Hello, everyone. Thank you for joining me for this presentation. My name is David Burnett. This is a presentation on the importance of data and data analysis. I will be speaking for two hours today. This is part one. So the topic again is data analysis and litigation. I'm pleased to be joining Quimbee for this CLE presentation today. This presentation is based on my personal experience. It's built on a law360 article that I wrote last summer on the same topic. I've been an attorney for 15 years, but I've never presented a CLE before other than this topic where I presented last month for the D.C. Bar. So this is a new experience for me. I appreciate you joining me and and giving me your attention. It's gratifying to be at a point in my career where I have this much to share with my colleagues. Um, so again, this data, this presentation is on data analysis. It's an opportunity to share some insights on the topic of using data in litigation that I've gathered over 15 years as a litigator, where for whatever reason, working with data has become a recurring theme in my career over a variety of different subject matter. I've worked with complex data sets in mortgage backed securities, securities, class actions, opioid litigation, and as a litigation funder. This presentation is probably most relevant to litigators and specifically people that work on the plaintiff's side, as I do. But within litigation and within plaintiffs work, data analysis spans lots of different types of cases. So I hope this presentation is broadly applicable to different people's work. And again, just because I'm a plaintiff's litigator does not mean that it's not also applicable on on the defense side because the same issues come up whether you're on the plaintiff or the defense side in these cases. So in this presentation, part one, I'm going to talk a little bit about my background. I'm going to explain this topic of data in litigation. And then in part one, I will describe how data has been used in mortgage backed securities litigation and in opioid litigation based on my experience. And then in part two, I will describe the use of data in three other subject matter. Matters and then end with some practical recommendations about how to use data in your cases. I should note right off the top that these slides and my comments today are entirely my own. I'm not purporting to speak for Motley Rice or any past employers. I'm just sharing my personal experience and observations. And obviously none of this information is work, product or privileged. Um, so without further ado, let's get into it. All right. So, um. Just briefly so you understand my background and how it relates to the topic of data. I'm currently a senior counsel at Motley Rice in Washington, DC. Motley Rice is one of the country's largest plaintiffs firms. These days, we're most known for being co-lead counsel of the national Opioids MDL on the plaintiff's side. So we represent states and counties and cities in litigation arising from the nation's opioid epidemic. I've done a lot of work in that area, including with epidemiologists and economists as experts. We also do a lot of securities class actions. We work on a lot of mdls. We do environmental cases, mass torts. So my work has touched on a lot of those different areas. Again, all on the plaintiff's side, mostly class actions. Before that, I worked at Burford Capital as a vice president. Burford is a litigation funder. It's one of the world's best known litigation funders. They, uh, they offer litigants generally plaintiffs money to fund cases. And if the case turns out well, then Burford or a similar litigation funder receives some portion of the the counterparties entitlements. So if you're a plaintiff, Burford would get some percentage of your contingency fee if you're the plaintiff's lawyer or some percentage of of your settlement or damages if you're the plaintiff. Anyway, so at Burford, I also did a lot of work with data, as I will explain. And then prior to that, I worked for ten years at Quinn Emanuel in New York, was an associate then, was enough counsel. My work there, as I'll touch on today, um a lot of it work involved mortgage backed securities litigation so residential and residential mortgage backed securities, i.e. RMBs. That was a whole wave of litigation coming out of the financial crisis of 2007 to 2008, where institutional investors like insurance companies lost a lot of money, um, from these investments. And, uh, law firms like Quinn Emanuel sued on behalf of those investors to recover some of their losses. So again, three jobs out of law school. All three of them have have. Have had a heavy data focus. So what is this topic all about? So I was inspired to write this because. Over over my lifetime, you know, leading into law school, during law school after law school, everyone jokes about how lawyers are bad at math or lawyers hate math. And that's why people become lawyers is because they don't want to work with numbers. They they leave that sort of work, you know, the number crunching to accountants or MBAs or engineers or people like that. A lot of lawyers come out of a liberal arts background. However, my experience has been that. Working with numbers is a critical, critical element to to being a litigator, particularly in the modern era as cases get bigger and bigger. Um. So in my experience, I have seen that, um, you know, working with numbers of all sorts, as I will explain, um, is critical to modern litigation because as I note here, lawsuits are becoming increasingly complex and ambitious. Um. It's hard to avoid numbers in