# Oral argument — City of Philadelphia v. Bank of America Corporation > Briefs, arguments, decisions, and more. URL: https://www.frixlaw.com/law-library/documents/argument%3A98208 ## Record - **Collection:** Oral argument - **Document type:** Oral argument - **Published:** April 7, 2025 ## Text Good luck to you and thank you for being here. Enjoy the rest of the day down in the Civic Education room on the 5th floor. Just wait one second, Mr. Wick, just while the room clears. All right, so Mr. Wick, you have reserved two minutes for rebuttal, so that gives you eight to start. The floor is yours. May it please the Court, the District Court order should be reversed for two principal reasons. First, the District Court analyzed the plaintiff's expert evidence only under a Daubert admissibility standard. And second, the Court improperly discounted the defendant's individualized defenses in its predominance analysis. Turning to the first question, the defendants made a series of showings that the expert evidence that the plaintiffs offered to satisfy Rule 23's predominance requirement is a blunt and indiscriminate instrument that cannot distinguish injured BRDOs from uninjured ones. For example, the defendants showed that when the plaintiff's expert models are applied to periods when everyone agrees there was no conspiracy, the models still try to tell us that BRDOs were harmed by a non-existent conspiracy. And among other things, when the models are applied to the five-year period the plaintiffs call their clean period, the models purport to tell us that up to 100% of BRDOs were injured by a non-existent conspiracy. Well, Mr. Wick, I think you said your argument is that the District Court only applied Daubert. But that's not what the District Court said, right? The District Court said it was applying a Daubert analysis to the extent that defendants seek to exclude testimony. And then said, after that, Daubert doesn't end the analysis and then considered sort of the dueling experts. What it seems to me is what you want is that you're saying the District Court had to resolve the dispute between experts before certifying the clients. Is that right? Yes, we are saying that the District Court had to resolve the dispute between the experts and that Teamsters and IPO require that, among other cases. But as to whether the District Court— Well, why would that be? That's like a motion for summary judgment then, isn't it? Well, if I may complete the answer to Judge Sullivan's question. On SPA-21, where the District Court said the Daubert inquiry does not end the analysis, it said the only reason the Daubert inquiry does not end the analysis is because the defendants made four other arguments independent of their challenges to the models. In evaluating the models, the only thing the District Court said was that the defendants' challenges don't preclude the admissibility of the expert evidence. The District Court did not take the further step of analyzing under a Rule 23 rigorous analysis standard whether the models are or are not capable of proving injury on a class-wide basis. But I guess I'm pushing back on your assertion that the District Court had to resolve the dueling expert issue. I thought the Supreme Court was pretty clear about this in Tyson Foods. There, the point is once the District Court finds that the expert evidence is admissible, then it's only appropriate to deny class certification based on the persuasiveness of expert evidence if no reasonable juror could have believed the expert evidence. So that's really your argument, right? No rational juror could have believed their experts over yours. Well, I would say two things. Tyson Foods and Angen are the cases that govern where the dispute does not bear on a Rule 23 requirement. The cases that govern where the dispute does bear on a Rule 23 requirement are Comcast and Goldman Sachs v. Arkansas Teachers and Halliburton II. Those cases say that where the dispute bears on a Rule 23 requirement, we can't just ask whether the jury would believe the plaintiff's evidence. The District Court has to weigh competing expert evidence and resolve disputes among the experts under a preponderance of the evidence standard. But the argument here is that causation, right? You've got dueling experts really who are disagreeing about causation, right? They're disagreeing about injury and causation, yes. So if the jury agrees with the expert, the plaintiff's expert, then there is predominance because that one issue will establish causation for the entire class, correct? Your Honor, leaving it to the jury to decide whether the models are or are not capable of distinguishing between Well, the jury's going to decide causation, aren't they? Well, only after If it gets that far, that would be a jury question, causation. Only after Predominance is not a jury question. The jury will decide who was injured, but only after the District Court makes a threshold determination of whether the case will proceed individually or as a class action. But I think the point is that it seems what the District Court found here is that the dueling experts are really dueling over causation. And if the jury credits their