Smilovits v. First Solar Incorporated

Opinion

WO

Mark Smilovits, individually and on behalf of No. CV12-0555-PHX-DGC all others similarly situated, ORDER Plaintiffs, v. First Solar, Inc.; Michael J. Ahearn; Robert J. Gillette; Mark R. Widmar; Jens Meyerhoff; James Zhu; Bruce Sohn; and David Eaglesham, Defendants.

This securities fraud class action is set for trial in January 2020. The parties filed nine motions to exclude expert testimony under Federal Rule of Evidence 702 and Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S. 579 (1993). The Court ruled on six of the motions in an earlier order. Doc. 674. This order addresses the remaining motions. I. Background. Defendant First Solar, Inc. produces photovoltaic solar panel modules. Its stock is publicly traded on the NASDAQ stock exchange. Plaintiffs purchased First Solar stock between April 30, 2008 and February 28, 2012 (the “Class Period”). See Doc. 171 at 22.1

1 The parties have filed multiple motions to file documents under seal, which the Court will address in a separate order. See Docs. 468, 511, 551, 589, 639, 649. The parties have lodged the proposed sealed versions of the documents on the Court’s electronic case filing (“ECF”) system pending rulings on the motions to seal, and have also filed redacted public versions of the documents on the ECF system. See LRCiv 5.6(b)-(c). Where The Individual Defendants are First Solar officers and executives who purchased or sold First Solar stock during the Class Period while allegedly concealing information from the market about manufacturing and design defects causing faster power loss in certain modules.2 Steep declines in First Solar’s stock price, beginning on July 29, 2010, followed the departure of First Solar’s CEO, disappointing financial results, and the release of quarterly financial disclosures reporting the product defects. See Doc. 401 at 7-8. First Solar’s stock fell from nearly $300 per share to less than $50 per share during the Class Period. See id. at 2. Plaintiffs allege that Defendants engaged in several acts of fraud during the Class Period, including concealing the product defects, misrepresenting the cost and scope of the defects, and reporting false information on financial statements. Doc. 93 at 7-20. Plaintiffs further allege that when First Solar later disclosed the product defects and attendant financial liabilities to the market, the stock price fell, causing economic loss to Plaintiffs. Id. at 124-41. Plaintiffs assert violations of §§ 10(b) and 20(a) of the Securities Exchange Act of 1934 and Securities Exchange Commission (“SEC”) Rule 10b-5. Id. at 135-36. Section 10(b) “makes it unlawful to ‘use or employ, in connection with the purchase or sale of any security[,] . . . any manipulative or deceptive device or contrivance in contravention of such rules and regulations as the [SEC] may prescribe.’” In re Oracle Corp. Sec. Litig., 627 F.3d 376, 387 (9th Cir. 2010) (quoting 15 U.S.C. § 78j(b)). Rule 10b-5 forbids, in connection with the purchase or sale of a security, “the making of any ‘untrue statement of a material fact’ or the omission of any material fact ‘necessary in appropriate, and as noted below, this order cites to some of the lodged documents, which will be filed on the ECF system (either under seal or in the public record) once the Court rules on the motions to seal. Citations to pages of a lodged document correspond to the same pages in the filed public copy (the documents are the same except for redactions). Citations are to page numbers added at the top of each page by the ECF system. 2 The modules affected by the manufacturing defect are referred to as “low power modules” or “LPM,” and the design defect is termed “hot climate degradation.” See id. at 3-6. order to make the statements made . . . not misleading.’” Dura Pharms., Inc. v. Broudo, 544 U.S. 336, 341 (2005) (quoting 17 C.F.R. § 240.10b-5). The scope of Rule 10b-5 is coextensive with that of § 10(b). See Oracle, 627 F.3d at 387; Stoneridge Inv. Partners, LLC v. Sci.-Atlanta, Inc., 552 U.S. 148, 157 (2008). To establish a violation of § 10(b) and Rule 10b-5, “a plaintiff must prove (1) a material misrepresentation or omission by the defendant; (2) scienter; (3) a connection between the misrepresentation or omission and the purchase or sale of a security; (4) reliance upon the misrepresentation or omission; (5) economic loss; and (6) loss causation.” Stoneridge, 552 U.S. at 157. Plaintiffs claim that the Individual Defendants are liable for the alleged § 10(b) and Rule 10b-5 violations as “controlling persons” under § 20(a). Doc. 93 at 136, ¶¶ 257-58. To establish a claim under § 20(a), a plaintiff must first prove a primary violation of § 10(b) or Rule 10b-5 and then show that the defendant exercised actual power over the primary violator. See In re NVIDIA Corp. Secs. Litig., 768 F.3d 1046, 1052 (9th Cir. 2014).3 II. Rule 702 and Daubert Standards. Under Rule 702, an expert may offer “scientific, technical, or other specialized knowledge” if it “will assist the trier of fact to understand the evidence,” provided the testimony rests on “sufficient facts or data” and “reliable principles and methods,” and “the witness has reliably applied the principles and methods to the facts of the case.” Fed. R. Evid. 702(a)-(d). The proponent of expert testimony has the ultimate burden of showing, by a preponderance of the evidence, that the proposed testimony is admissible under Rule 702. See Cooper v. Brown, 510 F.3d 870, 942 (9th Cir. 2007); Fed. R. Evid. 104(a). The trial court acts as a gatekeeper for expert testimony to assure that it “both rests on a reliable foundation and is relevant to the task at hand.” Daubert, 509 U.S. at 597; see Davis v. McKesson Corp., No. CV-18-1157-PHX-DGC, 2019 WL 3532179, at *3-4 (D. Ariz. Aug. 2, 2019). 