math litigation. Every every case is going to involve an analysis of damages and settlement value regardless of the subject matter. And in many cases, the subject matter also involves numbers, whether that's breach of contract, a securities case, any number of cases that involve complex expert analysis, which involves economists or scientists or accountants or people like that. So my experience has been it's unrealistic to avoid working with numbers unless perhaps you're a pure appellate lawyer who only works with case law. So you might as well try to understand data and if not, embrace it. Um, so I've only been a lawyer for 15 years and therefore don't have a broad historical context. But my sense is that cases are getting bigger and bigger in scope, in alleged damages, in settlement amount and in ambition. We still see lots of one off small cases between individual parties, but increasingly we're seeing more and more mdls huge class actions in waves of separate but parallel litigation across the country. So whether that's mdls on topics like opioids, breast implants, hernia mesh IUDs, which now make up a majority of cases in federal court, or whether it's big, ambitious one off cases involving antitrust securities, climate change or other wrongs. Lawsuits seem to be getting bigger and bigger and more ambitious. The bigger the cases get, the more we need data to organize information and tie the parts together. So. I'm going to talk about the benefits of using data in cases. Um, in my experience, data often acts as the foundational evidence for cases. So in mortgage backed securities cases, for example, where investors allege that banks misrepresented the quality of mortgage loans that were packaged together in investments, the case is at heart are about whether the loans were misrepresented and whether they were faulty. And that in turn depends on statistics about. The degree to which the loans were misrepresented and whether the numbers applied to each mortgage were correct or not. In opioids cases or topics in include addiction rates, death rates, volumes of prescription opioids shipped year over year and the cost of mitigating the epidemic. All of those topics involve a heavy data analysis in securities class actions. The harm is premised on a drop in stock value that can be attributed to misrepresentations. So that again, always involves expert analysis of damages involving numbers and an antitrust. Many cases involve alleged manipulation of securities markets, which involves heavy data analysis. So all of these cases also involve, of course, document discovery, as we all think of it, where the parties, um, exchange requests for production. They come up with search terms, they discuss topics, um, that they agreed to produce on. They negotiate search terms. They negotiate, um, uh, you know, search protocols. They run electronic searches on emails. Um, they pull hard copy documents and then both sides review emails and letters and PowerPoints and other documents like that to look for key documents that will help their case. They're looking for, you know, for example, juicy quotes from executives in an email which speak to their scienter. All of these cases also involve deposition testimony where both sides are looking for great soundbites that they can use at trial. Um, so the importance of data in these cases does not take away from the role of documents and testimony, but think the world is changing where cases are becoming more and more about the numbers. Um, and maybe the focus on hot documents and salacious details. It's still important, but when cases are so big, um, you know, when you're talking about market wide or population wide claims, you know, little snippets here and there, um, just, you know, perhaps pale in significance. Um. So data can be useful because it seems like hard evidence. It seems objective. Of course it can be manipulated and it's only as good as as the reliability of the numbers that you're working with and and the way it's explained. But. Nonetheless. Data. Data at least, has the appearance of objectivity. My mom, who was a math teacher, used to say that numbers don't lie, but you can lie with numbers. And that's where a lot of the challenge and the the. The skill involved in working with numbers comes from is is coming up with reliable explanations that withstand scrutiny from opposing parties and from the court. So people like numbers because they seem scientific theories will pay attention as long as the data is explained well. Judges are often persuaded by by notable statistics. Data can be the backbone for cases. The hard evidence, which is combined with qualitative evidence like juicy emails to together prove cases. As I will discuss, though, there are plenty of challenges and risks with using data. The other side, in whatever case you're working on, knows that data can be powerful. So there are going to push back hard and they will have their own data and their own experts in their own testimony to to challenge what you're saying. So. Um, as I will explain, data is also not only useful for helping to prove cases, it can also make cases more efficient, allowing you to use shortcuts to encompass a lot of evidence, like using statistical sampling to prove misrepresentations in a way that can streamline cases. Okay. So economic benefits of using data and litigation. We know that generally speaking of cases, value is proportional to its scope. So its value in terms of damages and settlement value, The bigger the case, the more the plaintiffs, the more serious the harm, the bigger the damages. Um. So, you