expert, then that is the predominant issue. That's going to be what decides the case, right? Your Honor, both sides agree that a Rule 23 predominance question that the District Court has to decide is whether injury and causation are susceptible to generalized proof. And the cases say that if the disputes that went unresolved here are resolved in the defendant's favor, then injury is not susceptible to generalized proof. The courts say that if the generalized proof does things like generate false positives or mask unharmed class members or is simply incapable of distinguishing the injured from the uninjured, then injury is not susceptible to generalized proof. In Lamechdal, the Third Circuit did not say it's up to the jury to determine whether these models mask or don't mask the existence of unharmed class members. It said the District Court has to decide that question under a preponderance of the evidence standard after weighing competing expert evidence because the answer bears on Rule 23's predominance requirement. If a model masks unharmed class members, then injury is not susceptible to generalized proof because the generalized proof cannot give us fair or accurate answers to the question which of the 12,000 VRDOs in this class were and were not injured by the alleged conspiracy. Similarly, in cases like Teamsters and Arkansas v. Goldman Sachs, the court doesn't ask would the jury be able to believe the plaintiff's evidence of price impact or would the jury be able to believe the plaintiff's evidence of market efficiency. The answer to that question bears on Rule 23, and so the District Court has to weigh competing expert evidence and decide that question for itself under a preponderance of the evidence standard. That's been the law of this circuit since Teamsters in 2008. Turning to the second ground for reversal, Your Honor, I guess I'd say one more thing on the judge-jury question. It does not make any sense to say, well, there's maybe a 10% chance that the jury will think that these models can distinguish the injured from the uninjured and a 90% chance that a jury would not think so, so send the case to the jury. Because if the 90% probability comes up, the jury is in a bind. It can't decertify the class. Its only option will be to do something fair, use indiscriminate expert models to render an indiscriminate class action verdict that either denies a recovery to injured class members or grants a recovery to uninjured class members. I guess, but Amgen itself seems to be saying, basically, although the class certification analysis has to be rigorous, nonetheless, there is going to be overlap with the merits for many of these inquiries, and Rule 23 grants courts no license to engage in free-ranging merits inquiries. It seems to me what you're saying is that, really, this is the equivalent of summary judgment on the experts, that no reasonable juror could believe what their experts say. Isn't that really what it was then? Well, in the passage from Amgen that Your Honor quoted, what the court said right before that was, the courts do have to engage in, do have to decide Rule 23 issues, even where they overlap with the merits. What it requires is showing that questions common to the class predominate, not that those questions will be answered on the merits in favor of the class. Exactly. And whether injury and causation is a common question or an individual question is exactly what we're arguing about. The disputes that the district court refused to resolve go exactly to that question. So what's your best case for saying that the court, the district court, has to resolve that in assessing predominance, even though it feels an awful lot like a free-ranging merit? Well, Lamech-Dahl is an excellent case for that proposition. Teamsters is an excellent case for that proposition. There, the district court weighed the competing expert evidence under preponderance standard and found that the defendant's evidence was more convincing under predominance standard. And in UFCW, a RICO class action decision by Judge Lynch, Judge Lynch said injury and causation are only, they are only common questions if they are truly susceptible to generalized proof. And if the generalized proof is prone to, if it's error prone in individual cases, injury is not susceptible to generalized proof. That's UFCW at 136. But that's where it's really clear, I guess. I mean, you seem to be suggesting that the court has to decide who wins. They've got to basically say, you know, which way these scales tip, whereas Judge Furman seems to be saying that, listen, if it's a close contest, if it's not a blowout, then this is enough to get past predominance under Rule 23B announced. That's the standard that this court rejected almost 18 years ago in INRAE-IPO. In INRAE-IPO, Judge Sinlin granted class certification because the plaintiffs had admissible expert evidence that, if believed, would have met their burden. She said as long as the expert evidence is not fatally flawed, it's good enough. A really cleared standard is just the not fatally flawed standard that this court rejected in INRAE-IPO. If I may turn to the second reason for reversal, the trial here