3 The Court has precluded Plaintiffs from asserting a new insider trading claim under § 10(b). Doc. 661 at 19-21. Evidence of insider trading by the Individual Defendants, however, may be relevant to the scienter element of the § 10(b) and Rule 10b-5 misrepresentation claims and perhaps for other purposes. See id. at 21; Doc. 401 at 7. III. Defendants’ Motion to Exclude Henderson’s Testimony (Docs. 532, 633). Plaintiffs retained M. Todd Henderson, a professor at the University of Chicago Law School, to opine on the sales of company stock by the Individual Defendants pursuant to Rule 10b5-1 trading plans (“10b5-1 plans”). Doc. 532-1 at 3. Henderson explains that 10b5-1 plans are meant to provide company executives with the ability to trade for diversification reasons and not on the basis of inside information. Id. at 4. Henderson opines that there are seven “hallmarks” of a valid 10b5-1 plan: (1) a significant time lag between the plan’s adoption and the first trade; (2) regularly scheduled trades; (3) sales over long periods of time; (4) sales of relatively consistent size and relatively small amounts compared with overall shareholdings; (5) sales at a range of stock prices; (6) no changes or cancellations of the plan; and (7) a high level of board and general counsel involvement in plan adoption, alteration, and termination decisions. Id. Henderson further opines the Individual Defendants’ 10b5-1 plans during the Class Period contained few or none of these hallmarks and are consistent with plans from his empirical research in which insiders earned large abnormal returns based on inside information. Id. Defendants move to exclude Henderson’s testimony under Rule 702, arguing that his opinions and purported hallmarks are based solely on speculation and fail every prong of Daubert’s reliability test. Docs. 532 (lodged sealed motion), 633 (public redacted version). Plaintiffs counter that Henderson is eminently qualified to offer his opinions, that the Daubert reliability factors do not apply to the non-scientific hallmarks, and that the hallmarks satisfy Rule 702’s reliability requirements and will assist the jury in deciding whether the Individual Defendants’ stock trades were made in good faith or to evade the prohibitions of Rule 10b5-1. Doc. 556 at 6-17. A. 10b5-1 Plans and the Need for Expert Testimony. Rule 10b5-1 provides an affirmative defense to securities fraud allegations of trading “‘on the basis of’ material nonpublic information[.]” 17 C.F.R. § 240.10b5-1. A corporate-insider may defend his trades by showing that they were made pursuant to a “written plan for trading securities” created before he learned of the material non-public information at issue in the suit. Id. at (c)(1)(i)(A). To be valid, the plan must list the amount and price of the securities to be sold or purchased, provide the date of the transactions, and prevent the insider from exercising any subsequent influence over how, when, or whether to effect purchases or sales. Id. at (c)(1)(i)(B). The trade must also have been executed according to the plan’s terms. Id. (c)(1)(i)(C). Finally, the plan must have been entered “in good faith and not as part of a plan or scheme to evade the prohibitions of [Rule 10b5-1].” Id. at (c)(1)(ii); see CSX Corp. v. Children’s Inv. Fund Mgmt. (UK) LLP, 654 F.3d 276, 303 n.13 (2d Cir. 2011) (“[T]he affirmative defense is available only when the plan to purchase or sell securities was ‘given or entered into in good faith.’”); Elec. Workers Pension Tr. Fund of IBEW Local Union No. 58 v. CommScope, Inc., No. 5:10- CV-00062-RLV, 2013 WL 4014978, at *7 (W.D.N.C. Aug. 6, 2013) (“[Rule 10b5-1] clearly requires a showing of good faith [that] . . . require[s] additional evidence to be presented by Defendants[.]”); Stocke v. Shuffle Master, Inc., 615 F. Supp. 2d 1180, 1193 (D. Nev. 2009) (“[A Rule] 10b5-1 trading plan does not provide an absolute defense to a claim of insider trading. Rather, it requires an additional factual finding of good faith.”). A plaintiff may present circumstantial evidence giving rise to an “inference of[] bad faith and scienter for § 10(b) and Rule 10b-5 purposes” by showing that the corporate-insider’s trades under a 10b5-1 plan are “unusual” or “suspicious.” In re Enron Corp. Sec., Derivative & ERISA Litig., 258 F. Supp. 2d 576, 593 (S.D. Tex. 2003). The Individual Defendants claim that their stock trades during the Class Period were proper because they were made largely pursuant to valid 10b5-1 plans. See Doc. 311 at 65-68. Plaintiffs counter that the Individual Defendants engaged in improper insider trading, and that their suspicious trading behavior is relevant evidence of scienter and materiality. Doc. 556 at 10. In order for the jury to make an informed analysis of the relevant trades, it will be helpful for them to understand the purpose of 10b5-1 plans and their legal requirements, the various features of such plans in practice, and trading behaviors that may be suspicious. See Doc. 556 at 10-11. Henderson explains that the purpose of his opinions, and his application of the seven hallmarks he identifies for valid 10b5-1 plans, is to help the jury understand such plans and determine whether the Individual Defendants’ trading behaviors are suspicious. Doc. 532-3 at 