know, yeah, the variables include the number of plaintiffs, the number of investments. If it's a securities case, the broader the time period, the more defendants, the greater the alleged damages and settlement value. So as plaintiffs lawyers, if you can prove a broader case, all else being equal, that's more desirable than a narrower case because it means more recovery for your plaintiffs. It means helping more plaintiffs and it means more attorney's fees or contingency. Um, so there is a self-interested reason for plaintiffs lawyers to want to scale up cases. But it also benefits more plaintiffs if you can if you can get the case proved at trial or if you can settle it. Um, so data sets are great because they allow lawyers to bridge the gap from individual claims to big society wide claims. So for example, you can go from a wrongful death case saying that John Doe died of an opioid overdose because his doctor over prescribed OxyContin. You can go from that to a case alleging that an entire state experienced an epidemic of thousands of opioid deaths and tens of thousands of people addicted because of a years long pattern of over supplying prescription opioids by Fortune 100 companies, these enormous pharmaceutical companies and pharmacies and distributors. So rather than talking about just one person, you're instead talking about the whole society and a whole pharmaceutical industry. Um, so whether we're talking about market data, about securities sales records for pharmaceuticals, aggregated health records for patients, demographic data about disease rates or scientific information about addiction and other harms. Data sets can aggregate huge swaths of information about individual people and events into a concise summary that allows you to scale up the case and still explain it in a manageable way. So the result is by being able to summarize broad information with concise numbers and doing so in a way that's reliable and admissible and understandable to a judge and a jury. Lawyers have been able to take on and get favorable outcomes in very ambitious cases. So we're going to talk about a few of those today. Litigation and opioid litigation. And then in part two, we'll talk about antitrust and securities and litigation finance. Um. All right. So how to use data in litigation. So the rubber meets the road when we talk about how to use data. It's one thing to have the data. It's another thing to understand it. Have experts help you with it, to analyze it, to present it well, and to have a fact finder, whether the judge or the jury accept the data as compelling. So data is only effective if it's number one collected in the first place. And then number two, analyze well and number three explained well. Um, so as a lawyer working with data, as I have, I've had to become very comfortable with Microsoft Excel, for example, as a way to organize data and then, you know, working with PowerPoint to think about what is the best way to present the data in, in spreadsheets, in tables, in bar charts. Um, we might romantically think that the work of a litigator is waxing poetic in court like Clarence Darrow. But the reality for me has been over the years, a lot of time, late nights in front of a computer after, you know, eating dinner at my desk, um, spending hours working on documents, working on spreadsheets, creating presentations, all in an effort to collect and understand and present data in a way that's understandable to other people. So a lot of the work of data analysis involves collecting the data, understanding it, understanding how it can help your case, how it fits into the rest of the evidence you have, what the data shows, and then determining how to organize it. So often that means distilling huge data sets with formulas and expressing the data in summary form. Could go without saying. Oftentimes working as a lawyer in data intensive cases involve working with consultants and testifying experts prior to filing suit. Even oftentimes, lawyers use consultants to help them develop the claims to determine who the defendants are, to determine the scope of wrongdoing, um, uh, before a complaint is even filed. Um, and then in discovery, lawyers will often use consultants to help them collect and process and explain the data. And then the lawyers will almost always use experts in these big cases to present the data. So those could be experts in epidemiology, economists, economics, statistics, marketing, medicine, etcetera. All right. Let's talk about a few challenges. Um. So that is right. When you have the right, you have enough of it. The predicate. But data can also be a pain and hard to work with. Sometimes a lack of data or problems with processing the data or explaining it can kill your whole case. If you go into a case hoping that you're going to get access to a data set and that it will be complete and that it will show certain things that will help your case. And then you can't get that data or it doesn't show what you thought it would. That that can be fatal to your case. Um, so, you know, datasets can be unavailable. Perhaps no one maintains the data on a certain topic that that you thought they would. Maybe that data has been deleted. Maybe it's incomplete. Um, maybe there's no one available to validate it. Um, maybe it's not stored in a way that can be exported and presented in a in a in a right in a understandable form. Perhaps your experts aren't comfortable using the information because they don't think it says what you hoped it would. Perhaps the judge doesn't admit the analysis. Um. The other the other challenge to think about is in