would involve thousands of individualized VRDO by VRDO showings of non-injury for particular VRDOs. The district court made two reversible errors in rejecting that showing in its predominance analysis. First, the court held that it was, quote, sufficient at this stage, close quote, that the plaintiff's evidence was admissible under Daubert, SPA 23. That was reversible error under this court's decisions in Haley and Myers, which both hold that a district court cannot give less weight to the defendant's evidence in the predominance analysis simply because it comes from the defendant's and not the plaintiff's. Class actions are not tried on the evidence of one party. That was a legal error that requires reversal and remand. It seems to me the district court explicitly acknowledged that the moving party had the burden and expressly considered the experts that you put forward. What the district court did was it said, sure, the defendant has some individualized evidence, but it is sufficient at this stage that the plaintiff's expert is common and survives Daubert. And under Haley and Myers, it is error to say that the defendant's evidence weighs less in predominance simply because it comes from the defendant's. The second reversible error the court made was when it faulted the defendants for failing to estimate how many individualized inquiries would be necessary at trial. That was wrong both factually and legally. It was wrong factually because the court completely ignored repeated estimates that thousands of individual inquiries would be required. Those are at Appendix 903, Appendix 1823, and Appendix 1836 and 1837. So what a trial here would look like would be the plaintiffs would put on their common evidence, and then the defendants would make literally thousands of one-by-one VRDO showings. The evidence on all of those showings would be different. Each of the 12,000 VRDOs in this class is unique. They all have their own individual facts, features, and circumstances that can be analyzed to determine whether or not that particular VRDO was injured by the alleged conspiracy. And what expert laid all of that out in front of Judge Furman? Professor Chalmers and Professor Hubbard laid that out, and the defendants argued that at length in their brief. That was Roman 1A of the class certification opposition, and the district court said— to make a generalized statement across the class as to injury, that each of the injuries were particularized. Let me finish. I'm sorry. That the nature of these VRDOs was such that the injury could only be understood by the specifics of the VRDO and the market in which it operated and the time in which it operated. That's not exactly our argument. Our argument focuses on what a trial would look like. The plaintiffs would put on their common evidence of injury and rest. The defendants— The plaintiffs offered proof during the hearing from two experts that said, A, they attempted to filter out other causes or other reasons why the rates might vary, and then attempted to identify and draw a conclusion with regard to the fact that collusive or conspiratorial conduct was a reason for the inflation of the rates and not other—either catastrophic events or any other reason that would normally be accepted. But I didn't see—I didn't see—and you say that—and so they presented a case that said, Our proof will do two things. It will, A, show that a conspiracy could affect VRDOs, and, B, that the effect would be uniform as to how it affected these people in that— these VRDOs in that artificially kept rates high. Correct? I mean, is that in essence what their case is? If you're asking me for is that a summary of their case, that is a summary of their case. Yeah, that's what I just said. Yes. Now, and the summary of your case was what? No, these are so unique that you can't—you can't consider them all to be the same? We—yeah, in essence— And so—and tell me the pages that you presented that. Sure. Because I don't see that in—I don't see that in Judge Furman's evidence. Sure. Well, it's true. Judge Furman did not engage with this evidence. That was Appendix 850 through 903 and Appendix 18 through— Hang on. Hang on. Sure. 850 through what? 850 through 903 and 1816 through 1836. And when you look up those pages, what you will see is 70 pages of painstaking detailed VRDO by VRDO showings that those particular VRDOs, based on their particular facts, circumstances, and features and rates, were not injured by the alleged conspiracy. Let me stop for a second. Yes. How many? That was 20—that was about 24 representative showings for the named plaintiff VRDOs and VRDOs of other absent representative class members. And was there an extrapolation from those 24 as to what the population of those would be within the entirety of the VRDOs that would be captured by the class? Yes, Your Honor, there was an extrapolation that it would be thousands. That's Appendix 903, Appendix 1823, and Appendix 1836 and 1837. Thousands isn't very specific because 1,000 is one thing, but 20,000 is another. So did you give him a number? Because I would take it that your clients knew how many VRDOs they had out. Well, there were 12,000— I would