6-7. Because 10b5-1 plans generally are beyond the ken of the average juror, Henderson’s expert testimony on these issues will assist the jury in understanding the evidence and determining facts in issue. See In re Novatel Wireless Sec. Litig., No. 08CV1689 AJB RBB, 2012 WL 5463214, at *4 (S.D. Cal. Nov. 8, 2012) (rejecting a Rule 403 challenge to “Henderson’s opinion setting forth the ‘seven hallmarks’ [because it] describes for the jury the best practices when employing 10b5-1 trading plans”); United States v. Robinson, No. 98 CR 167 DLC, 2000 WL 65239, at *3 (S.D.N.Y. Jan. 26, 2000) (expert testimony on the “‘hallmarks’ of financial fraud schemes” would help the jury understand the evidence); United States v. Jergensen, No. 8:16-CR- 235(BKS), 2017 WL 11458077, at *4 (N.D.N.Y. Oct. 6, 2017) (same). B. The Reliability of Henderson’s Hallmarks. Defendants contend that Henderson’s hallmarks are not reliable because they cannot be tested, have never been published, have no known error rate, and lack general acceptance in the securities regulation field. Doc. 532 at 9-11. But these factors are neither exclusive nor dispositive in a Rule 702 inquiry, see Daubert, 509 U.S. at 593-94, and “may not be pertinent in assessing reliability, depending on the nature of the issue, the expert’s particular expertise, and the subject of his testimony[,]” Primiano, 598 F.3d at 565. As the Supreme Court has explained, although some expert testimony “rests upon scientific foundations,” in other cases “the relevant reliability concerns may focus upon personal knowledge or experience. Daubert makes clear that the factors it mentions do not constitute a definitive checklist or test.” Kumho Tire Co. v. Carmichael, 526 U.S. 137, 150 (1999) (citations omitted; emphasis in original). The Ninth Circuit likewise holds that “[t]he Daubert factors (peer review, publication, potential error rate, etc.) simply are not applicable to [testimony] whose reliability depends heavily on the knowledge and experience of the expert, rather than the methodology or theory behind it.” United States v. Hankey, 203 F.3d 1160, 1169 (9th Cir. 2000). Defendants do not dispute that Henderson is a highly qualified expert in securities regulation. See Doc. 532-1 at 6-7 (Henderson’s qualifications); In re Novatel, 2012 WL 5463214, at *4 (“Henderson is qualified pursuant to Rule 702 to testify as an expert regarding the use of 10b5-1 trading plans.”). Henderson’s opinions are based on a review of the documents in this case, his position on the Financial Industry Regulatory Authority, his experience as a management consultant advising companies on security regulation matters, his extensive research on thousands of 10b5-1 plans used by various companies over the past decade, and his specialized knowledge in the areas of corporate governance, executive compensation, and insider trading. Id. at 3. Under Rule 702 and Daubert, expert testimony “is reliable if the knowledge underlying it has a reliable basis in the knowledge and experience of the relevant discipline.” Primiano, 598 F.3d at 565 (citation omitted). The Court finds that Henderson’s experience in the field of securities regulation generally, and his knowledge about the creation and use of 10b5-1 plans in particular, form a sufficient foundation for his opinions on the purported hallmarks of a valid plan. See In re Novatel, 2012 WL 5463214, at *5 (“Henderson’s training, education and practical experience provide a sufficient basis for his opinions regarding 10b5-1 trading plans”); United States v. Barreiro, No. 13-CR-00636-LHK, 2016 WL 259590, at *2 (N.D. Cal. Jan. 21, 2016) (testimony on whether “notes of expenditures have the hallmarks of business or personal expenditures” was within the expert’s “specialized knowledge and training as a CPA and a certified fraud examiner”); S.E.C. v. Lines, No. 07 CIV. 11387 (DLC), 2010 WL 2926591, at *1 (S.D.N.Y. July 26, 2010) (rejecting the argument that the expert identified no reliable methodology to support his description of the common characteristics of a “pump and dump scheme” where the expert was “well qualified by years of experience as a regulator to describe this species of securities fraud, and [gave] an extensive and detailed report explaining the hallmarks of such a scheme”). Defendants’ criticisms of Henderson’s methodology and his purported hallmarks for valid 10b5-1 plans are matters for the jury’s consideration in weighing the evidence. See Lines, 2010 WL 2926591, at *1 (the arguments that “many of the characteristics that the expert identifies as associated with pump and dump schemes can also be found in legitimate transactions, and [that] he did not adequately consider factors that tend to show that a transaction is not a pump and dump scheme, may be proper as lines of cross-examination but do not render the testimony inadmissible”). Defendants note that “[t]he first time [the] hallmarks appeared in a bulleted list was for an expert opinion on behalf of Plaintiffs[.]” Doc. 532 at 10. But Henderson explains that “the general impression of [the hallmarks], each of them in different forms or different words[,] was used by [him] countless times before then[.]” Doc. 532-3 at 16. The Court cannot conclude that Henderson developed the hallmarks solely for purposes of litigation. Defendants further contend that Henderson’s opinions are replete with improper legal conclusions and that his hallmarks are contrary to law because Rule 10b5-1 offers an affirmative defense for individuals who comply with the Rule’s three express legal requirements. Doc. 532 at 12, 14-16. But to avail themselves of the affirmative defense, Defendants must show that their 10b5-1 plans were entered “in good faith and not as part of a plan or scheme to evade the prohibitions of [Rule 10b5-1].” 