these big cases, big cases are inherently more expensive. They take more time. They can drag on longer. They involve more attorney hours. They also involve more money typically spent on consultants and experts. So you can spend, you know, tens of millions of dollars in a in a huge wage wave of cases just on experts and consultants. So you have to be mindful of that with these big cases where you're scaling up the scope of the claims and the scope of the evidence and bringing in complicated data as part of the way to prove your case and hiring expensive experts and consultants to help you do so, you just have to be mindful of the cost and make sure that it's still proportional to the expected settlement or damages value. So assuming you have complete data sets and it's reliable and you have well qualified experts to help you analyze it and explain it, then you need to present it. Um, the more helpful your evidence is to your case, the harder the other side will fight to undercut it and dismiss it and have it excluded. So that's particularly true in high stakes litigation with huge damages at issue where data can be critical to making the broad allegations work if the other side can get you knocked out on Daubert. That could potentially kill your whole case or leave you scrambling to find another expert if you have time. So we won't get into this would be a whole separate topic and won't get into it today. But I will just note that when working with data and working with experts, you of course need to be aware of the Daubert or Frye standard depending on your jurisdiction. Um, I'll just describe them briefly. So the. A federal rule of evidence. 702. Testimony by expert witnesses touches on the qualifications of an expert witness. There are four factors a the expert, scientific, technical or other specialized knowledge will help the Trier of fact to understand the evidence or to determine a fact in issue. B testimony is based on sufficient facts or data. C The testimony is a product of reliable principles and methods. And D the expert has reliably applied the principles and methods to the facts of the case. And then the Daubert standard elaborates on that. And then separately, there is the five standard for admissibility of evidence. Again, I won't get into those today, but just noting that when you're working with experts and you're working with data, just be mindful of that. Gatekeeping step that you'll need to get past. So you always need to be thinking not only about the persuasiveness of the data to the factfinder, but also the admissibility. All right. So let's move into the different examples from my work experience of how data sets are used in litigation and what I've seen pros and cons, mostly pros, but but some life lessons of the challenges of using data and how I overcame those that the lawyers and I that I worked with. So in part one, we're going to talk about the first two of these structured finance litigation, what I just what I call residential mortgage backed securities or RMBs litigation. And then I'll also talk about opioid litigation. So product liability. And then in part two, I'll describe I'll talk about securities litigation, antitrust and litigation finance. Okay, so example one litigation. This is a subject matter that I fell into as a first year associate at Quinn Emanuel because when I joined the firm in 2007, the financial crisis was just starting. Um, the. The economy was was was tanking in large part because of problems in the housing market, not only in the value of houses, but also delinquency and default rates, bankruptcy rates, um, and also a loss in value of the securities backed by those. Buy those mortgages. So. Um, residential mortgage backed securities are bonds where the collateral backing the bonds is mortgages. And typically in a single offering you would have anywhere from 500 to 10,000 mortgages that are all pooled together. And the investors who buy those bonds, they're entitled to the, um, the payments by those homeowners on their mortgages. The problem is that if people stop paying on their mortgages, um, that revenue stream from the mortgages goes down. And often we learned that the mortgages that were pulled together in these offerings were often very risky. So they it turned out that a lot of these people ended up defaulting on their mortgages. So the principal and interest that was being paid out to the investors went down oftentimes. And the value of the bonds also went down because people realized investors realized they were riskier than they thought. So a lot of these. Institutional investors that had bought mortgage backed securities such as my clients. My two main clients that I worked with were Allstate Insurance Company and Prudential Insurance Company. They had bought billions of dollars of bonds and those bonds tanked in value. And so they hired Quinn Emanuel to come in and help them try and recover some of those losses through litigation. So in our case, we filed dozens of lawsuits against the, um, the investment banks that had pooled together the, the, the mortgages and sold them to investors. So, um, Allstate and Prudential each had a lawsuit against Citibank and Credit Suisse and Deutsche Bank and RBS and. Ubs and Bank of America and JP Morgan, Deutsche Bank. All of the big banks have been involved in this in this market. Um, and the bonds that Allstate and Prudential bought from each of those banks. Um. Tanked in value. So we sued the the sponsors, the the banks that had pulled together these these mortgages. There were also later there were lawsuits against the the servicers