expect that they would, wouldn't you? There are 12,000 VRDOs in the class, Your Honor. We do know the number. It's 12,000. Now we're making progress. Thank you. And of the three appendix sites that I just gave you, one of them says we need to do this individualized analysis for every one of these VRDOs. The other two say it would be thousands and that many of the showings would find the VRDOs were uninjured. Right, but can I interrupt? Yes, Your Honor. It seems to me the disconnect here is that what the plaintiff's experts are saying is that there's a simple explanation for all of this. And whether this is a one plaintiff trial or a thousand plaintiff trial, the defense is going to be poking holes in that argument. And that's going to be—the way to poke holes in that argument is going to be with individual factors and experts that show this is wrong. But that's different than saying the plaintiff's theory requires a consideration of thousands of variables. That's not their theory, right? Correct, Your Honor, but a trial will not be tried on the evidence solely of the plaintiffs. Defendants have a Seventh Amendment and due process right to make all the individualized— Right, you'll get to do that. But I think the point is whether there's predominance. And so whether this is a one plaintiff trial or a hundred plaintiff trial, it's going to involve the same theory. It's going to be the same experts, the same theory, that is going to get holes poked in it either way. And so doesn't that mean there's a question that predominates? No, Your Honor, because for all 12,000 VRDOs in the class, the defendant's showing of non-injury will rely on different evidence. The evidence will be different for every one of those VRDOs. But it doesn't matter. You just have to disprove their theory a couple of times. You just need a sample to show these guys are out of—these guys are crazy. We'll take a random sampling. We'll take VRDO 16 and use that as your example to show this is a dumb theory that doesn't hold up. That's not correct, Your Honor, because I could lose some individualized showings and win others. When the jury hears 12,000 individualized showings of non-injury, it could find that the common evidence is more persuasive as to some of those 12,000 VRDOs, and the individualized counter evidence is more persuasive as to other VRDOs. No, but the point is their experts are wrong. They've given an overarching theory that is not correct. And that's going to be proven with experts. But it's not going to require a thousand-variable consideration to prove the plaintiff's theory, right? I think the answer to your question is in Amgen at pages 473 and 474. In Amgen at pages 473 and 474, the court said the way we can tell the difference between a common question and just an argument that the common proof is defective is to ask the following question. If there is a finding that the common evidence is flawed here, if there is a finding that these models generate false positives or can't distinguish the injured from the uninjured, would that still leave open the possibility that individual plaintiffs could use individual proof to win their claim? And if the answer is yes, the door would still be open to individualized proof. Then what we're looking at is a Rule 23 predominance dispute that the court has to decide and not a common question or a fatal similarity question. I didn't see the plaintiff says relying on something other than their experts. You're suggesting that they're going to get class certification with their experts here and then at trial they're going to then advance different theories? Amgen says it for certain. I'm asking about this case. Is that your understanding of their theory? I didn't see that. They may be willing to bind themselves to the mast and live or die on their common proof. But Amgen says that's not the dispositive question. The dispositive question is what could an individual plaintiff do? What could an individual plaintiff do? And if the door would be open for an individual plaintiff to come up with individualized proof, then what we have is not a fatal similarity but an argument that the common proof is defective. And if the common proof is defective, you deny class certification, and the case reverts to the traditional mode of individualized proof. It doesn't matter whether these class action lawyers don't want to bring the individualized proof. Amgen says it 473 to 474. The question is whether an individual plaintiff could offer individualized proof. That is clearly true here. Individualized evidence exists in this case in abundance. Every one of the 12,000 VRDOs in this class has its own individual facts, features, and circumstances that can be analyzed to determine whether its rates were inflated. And a jury would not have to render the same decision on all 12,000 VRDOs. It could find the common evidence more persuasive sometimes, the individualized evidence more persuasive other times, based on the particular degree of fit between the models and that particular VRDO, or simply because the defendants have shown, as we've proffered to show, that the plaintiff's expert models generate nonsensical estimates for