17 C.F.R. § 240.10b5-1(c)(1)(ii). Henderson’s opinions about 10b5-1 plans generally, his purported hallmarks of valid plans, and whether the Individual Defendants’ trading behavior is suspicious are relevant to the affirmative defense. Henderson states that he will offer no legal opinion on the Individual Defendants’ trading plans and behavior or any other issue in the case. Doc. 532-3 at 6, 19-21. Defendants may object at trial if they believe Henderson crosses the line into inadmissible legal opinions. The same is true with respect to any potential opinions on the mental state or credibility of witnesses, which would be improper. See Doc. 532 at 17-18. C. Henderson’s Analysis of Empirical Studies. Henderson discusses empirical studies purportedly showing that “insiders systematically abuse [Rule 10b5-1].” Doc. 532-1 at 5; see id. at 36-43 (citing M. Todd Henderson, Alan D. Jagolinzer & Karl A. Muller, Offensive Disclosure: How Voluntary Disclosure Can Increase Returns from Insider Trading, 103 Geo L.J. 1275 (2015); Jagolinzer, Alan D., SEC Rule 10b5-1 and Insiders’ Strategic Trade, Management Science Quarterly (Feb. 2009)). Henderson opines that “[t]he characteristics of the Individual Defendants’ trades under their plans are consistent with plans that this empirical research shows are abusing the Rule and trying to shield trades that are based on material, nonpublic information.” Id. at 5. Defendants argue that Henderson does not reliably apply the empirical studies to the facts of this case. Doc. 532 at 13-14. Plaintiffs do not meaningfully respond to this argument, claiming instead that Henderson’s general discussion of the empirical studies will assist the jury in understanding why scrutiny of Defendants’ Rule 10b5-1 plans is appropriate. Doc. 556 at 17-18. More importantly, Henderson made clear in his deposition that he has not “taken the trading behavior of the First Solar executives and compared them to other companies in an empirical study.” Doc. 532-3 at 23. Plaintiffs have not demonstrated the reliability of Henderson’s opinion that the characteristics of the Individual Defendants’ 10b5-1 trades are consistent with abusive plans in the empirical studies. Henderson is precluded from offering this opinion at trial. Defendants’ motion to exclude Henderson’s testimony (Docs. 532, 633) is granted in part and denied in part as set forth above. IV. Defendants’ Motion to Exclude Dr. Feinstein’s Testimony (Docs. 552, 632). Plaintiffs retained Dr. Steven Feinstein to opine about market efficiency, loss causation, and damages. Based in part on an “event study,” Dr. Feinstein concludes that First Solar’s alleged fraud caused its stock price to be artificially inflated and that corrective disclosures caused stock price declines resulting in investor losses. Doc. 545-1 at 9. An event study is a statistical analysis used to “‘disentangle[] the effects of two types of information on stock prices – information that is specific to the firm under question . . . and information that is likely to affect stock prices marketwide.’” In re Vivendi, S.A. Sec. Litig., 838 F.3d 223, 253 (2d Cir. 2016) (quoting Mark L. Mitchell & Jeffry M. Netter, The Role of Financial Economics in Securities Fraud Cases: Applications at the Securities and Exchange Commission, 49 Bus. Law. 545, 556-57 (1994)). The study includes a regression analysis of daily stock returns to determine the typical level of stock price volatility and the relationship between a company’s stock price movements and market and sector index movements. It allows an expert “to make inferences about the degree to which the company’s stock price may have been artificially inflated on the basis of the market’s misconception as to the truth prior to the release of that information.” Id. at 253-54. Event studies are widely accepted and have been characterized as “standard operating procedure in federal securities litigation.” Id. at 253; see also FindWhat Inv’r Grp. v. FindWhat.com, 658 F.3d 1282, 1313 n.31 (11th Cir. 2011) (“The methodology of event studies has been sustained by many circuits.”). Defendants do not contend than an event study is improper methodology. In fact, their own expert uses an event study as discussed below. Defendants move to exclude Dr. Feinstein’s testimony because he fails to evaluate when inflation entered First Solar’s stock price and analyze how much of the price decline is attributable to the alleged fraud. Docs. 545 (lodged), 632 (public). A. Stock Price Inflation. Dr. Feinstein describes artificial inflation as “the difference between what the stock price actually was at a point in time, and what the price would have been absent the alleged fraud.” Doc. 545-1 ¶ 232. To calculate artificial inflation in this case, Dr. Feinstein worked chronologically backwards from the final corrective disclosure and cumulated the fraud- related residual stock price declines. Id. ¶ 234. Dr. Feinstein opines that (1) all price declines attributable to corrective information regarding the low power module issue reflect inflation that was in place at the start of the Class Period (the date of the first alleged misrepresentation regarding that issue), and (2) all price declines attributable to corrective information regarding heat degradation reflect inflation that was in