of of the trust and the trustees. Um, the first lawsuits were typically fraud. Um, some were federal securities fraud. Some were common law fraud, typically under New York law, some were breach of contract. Um, and as I will explain in, in large part, not in large part, but in part because of successful use of data, these cases turned out very well. So I had the unfortunate experience of being involved in one of the earliest fraud cases that were filed and that reached resolution. Um, and unfortunately, it turned out badly. So that became a learning experience for me and for all lawyers representing plaintiffs in these cases. So this was a case called Footbridge V Countrywide. This was brought on behalf of a small hedge fund that lost money through its investment vehicle called Foot Bridge. Um, lost money on that. It bought from Countrywide, where Countrywide had originated the mortgages and pulled them together and sold the bonds backed by those mortgages. Um, Judge Castel of the FDNY dismissed the complaint. Um, he so in our complaint, we had, uh, we had included a lot of information about how bad Countrywide's mortgages were as a whole company wide. Uh, you know, we included newspaper articles and, um, I don't remember if yet at that point there were, there were federal investigations, but we included everything we could find publicly about Countrywide. And we described how these specific bonds, um, had decreased in value and how their credit ratings had gone down. Um, nonetheless, um, Judge Castel granted summary judgment to Countrywide because he held that we had not sufficiently tied the general allegations about the company to this specific investment. And part of the investment was sorry, part of the debate was whether rule nine of the federal rules of Civil procedure, which requires pleading with particularity, whether that required us to plead more and more about the individual investments. Um, we felt that, um, Judge Castel was, was applying to too high a standard. Oh, and sorry, I said granted summary judgment. I meant that he granted. Um. Uh, granted a motion to dismiss, but he seemed to require, by our standards, a summary judgment like degree of proof at the pleading stage. So we learned our lesson from this. What we did afterwards in the Allstate and Prudential cases that followed and what other plaintiffs lawyers did and other plaintiffs cases in, was to include as much information about the specific investments as we could at the complaint, at the at the pleading stage. So what we did specifically was we found a vendor that had a database of mortgage loans, and we found a way of connecting that public information about nationwide mortgage loans to the limited information that our sponsors that the banks provided to investors. So they would give you summary information about. Uh, the individual mortgage loans that were underlying each trust. And what we did was we matched up that summary, anonymized information to public records, um, in what we called a forensic analysis. And we used proprietary formulas to do so and proprietary methodologies that we came up with. And that allowed us to try and find which real world mortgages collateralized those, um, those trusts. And then we were able to do our own analysis of representations that the sponsors made about those trusts. So in other words, we came up, we found the data and we came up with a whole mathematical analysis to second guess the numbers that the banks provided to investors about key characteristics of the mortgages, including whether the properties were were owner occupied and the loan to value ratio of these properties. The details don't really matter. But the point is, um, to get passed a motion to dismiss, we found data and analyzed it using consultants in a way that in all subsequent cases that I worked on, we were able to get past every single motion and. Smith After that first footbridge lost. So we learned, we learned from our experience with footbridge. We beefed up our complaints, um, by doing this proprietary forensic analysis. Um, and that analysis coupled with information about loss in market value of the bonds and decreases in credit ratings plus, um. Information about company wide wrongdoing pulled from newspaper articles and governmental investigations. All of that information together was sufficient in each of the Allstate and Prudential cases to get past motions to dismiss and believe the same is true for all of the other plaintiffs out there that Quinn Emanuel represented, including MassMutual, AIG and Fannie Mae and Freddie Mac. So it was the layering. It was a combination of country company wide allegations, plus investment specific allegations that really helped us. So this is an example. This is an excerpt of tables that I prepared that were included in our complaint for Allstate against Credit Suisse. And these are representative of the kinds of tables that we included in every Allstate and Prudential complaint. Um, so Allstate had eight cases. Um, Prudential had about 12. So these are summary spreadsheets that I created based on much more extensive Excel spreadsheets. These are the key extracted data points. This shows the asset is each of the trusts. The second column is what the defendants represented. The third column is what we are alleging. The true information about these trusts was based on our forensic analysis. And then the fourth column shows the overstatement, which shows, as we alleged it, a material misstatement of these key characteristics. Um, the first is whether the properties were owner occupied. The second is the loan to