that particular plaintiff. Unless the court has further questions for him. We've got our money's worth out of you. We've got 13 extra minutes. I appreciate your indulgence. That's an interesting question. So we will give you two minutes for rebuttal, but we're now going to hear from Mr. Cooper. Thank you, Your Honor. May it please the court, David Cooper on behalf of the Plaintiff's Appellates. The district court carefully considered all of the evidence in concluding that common issues predominate and acted well within its discretion in certifying the class. Defendants largely rely on an extreme legal position whereby, as has been noted prior in this argument, that the district court should have resolved the dispute among the experts rather than simply considering whether or not common issues would predominate. And that is incorrect as a matter of law. Under Tyson, it's very clear that the district court could decide this issue and take this issue away from the jury only if the evidence failed Daubert or if it failed essentially the summary judgment standard, that is, that no rational juror could accept it. And Your Honor asked the question about this standard. Is this ultimately what it comes down to? And it is. But the problem for defendants is that they did not argue below and did not argue in their opening brief that, in fact, no rational juror could accept our expert evidence. Nor did they, on appeal, challenge the Daubert ruling. The only mention of this at all is literally in one sentence footnote, page 20, note 4 of their reply brief, where they suggest unexplained that no rational juror could accept it. Well, I guess they'll get another bite at this apple because the summary judgment hasn't happened yet, right? Absolutely. Absolutely. They have not waived this for all purposes. We are not suggesting that they cannot argue to the district court on summary judgment or, of course, to the jury that the evidence is wrong. They have every right to do that. But for purposes of class certification, they have not argued that it fails the correct legal standard under Tyson, which is that there is no rational juror could accept this. So I'm happy to answer further questions about that first issue on the legal standard. But I just want to make one further point on that before I get to the individualized evidence. And that is, the idea that you could resolve this issue is fundamentally wrong, not just because it conflicts with Tyson and Amgen and this court's precedence as well, but also because fundamentally the district court cannot, cannot resolve this issue before it reaches the jury. So if the district court were to say that our expert was correct, that wouldn't bind the jury. They accept this on page 11 of their reply brief, that the judge's determinations at class certification cannot bind the jury. So the jury, even at that point, could say, well, we still think this model is wrong. I mean, the point is, if it's a blowout, then you effectively not pass the summary judgment standard. Exactly, exactly. But if you're not in that blowout scenario where it doesn't pass the summary judgment standard that they haven't argued, then the jury can do what it wants to do. And so this idea that you can resolve it is sort of a contradiction in terms, because it won't be resolved. The jury will still get to do whatever it wants to do with respect to that evidence that reaches them legitimately. And so therefore, this can't be resolved on class certification. There's no point in even attempting to guess at what the jury does, again, with the exception of Daubert and summary judgment. And that is why the Supreme Court correctly held in Tyson that the question is simply whether or not issues are capable of resolution on a class-wide basis. And again, quoting Abgen, not whether the answers ultimately will be in favor of the class or not. To get to the second issue of individualized evidence, they claim that the district court did not consider this evidence on pages 22 to 25 of the special appendix. The district court did, in fact, consider specifically all of the individualized evidence. And the weighing of common evidence versus individualized evidence in deciding predominance is subject to great deference that's under this court's precedent and sites. And they present no basis to believe that the district court abused its discretion in how it balanced and considered the common evidence and the individualized evidence in this case. I'd also note that I think it's important to talk about what is a common issue and how the individualized evidence factors in. And Tyson addresses this exact point. And what Tyson says is, and this is a direct quote, a common question is one where the same evidence will suffice for each member to make a prima facie showing. And there's no question, there's no real even dispute here, that the common evidence is sufficient to make a prima facie showing. What their individualized evidence goes to, and these aren't individualized offenses to be clear, this is individualized evidence that's attempting to rebut whether or not the model is correct or not. But that does not go to the question of commonality and