place on February 22, 2010 (the date of the first alleged heat degradation misrepresentation). Id. ¶ 235. He concludes that “the artificial inflation in First Solar’s stock price at the start of the Class Period was at least $17.93 per share, measured using the stock price declines that occurred when the corrective disclosures were made regarding the LPM issues,” and that “the heat degradation-related artificial inflation of $11.48 per share that dissipated following the corrective disclosures . . . represented artificial inflation that was in the stock as of 22 February 2010 at the latest.” Id. ¶¶ 237, 243; see id. at 389 (graph showing LPM and heat degradation inflation during the Class Period). Defendants contend that the premise of Dr. Feinstein’s damages analysis – that the entire inflation for both LPM and heat degradation was in place at the outset – contradicts Plaintiffs’ theory of liability. Doc. 545 at 10-11. Defendants assert that Dr. D. Paul Regan, whom Defendants characterized as Plaintiffs’ “liability expert,” offers opinions fundamentally at odds with Dr. Feinstein’s analysis because Regan opines that First Solar learned of the LPM and hot climate problems incrementally over time and should have made incremental accounting disclosures to reflect its growing knowledge. If the problems were understood only incrementally, Plaintiffs suggest, they could have been disclosed only incrementally, and, it follows, their disclosure would have affected stock prices only incrementally. As a result, Defendants’ alleged nondisclosure likewise could have had only an incremental effect on any fraud-related inflation in the value of First Solar stock, and all such inflation would not have been in place from the beginning of the Class Period. Defendants argue, therefore, that Feinstein’s “constant dollar inflation ribbon” is unreliable, results in overstated damages early in the class period, and does not correspond with Plaintiffs’ theory of liability. Id. at 14-16.4 Defendants have a point. The Second Circuit has described the problem of assuming that fraud-related inflation ultimately eliminated from a stock’s price by a series of truthful disclosures was all present on the first day of the class period: A key question was how that inflation entered the stock or, more aptly, when. Given that the maximum amount of inflation in the stock was €22.52, one approach to determining how inflated the stock price was throughout the Class Period would have been to say that all €22.52 of inflation entered into Vivendi’s stock price from the very beginning of that period, on October 30,

4 An inflation ribbon is a time series indicating how much artificial inflation caused by the alleged fraud was in the stock price on each day of the Class Period. Doc. 545-1 at 8. 2000, and remained at that level until the date of the first negative residual return, January 7, 2002. There is an obvious downside to this approach, however. Namely, the full amount of the inflation reflects the value of the truth about Vivendi’s liquidity problem at the apex of that problem. But the magnitude of Vivendi's liquidity risk—and by extension, the amount of liquidity-related inflation in Vivendi’s stock—presumably had not reached its peak at the start of the Class Period. Rather, it grew over time as Vivendi’s liquidity situation worsened, and as the distance between the truth and the deception thus widened. Ascribing the full value amount of loss to the very first alleged misstatement would therefore tend to overstate the degree to which Vivendi’s stock was inflated due to the market’s lack of knowledge about Vivendi’s true liquidity risk, at least toward the beginning of the Class Period. Such an approach might thus lead to an inflated recovery for class members who purchased the stock earlier in the Class Period. In re Vivendi, 838 F.3d at 255 (emphasis in original). Here’s another way to describe the issue. Assume that on the first day of a class period a company’s stock is selling for $100 per share, the company’s product has a serious defect which is unknown inside and outside of the company, and, if the defect was known publicly, the stock price would drop $30 per share. Although the price of the stock can be said to be inflated on that date, it is not inflated because of any fraudulent actions of the part of the company – the company does not know of the defect. A person buying the stock on this date could not claim to have been defrauded in the amount of $30 because no false statements have been made. Assume a week later the company gains the first intimation of a defect in its product, the full extent of the defect and its eventual cost to the company are not understood, and, because the information is incomplete and does not reflect the full extent of the actual problem, even full disclosure of the known information would not cause the stock to drop $30 per share. If full disclosure was made, the market would learn of a potential problem and would take the potential problem into account in the stock price, but would not likely reflect the full $30 extent of the problem because the full extent is not known. If disclosure of this incomplete information would cause the stock price to drop, say, by $10 per share, then a fraudulent nondisclosure of the information would result in $10 of inflation remaining in the stock due to fraud – the fraud-related inflation would be $10 per share. There would remain another $20 per share of inflation in the stock price due to the undiscovered extent of the defect, but