, under a more thorough predominance analysis, the class of predominantly uninjured consumers should have been certified in the first place.