predominance. Because in every case, in every case, a defendant could say, well, there's individualized evidence that undercuts the model. That would be true of virtually any case. They did identify certain types of VRDOs, the fixed rate, or there's certain types of them where they said that the conspiracy in essence wouldn't really be able to reach because it reached them. What's your response to that? Because the district court dealt with that, but in a very cursory manner. What's your response to that? So our response to that is all of their individualized evidence, when they're talking about different aspects of different VRDOs, all of that is responded to. This is page 1970 of the joint appendix, paragraph 106-07 of our experts' reply report. And I think it's important because what it explains is that all of this individualized evidence, we have a common response to, and that is the model already accounts for it. So the model accounts for VRDO-specific issues by taking account of that through the credit ratings and the credit watches for specific VRDOs. And what our expert explains is far better than attempting sort of in an ad hoc, post hoc kind of way to decide whether or not a VRDO was affected in a particular way or not, look at what the market thought of the VRDOs at the time. And that's what my model does, and this is why the model, in fact, takes into account all of this individualized evidence. And that's why ultimately, now the jury can accept that or not. The jury could say, well, no, I don't think the model sufficiently accounted for that, in which case they'll reject the class-wide evidence and we'll lose. Or the jury could say, yes, in fact, the model accounts for all of that individualized evidence. But either way, it is a common answer to that question. And because there can be a common answer to that question based on the evidence we presented and, frankly, even based on the evidence that they presented, it can be resolved on a class-wide basis. And that's all that's required at the class certification stage. The last point I want to make is on this question of how much individualized evidence would actually have to go before the jury. And I think my prior answer sort of goes to this question. Because the point is, if they're going to say that the model is wrong because it doesn't account for all these VRDO-specific issues, whether they put in two examples of that, 10, 100, it doesn't change the fundamental point, which is either the model, through using these VRDO-specific variables, accounts for those individual variations in VRDOs, or it does not. But that answer is not going to change based on the number of examples that they provide to the jury. And so that is why, when the district court says, you haven't shown me why thousands are necessary, they haven't shown why thousands are necessary because the thousands would be the same as the two or the five or the 10. It would still make the exact same point, and we'd have the exact same argument and response, and the jury could decide that issue for the entire class as a whole. If you get past summary judgment. Absolutely, if we get past summary judgment. Is it that the interest rates may not necessarily be the same on each VRDO, but your theory is that the inflation of the interest rates was kept artificially high because of the alleged conspiratorial conduct? Correct. They were all inflated by the conspiratorial conduct, not necessarily in exactly the same amount. In other words, a VRDO with regard to San Diego might be at a different rate than one from St. Petersburg. Absolutely, and the expert specifically accounted for that through, I think, at least three VRDO-specific variables to account for the differences between them. I mean, you read the opinion and the brief, and there's nothing that kind of identifies the uniqueness of any particular VRDO. And why, you know, having lived with Goldman for a number of years, the basic presumption made that an easy thing to understand, and so whether something was supporting an inflated price and not that. Goldman's really a different kind of case in this, because it was uniform. He had a market, whereas here these are all particularized debt obligations with varying terms to some degree. Are they all daily resets, or are they varying terms in that regard, too? And was that factor present? I know that the opinion talks about it briefly. I don't know if all of them were daily resets. I believe that that is generally true, and I know that, in fact, the expert did consider those kinds of variations within the VRDOs in the modeling. But I'm sorry, I don't know all of them. But I don't see them always as a reason to reject them. Right. And I'll just close by saying I recognize that, in fact, you know, there were these cases like IPO in the basic presumption context, which is different for the factual reason you state, but also for the legal reason that there you're considering whether or not there's a presumption of reliance. And in the absence of that presumption, there would be individual issues of reliance that would predominate. That is what the court recognized in Goldman and Halliburton. But that is