value ratio of the properties. So I remain very proud of this analysis that we did because it involved a lot of work, it involved a lot of creative number crunching, and it was effective because, you know, these are at least on their face, very hard numbers and very persuasive numbers that a judge can look at and say, well, you know, at the pleading stage where I, um, I must accept, well, pled allegations is true. You know, a judge can't fault this, at least on its face. Um. By the way, one consequence of all of this work to get past the pleading stage in these cases is that every complaint of ours got longer and longer because what we would do is we would respond to all of a given defendant's arguments. In one case, we would respond to them in opposing a motion to dismiss. And then whatever they argued in their motion to dismiss, we would incorporate more allegations into each subsequent complaint to to counter that. So whether on the topic of data analysis or any other topic, these complaints just got longer and longer, longer and more and more complex because we were throwing in more and more information and more and more allegations to try and combat. Um, defendants arguments. And as we ramped up our data analysis, defendants ramped up their, um, their critiques. But as I'll show you on the next few slides. Judges accepted this forensic analysis of ours as sufficient to get past motions to dismiss. So this is an example from New York Supreme Court. That's New York's state trial court in southern Manhattan. This was Allstate's case against Deutsche Bank. Um, Justice Bransten denied Deutsche Bank's motion to dismiss noted, citing other cases that misrepresentations of such data has been held to be actionable. And noted that while defendants try to downplay the data discrepancies as merely a difference in valuation methodologies, that theory rests on factual issues outside the scope of this motion. So in responding to the elaborate data analysis that we came up with in these cases, defendants would try to quibble with the numbers or they'd try and quibble with the analysis. But judges rightly noted that, you know, they were trying to apply. The defendants were were trying to apply a summary judgment like standard. And they were suggesting that we basically had to prove our cases at the pleading stage and that the numbers that we were alleging in the complaints did not prove our case. But the judges rightly said that that's. Getting into methodology and quibbling with the details of numbers is something best left for expert discovery and summary judgment. Um, here's another example. Justice Friedman of the New York Supreme Court, the which is in New York Trial court. Um, Allstate's case against Credit Suisse again denied the motion to dismiss. Um, Justice Friedman said allegations regarding specific misrepresentations as to loan to value ratios, owner occupancy and credit ratings are sufficient to support the fraud causes of action. And then referring to our, um, uh, our forensic analysis, Justice Freedman said plaintiff's own development of complex methodologies enabled them to conduct a loan level analysis of the mortgages. That term loan level analysis kept coming up. That was a point of why our work was reliable, because it, it, it went to, um, specific loans. So it was a granular analysis that was also explained in a summary fashion. One more example. Prudential. The Bank of America. This was Judge Chesler of federal court in New Jersey, also denied the motion to dismiss. So in this case, we had analyzed our complaint. The summary tables included analysis that encompassed over 20,000 individual mortgage loans from 30 of the trusts. And we didn't always do a loan level analysis on every single trust in a case because that would be overkill and mean. Our position all along was that this type of detailed forensic pre, you know, pre-discovery analysis at the pleading stage should not be necessary because this degree of detail should not be necessary even under a rule nine standard in federal court. Nonetheless, we felt obligated to because of the footbridge. Bad decision. But nonetheless, it was gratifying over and over to note that, you know, as you would hope, in a case like this, Prudential Bank of America case, where we looked at over 20,000 loans, the judge, of course, hope, you know, you would hope, of course, found that 20,000 mortgage loans was a sufficiently large sample to make plausible inferences about all of the offerings, particularly when it is considered, together with the other factual support offered. And again, this is in a complaint that by this point, this was one of the later complaints that we prepared. The team that I worked with. This complaint was over 200 pages long and had over 600 paragraphs of allegations. So you would hope that, you know, with looking at 20,000, with making allegations about 20,000 mortgage loans plus all of the company wide information, plus all the other categories of of facts that we allege, it's gratifying that that he correctly found that our allegations were sufficient. One other way that data was used, and this is in discovery. So the challenge with these cases is that like in the in the Bank of America case that we just looked at, there were 45 sorry, 54 individual trusts. Each one of those, as I noted, could have up to 10,000 mortgage loans. So. A case like that involves. Hundreds of thousands of individual loans. And if you were to try and prove. That, you know, some of those individual phones among the 100,000 or more represented that would take forever mean you know. Holding a trial