it could not be claimed as fraud damages by a person purchasing on the date of the nondisclosure because it did not remain in the stock due to nondisclosure. If more was learned by the company but not revealed on later dates, the fraud-related component would increase on each of those dates by the amount the stock would have dropped if the information was disclosed, until the full $30 becomes fraud-related inflation. Defendant argue that Plaintiffs’ own accounting expert, Dr. Regan, opines that the extent of the LPM and hot climate defects became known to First Solar only gradually over time, and that GAAP therefore required First Solar to make incremental and increasing accounting disclosures over time. Defendants argue that this opinion conflicts with Feinstein’s conclusion that the full extent of the fraud-related inflation was in First Solar stock from the start. The Court views this as a fair criticism of Dr. Feinstein’s opinion. The question is whether it renders his opinion inadmissible under Rule 702. Defendants do not challenge Dr. Feinstein’s qualifications. He is a Yale Ph.D. economist, a professor of finance at Boston College, has authored numerous publications in the fields of economics and finance, and has testified in numerous securities fraud cases. Doc. 552 at 8. Nor do Defendants challenge the principles and methods he applies – “event studies or backcasting generally.” Doc. 654 at 9. The question appears to be whether he has applied the principles and methods reliably to the facts of this case. Fed. R. Evid. 702(d). Plaintiffs respond, in part, by rejecting Defendants’ premise concerning the nature of their fraud case. They contend that Dr. Regan is not their liability expert and that his incremental accounting opinion does not reflect the liability case they will present to the jury. They contend that Dr. Regan’s role is to explain why First Solar’s accounting was defective and failed to reveal the fraud in which Defendants were engaged. Plaintiffs assert that their fraud case will consist of evidence that Defendants had enough information to disclose the LPM defect on the first day of the Class Period and the heat degradation defect on February 22, 2010. Doc. 552 at 11. Feinstein assumes Plaintiffs will be able to provide this proof, and concludes that such disclosures would have lowered the valuation of First Solar stock to the same levels that resulted from later disclosures. Id. at 11-12. Plaintiffs also argue that their theory of the case is that the mere existence of the defects, not the extent of the defects, was enough to cause the full economic damages they claim. Id. at 14. Feinstein opines that the effect of unfavorable disclosures would have been greater earlier in the Class Period – when the value of the stock was higher and the prospects for the company were brighter – than in the later periods when more difficulties had arisen in the industry. As Feinstein states: “The valuation impact of the LPM issues would have been larger at the start of the Class Period on account of, among other things, the higher expected earnings and growth prospects for the Company at that time. . . . Moreover, had the market been aware of management’s alleged intent to mislead investors, a discount for reputation effect would have been impounded in the stock price at the start of the Class Period.” Doc. 545-1 ¶ 239.5 Finally, citing two scholarly articles,6 Feinstein also asserts that “[b]ased on the fundamental principles of market efficiency and rational expectations, because investors generally do not make systematic errors in forecasting when they are equipped with full truthful information, one can estimate that investors would have centered their forecasts on a valuation that reflected the ultimate outcome.” Doc. 469-3, ¶ 238.7 The Court’s decision is guided by two principles. First, the Court cannot exclude Plaintiffs’ expert by concluding that Defendants’ factual recitation concerning the

5 The Court does not find this last point very persuasive because if Defendants had made a full disclosure of what they knew from the start, they would not have been acting fraudulently and the effect on First Solar’s stock value simply would have been the value of the information they disclosed, not some additional fraud premium. 6 “Rational Expectations and the Theory of Price Movements,” by John Muth, Econometrica, 1961; and “Asset Prices in an Exchange Economy,” by Robert Lucas, Econometrica, 1978. 7 The Court is skeptical of this last argument because Plaintiffs do not appear to contend that First Solar could have disclosed “full truthful information” about the defects on the first day of the Class Period or February 22, 2010. They seem to acknowledge that more became known about the defects over time. unfolding of the First Solar product defects is correct and Plaintiffs’ is incorrect. As the Advisory Committee on the Federal Rules of Evidence noted when it adopted the current version of Rule 702: “The emphasis in the amendment on ‘sufficient facts and data’ is not intended to authorize a trial court to exclude an expert’s testimony on the ground that the court believes one version of the facts and not the other.” Fed. R. Evid. 702, 2000 Advisory Committee Note. Second, the Court’s focus “must be solely on principles and methodology, not on the conclusions that they generate.” Daubert, 509 U.S. at 595; see also Wendell v. GlaxoSmithKline LLC, 858 F.3d 1227, 1232 (9th Cir. 2017) (same). The Court’s task is not to determine whether Dr. Feinstein’s opinions