not the situation we have here, where we're going to the merits, the elements of the case. And as Halliburton recognized, even within the basic presumption, where one element of it, materiality, went to the ultimate merits of the case, the court does not decide that issue on class certification, just as it can't just sort of free-roving decide any of the elements of the claim on class certification. That's ultimately for the jury to determine or the judge on summary judgment. If there are no further questions. All right. Thank you, Your Honor. Thank you very much, Mr. Cooper. Mr. Wick, you've got two minutes to rebuttal. Thank you, Your Honor. First, as to the individualized defenses to injury and causation, they say they will argue the model accounts for all the individualized variation. But the question is not, after the jury hears 12,000 individualized disputes at trial, is there a chance the jury might resolve all 12,000 of those one-by-one disputes in the plaintiff's favor? The question is, what will the trial look like? How many individualized disputes will the jury have to hear and consider? Here it would be thousands. They would overwhelm common issues at trial. And the district court never weighed that under the correct legal standard. It erroneously discounted the defendant's evidence just because the plaintiffs happened to have admissible evidence. That's error because cases aren't tried on the evidence of one party. Well, what if the district court found that both sides had presented credible explanations? Did it have to choose? The dispositive question is, what will the trial look like? If the district court says both sides have credible positions, they have a credible position that common evidence is enough. Defendants have a credible position that they can make individualized one-by-one rebuttals. Then a trial will look like 12,000 individualized disputes. I agree with you that, in essence, Judge Furman had to decide that. But that doesn't preclude the defendant from offering proof, saying, this is crazy. It makes no sense because it misunderstands how the RDOs work. And we get to say that- Am I right about that? You are, Your Honor. And we get to say this is crazy on an individualized basis one-by-one. We don't just have to say- That's not true. Look, if it was a one-plaintiff case with their same plaintiff's expert, you wouldn't be limited to the one transaction, the one instrument that's being challenged, right? You'd be able to bring in lots of other examples to show that this is a bad theory. So it seems to me whether the number of plaintiffs here is not necessarily going to dictate the number of holes that you're going to try to poke in their theory. You don't need to do 12,000, and you probably are not limited to one if it's a one-plaintiff case. Well, let's just assume there are 12,000 one-by-one individual trials. At each of those trials, we would not be limited to saying put on evidence about 10 representative VRDOs that are different than the ones in this individual trial. Well, you would be free to do that because you would be able to show that this expert is nuts, that this doesn't work. And my hunch is you probably would do that. I would say this, Your Honor. The underlying assumption here seems to be that the jury will come out the same way as to every one of these 12,000 showings. That if it accepts the individualized evidence the first 10 times, it will accept it all the time. Because there's this riding on a theory that is expert-driven that is either right or wrong. But it's not right sometimes and wrong sometimes. That is their plaintiff's theory, right? The point, Your Honor, is that it can be right sometimes and wrong sometimes. The jury could find on a case-by-case basis that in some cases the common evidence is more persuasive, and in some cases the individualized evidence is more persuasive. Maybe in some cases the model is a pretty good fit for the VRDO. Defendants have a Seventh Amendment and due process right to present all the individualized offenses at a class action trial that they would present in individual trials. And in individual trials, they would make a customized showing of non-injury for each one of the 12,000 VRDOs. The district court never weighed those showings under the correct legal standard. First, it said it was enough that the plaintiffs have common evidence, but class actions aren't tried on the basis of the evidence of one party. That's Myers and Hertz and Vann v. LLR. And secondly, the court was under the wrong impression that the defendants did not estimate the number of individualized showings. The defendants repeatedly estimated that it would be thousands, that it would be each VRDO in the class. Turning to the earlier argument about whether the models are defective, Your Honor has suggested that if it's a blowout, well, then the plaintiffs lose. But short of a blowout, we send it to the jury. This court squarely rejected that it's a blowout standard in IPO and Teamsters. In IPO, the court said the test is not do the plaintiffs have admissible evidence that a jury could believe. The test is not have they put forward some expert evidence that simply