where you talked about hundreds of thousands of loans or even thousands of of individual loans would take months or years. So it's not practical. So one methodology of data analysis that plaintiffs used was to enlist experts in sampling to come up with a statistically reliable method of aggregating information about the loans and describing. Um. Uh, just. So that we are able to describe wrongdoing about a huge set of mortgage loans by only looking at a small subset of mortgage loans. So, for example, the expert might say that you can come up with a reliable analysis by looking at only 300 mortgages in every trust. And so rather than having to look at 10,000 mortgages, you're only looking at 300. And that streamlines the case because for every one of those individual loans, if you're going to quibble with the numbers, you're going to have to look at the original mortgage. You're going to have to pull out that information and you're going to have to have a contractor, um, like a retired mortgage broker, um, or retired mortgage underwriter. You're going to have to come back and look at those files. And that can take a lot of time for each one of these loans. So if you can cut down the total number of loans you have to re-underwrite, that can dramatically save money. And because statistical sampling was so important as a way to shortcut the analysis, plaintiffs would often. Uh, try to get this issue teed up early in discovery, and they would ask for special dispensation to brief this issue of statistical sampling and expert analysis related to it early in in the case rather than after effect discovery. So this is the case. The decision that I've cited here, the Countrywide. This was another manual case with Justice Branston. Um. New York State Court, where the justice granted a motion in limine on sampling. What I had sought was to be allowed to use and present evidence for its case through statistical sampling, i.e. to present evidence about misrepresentations of the individual mortgage loans through this aggregated means of using sampling. The court found that the use of sampling is widespread as a valid method to prove cases with large amounts of underlying data. So that was a huge win in that case because it meant that NBA had a much easier job of proving its case than it would have otherwise. Whoo! All right, so last slide about, um, uh, just as a testament to how, um, these cases were and how data allowed for that success by getting cases past the pleading stage and getting them, um, using sampling to streamline the cases. Those are just two, two ways in which these cases, uh, did well. Um, but. The end result is that plaintiffs recovered billions of dollars on their losses. So I worked mostly on Allstate and Prudential, but I've listed here some other examples. Um. The DOJ recovered more than 10 billion with settlements with individual banks. Quinn Emanuel working with the FHFa on behalf of Fannie and Freddie as as as their receiver recovered over $20 billion. So these cases were hugely successful, thanks in no large part to to data analysis. All right. In our remaining time in this hour, 12 minutes, I'm going to talk about opioid litigation. So litigation involved market wide problems with the market for mortgage backed securities where you have market wide losses. And the claims were huge, um, particularly for Fannie and Freddie, which bought millions of mortgage loans. So that's why their damages were so huge because the scope of their cases was so huge. Um, opioids litigation is just as big if not bigger. Um, the, the damages settlements in these cases are. $50 billion in total across all plaintiffs and all defendants. So, you know, the scale of these opioids cases is just off the charts because the opioid epidemic is a nationwide problem. Every municipality, every state, every county, every city has. Um, and we've seen thousands in every state in the country. Many, most are in federal court. The defendants are every type of company involved in the market for prescription opioids. So that's manufacturers of drugs like OxyContin. Um, so, you know, Purdue and Teva and Allergan, companies like that. Um, it's distributors of pharmaceuticals, including prescription opioids. So that's AmerisourceBergen and Cardinal Health. And McKesson and then pharmacies, which are the retail distributors such as CVS and Walmart. Um, and then even consulting companies like McKinsey that help strategy for marketing prescription opioids. So the claims are nationwide. The defendants involve every link in the distribution chain. Um. Uh, there is a federal MDL which motley rice my firm is co-lead in. Um, where the federal cases have been aggregated. I was fortunate enough to work on several of the bellwether cases, and I went to trial in West Virginia, in Charleston, for a bellwether case on behalf of an individual county and city in western West Virginia that have been hit particularly hard by the epidemic. So the claim in these cases, the primary claim is public nuisance, where the allegation is that all of these defendants have contributed to the opioid epidemic. The epidemic is the public nuisance. Um, and these companies, the allegation is, um, contributed to the epidemic by oversupplying opioids, ignoring concerns with addiction and overdose, by deceptively marketing the opioid opioids and playing down their addictiveness and their harms. Um, and, you know, the, the harms that we talk about are the epidemic of overdoses and addiction and all the, the secondary harms that come with that, including homelessness and loss of jobs and tearing apart the fabric