are correct, but whether they satisfy the requirements of Rule 702. In re Bard IVC Filters Prods. Liab. Litig., No. MDL 15-02641- PHX DGC, 2018 WL 495607, at *4 (D. Ariz. Jan. 22, 2018) (“It is not the job of the court to insure that the evidence heard by the jury is error-free, but to insure that it is sufficiently reliable to be considered by the jury.” (internal quotations and citation omitted)); see also Fed R. Evid. 702, 2000 Advisory Committee Note (proponents of evidence “do not have to demonstrate to the judge by a preponderance of the evidence that the assessments of their experts are correct, they only have to demonstrate by a preponderance of the evidence that their opinions are reliable” (citation and quotation marks omitted)). The Court concludes that Defendants’ arguments go more to the correctness of Feinstein’s conclusions than the method by which he reaches them. A well-qualified expert has used an accepted methodology to reach conclusions with which Defendants disagree. Defendants contend that his opinions are contrary to Plaintiffs’ liability case, but the Court must look to Plaintiffs, not Defendants, for the nature of their liability case, and Plaintiffs claim they can prove an extent of knowledge at the beginning of the Class Period and on February 22, 2010, and a sensitivity to misrepresentations at that time, that will support Feinstein’s opinion that the full amount of the fraud-related inflation was in the stock from the start. Given this situation, the Court concludes that the jury must assess the validity of Feinstein’s conclusions. See In re Novatel Wireless Sec. Litig., No. 08CV1689 AJB (RBB), 2013 WL 12247558, at *3 & n.3 (S.D. Cal. Nov. 19, 2013) (finding that the two alternative calculations for calculating damages – the constant dollar and constant percentage methods – “may properly be presented to a jury for determination as to which is the more reasonable measurement of damages”). B. Confounding Information. Stock price declines can result from corrective disclosures concerning earlier misrepresentations and various confounding information unrelated to the alleged fraud, such as “changed economic circumstances, changed investor expectations, new industry- specific or firm-specific facts, conditions, or other events[.]” Dura Pharms., 544 U.S. at 343. As part of his event study, and for each corrective disclosure he identifies, Feinstein considers and accounts for potentially confounding information. Doc. 545-1 ¶¶ 20, 39, 161-231. Defendants argue that Feinstein’s opinions on loss causation and damages are unreliable because his approach to confounding information is arbitrary and inconsistent. Docs. 545 at 17-20, 654 at 11. Feinstein’s methodology – conducting an event study that purports to account for confounding information – meets Rule 702’s admissibility threshold. See In re Vivendi, 838 F.3d at 253; Angley v. UTi Worldwide Inc., 311 F. Supp. 3d 1117, 1129 & n.20 (C.D. Cal. 2018) (collecting cases that “have recognized the event study/out of pocket method . . . is an accepted method for calculating damages in securities fraud class actions”). Defendants’ criticisms that Feinstein discounted the confounding information and exaggerated the impact of the alleged fraud go to his credibility and the weight of his opinions, not their admissibility. See In re Novatel, 2013 WL 12144150, at *8 (“[T]he failure to include one of the relevant [non-fraudulent] variables does not dictate exclusion of the [loss causation] report as unreliable.”); In re Homestore.com, Inc., No. CV 01-11115 RSWL CWX, 2011 WL 291176, at *8 (C.D. Cal. Jan. 25, 2011) (“[The] expert . . . expressly states that her regression analysis distinguishes between loss attributable to alleged fraud and loss attributable to non-fraud related news and events. . . . As a result, [she] makes an attempt to account for other possible causes, and her opinions therefore meet the reliability threshold requirement.”); S.E.C. v. Leslie, No. C 07-3444, 2010 WL 2991038, at *15 (N.D. Cal. July 29, 2010) (“While [defendant] certainly may argue to the jury that [the expert] did not reliably filter out other confounding variables that would have affected the stock price[,] . . . he has not shown that [the] opinion is so unreliable that it must be excluded.”); In re Sci.-Atlanta, Inc. Sec. Litig., No. CIV.A.1:01-CV1950RWS, 2009 WL 4281256, at *3 (N.D. Ga. Nov. 24, 2009) (“[The expert] conducted an event study which is an acceptable method for analyzing loss causation. . . . The failure to disaggregate the fraudulent and non-fraudulent statements does not require exclusion of [his] opinion[.]”).8 Defendants’ motion to exclude Feinstein’s testimony (Docs. 545, 632) is denied. V. Plaintiffs’ Motion to Exclude Dr. Kleidon’s Testimony (Docs. 469, 485). Defendants retained Dr. Allan Kleidon to rebut Dr. Feinstein’s opinions on loss causation and damages. Doc. 469-1 at 6. Plaintiffs move to exclude Dr. Kleidon’s testimony based on the method of his event study and his “peer comparison” opinion. Docs. 469 (lodged), 485 (public). A. Dr. Kleidon’s Event Study. Dr. Kleidon conducted an event study and found no evidence that company-specific information caused statistically significant price declines in First Solar’s stock on the corrective disclosure dates. Doc. 469-1 at 7, 26-75. Dr. Kleidon claims that Dr. Feinstein’s event study is flawed in finding statistically significant declines, in part because Feinstein fails to account for the increased “beta” of First Solar stock on those dates.9 Id. at 7.