isn't fatally flawed. The not-fatally-flawed test that this court rejected in IPO is exactly the blowout test. And so if this court were to say, as long as it's not a blowout, send it to the jury, it would be breaking with IPO, it would be breaking with Teamsters, it would be breaking with the Third Circuit's decisions in Lamictal and hydrogen peroxide, and it would be breaking with West v. Prudential Securities. The cases say if a model generates false positives, it is not capable of providing class-wide proof. And here, the district court refused to resolve disputes about whether these models generate false positives. It had to resolve the disputes because if those disputes are resolved in the defendant's favor, everyone agrees false positives would be disqualifying. Then injury would not be susceptible to generalized proof. Two things are clear, Your Honor. First, the district court ended its analysis of the plaintiff's expert evidence after a Daubert inquiry. The only thing it— I guess you keep saying that, but I mean, the district court explicitly said it wasn't doing that. The district court said— Do you think the judge is lying? No, I think if you look closely at SPA 21, Judge Furman is very clear about what he's saying. He's saying Daubert is largely dispositive, but it doesn't— One of your principled arguments was that this doesn't meet the Daubert test, right? That was not our only argument. It wasn't your only argument, but that was the argument that he's responding to at that point in his decision. What he missed in our brief, and it was clear in our brief on pages 18 and 19, was even if the expert evidence survives Daubert, it can still fail to satisfy Rule 23. We made that argument. Comcast says that in footnote 4. The D.C. Circuit said that in rail freight. This Court essentially said that in IPO and in Teamsters and in UFCW, where the plaintiff's had admissible expert evidence, and if that admissible expert evidence had been believed, it would have proved classified injury or causation. And Judge Lintz said in UFCW, the Court has to figure out whether the proposed common proof is error-prone in individual cases, because if it is, injury and causation are not susceptible to generalized proof. That's page 136. Yes, Your Honor? If you're right, if your standard's right, what would there be left to prove after a judge made a determination? Say in this case, say Judge Furman had gone the extra step and said, okay, there is proof. The plaintiffs can prove their case to commonality. I accept. Not only do I accept the experts under Daubert, but that by a preponderance of the evidence, the plaintiffs have established that they can prove their case that way. What else is there left to do in the case, then? Then we would—so if the Court finds what we're asking for, is it finding— Would it preclude your attempt to show that the DRDOs are actually individualized and not capable of being subjected to conspiratorial conduct? No, Your Honor. The Court addressed that in IPO. In IPO, the Court said it's kind of like the difference between a preliminary injunction and a final injunction. So the Rule 23 determination would be, by a preponderance of the evidence, can these models prove—have the plaintiffs proven by a preponderance of the evidence that these models can distinguish the injured from the uninjured? If the answer is— Isn't that exactly what Judge Furman did? No, he didn't resolve that question under a preponderance of the evidence standard. If you're wrong about that, is this an affirmance? I beg your pardon? If you're wrong about that, is this an affirmance? If Judge Furman— If he did make that finding, if he did, in essence, do that work— By a preponderance of the evidence— Yeah. If he found, by a preponderance of the evidence, after— Because a preponderance of the evidence would—you're saying it wouldn't preclude you from then still providing—attempting to prove that these are individually—individualized injuries. Correct, Your Honor. IPO says that the Rule 23 determinations bind the procedural question, is the case going to proceed individually or as a class. It doesn't bind at the merit stage. Here's where Judge Furman was— Binds by a preponderance of the evidence what? That the plaintiff has established that there's a commonality with regard to injury? In order to prove that injury is susceptible to generalized proof. Right. The plaintiff would have to prove, by a preponderance of the evidence, that their models are actually capable of distinguishing injured VRDOs from uninjured. Okay. And you're saying that Judge Furman did not do that? Correct. At SPA, I want to say 13, he said the defendants have made forceful, compelling— I've read Judge Furman's opinion several times. He said perhaps even meritorious challenges to the models, but those perhaps meritorious challenges are not enough to preclude admissibility. Do you want us to remand to Judge Furman for the appropriate findings? Yes, Your Honor. All right. Thank you, Your Honor. All right. Thank you both. We will reserve decision. --- Source: Frix Law Library, https://www.frixlaw.com/law-library/documents/argument%3A98208. Public record. Not legal advice.