of society, really. Um, and then the remedy in these cases is the remedy in a public nuisance case is to abate the nuisance which we interpreted as plaintiffs in these cases to mean coming up with a abatement plan, which is a a broad set of. Social services programs that would be needed to mitigate the harms that the epidemic. So needless to say, these extremely broad opioids cases have involved a lot of data. You're talking about population level cases. That require population level proof. And the way to do that in an efficient way is to use data mean. These cases involve individual overdoses. You know individual people that that overdose on on opioids or individual people that become addicted to opioids or family members that are adversely affected by their loved ones being addicted. But just like with R&B, it would be basically impossible to prove these cases on a one on one one by one level. So instead we aggregate the information so that we're talking about society as a whole and the plaintiffs as a whole. You know, whether that's a state or a county or a tribe or a city. So some categories of data that we've used in these cases, um, the amount of marketing to doctors such as payments to doctors by pharma companies to get. The volume of prescription opioids sold and shipped to these different jurisdictions, rates of overdoses. Addiction rates. Disease rates. The cost of programs recommended to abate the epidemic, such as inpatient addiction treatment and outpatient treatment. All of these involve experts. All of these involve lots of spreadsheets, lots of numbers, lots of lots of number crunching. Um, so different types of experts that we've used. Epidemiology. That's a big one. We need them to describe the harms. Uh, describe how, uh, you know, the link between prescription opioids and illicit opioids such as heroin and fentanyl. Um, we need those epidemiologists to come up with the abatement plans for what programs are needed to reduce the epidemic. We have economists who describe the harms and the costs of abatement. We have experts in the history of opioids, the marketing of opioids who both use data to describe that. Um. And then we have people who sort of process and describe data from from the government and other sources. So this is a demonstrative from the West Virginia trial that I mentioned that I worked on. This is a public demonstrative that was used in trial. That is an illustration of how we try to use, um, uh, demonstratives in this case, a chart to aggregate a bunch of data in a way that's easy to understand for everyone. You know, the judge, if there had been a jury, uh, for the jury, you know, for members of the public, for people writing news articles about this, we want everyone to be able to understand this information and important information. So we use a colorful chart to to bring together, in this case, information about different types of opioid overdose deaths in this county. This is, um, the summary spreadsheet. It's, it's super dense and you can't really read it. But this, believe it or not, is the most simple summary spreadsheet, uh, page that we introduced in a slideshow for our economists in that case, which shows on the left each of the programs that our epidemiologists came up with, um, to, um, to which he recommended to abate the epidemic. And then the rest of the spreadsheet is the cost of each of those programs year by year over a 15 year time period. So even though this is a bunch of numbers. Um, nonetheless, this was kind of the best we could do because of the, the amount of information and the complexity. This is the best we could do to summarize, um, the economists analysis, and this is just a small expert excerpt from an Excel spreadsheet that the economists created, which has more than 100 tabs. So we only showed one tab here. So Judge Polster, um, from the Northern District of, of Ohio who's overseeing the MDL, um, he denied Daubert motion seeking to exclude our expert's testimony about various data related topics. He found that. The arguments that defendants raised on Daubert motions were better left for trial. So this was a huge win for us to be, um. To be able to get past this. Stage and have our data analysis be used at trial. Um, conversely, we in the MDL bellwether that I worked on in West Virginia, we actually had a bad outcome where Judge Faber, the judge overseeing that case, um, ruled against us. Um. On a judgment following trial. But the silver lining is that my work involving the epidemiologist. Uh, testimony about overdoses and addiction rates. That information was accepted, but he rejected the overall theory. Unfortunately. Um, All right. And then last slide about opioids. Is that just as with where we saw huge recoveries, opioids is also had huge recoveries in no small part because we've been able to plead and prove those cases on a broad level. So in opioid mean sorry, in. Uh, plead and prove the cases as to entire trusts involving thousands of mortgage loans. And in the opioids cases, we've been able to plead and prove the cases, um, on a society wide level. So that has led to these huge settlements. Um, some of them are listed here. They're all in the billions. Um, except for McKinsey, because each one of these settlements is with one or more defendants across all of all of the cases nationwide or almost all the cases. So you see huge, huge settlements. Um, thanks. In no small part to data analysis. And that is the end of my presentation for part one, and I will look forward to seeing you in part two of this discussion of the use of data in litigation. Thank you very much.