8 The cases Defendants cite in their reply (Doc. 654 at 11-12) are distinguishable. See Bricklayers & Trowel Trades Int’l Pension Fund v. Credit Suisse First Bos., 853 F. Supp. 2d 181, 191 (D. Mass. 2012) (“Given . . . the pervasiveness of [the expert’s] methodological errors and the lack of congruity between his theory and the data, the Court is compelled to exercise its role as gatekeeper and to exclude his event study as unreliable.”); In re Omnicom Grp., Inc. Sec. Litig., 541 F. Supp. 2d 546, 554 (S.D.N.Y. 2008) (finding that “there [was] simply no way for a juror to determine whether the alleged fraud caused any portion of Plaintiffs’ loss” where the expert’s “event study at best incorrectly identifies several corrective disclosures and at worst fails to identify any at all”). 9 A “beta” measures stock price volatility – the degree to which a single company’s stock return fluctuates in response to forces that affect the industry or market as a whole. See Docs. 469 at 11, 622 at 5. A stock with a beta above 1.0 is more volatile than the Plaintiffs contend that Dr. Kleidon’s use of a date-specific beta improperly alters the widely-accepted event study methodology by applying a manual adjustment that attributes nearly all of First Solar’s stock price declines on corrective disclosure dates to industry- wide, rather than company-specific, information. Doc. 469 at 5; see Doc. 469-3 at 33. Plaintiffs claim that this adjustment is invalid because it measures “the impact of First Solar’s company-specific information on First Solar’s industry peers’ stock – an inquiry that has no relevance to loss causation and damages in this case – and then treats the result as a measurement of the industry’s impact on First Solar’s stock.” Doc. 469 at 6 (emphasis in original). Plaintiffs have a point. Dr. Kleidon seems to assert that First Solar’s corrective disclosures on the key dates, which revealed negative information about the LPM and hot climate defects, reduced stock prices in the entire solar industry, which resulted in an increased correlation between First Solar stock prices and industry stock prices (because First Solar stock also decreased that day) and therefore increased First Solar’s beta, at least on that date. Applying the higher daily beta, Dr. Kleidon’s event study seems to suggest that the reduction in First Solar stock was a reaction to downward price movement in the solar industry and cannot be attributed to any company-specific factors at a statistically significant level. This seems to be a kind of economic boomerang. First Solar’s corrective disclosures depressed the stocks of other companies, which in turn depressed the price of First Solar stock – at a higher rate than usual due to the increased daily beta – leading to the conclusion that First Solar’s negative disclosures were not the cause of its stock price decline. But it is difficult to understand how a process that begins with allegedly fraud- related disclosures, and eventually reduces First Solar stock, cannot be said to have been caused by the fraud-related disclosures. Plaintiffs argue that Dr. Kleidon uses unreliable methodology to reach this conclusion because event analysis does not use a daily beta as he does, the article Kleidon overall market, while a beta below 1.0 is less volatile. See United States v. Gushlak, No. 03-CV-833 NGG, 2012 WL 1379627, at *4 (E.D.N.Y. Apr. 20, 2012); ASARCO LLC v. Ams. Mining Corp., 396 B.R. 278, 354 n.72 (S.D. Tex. 2008). relies on for his daily beta does not support his approach, the effects of a company’s negative disclosures are unidirectional – they affect prices in the market, not vice versa – and Dr. Kleidon errs in concluding that the increased daily beta he observed was caused by First Solar’s negative information.10 After considering these arguments, the Court concludes that Plaintiffs, like Defendants with respect to Dr. Feinstein, are quarrelling more with the results of Kleidon’s analysis than his methodology. Plaintiffs do not dispute Dr. Kleidon’s qualifications. After receiving a Ph.D. in economics and finance from the University of Chicago, Dr. Kleidon taught at Stanford University and the University of California at Berkeley. Doc. 485-2. He has served as an associate editor for the Journal of Finance and the Journal of Financial Economics, and has conducted event studies and testified about causation and damages in numerous securities actions. Nor do Plaintiffs dispute the reliability of his general methodology – an event study. It is the same method used by Plaintiffs’ expert Dr. Feinstein. Plaintiffs disagree with how Dr. Kleidon applies the event study methodology. They claim that he errs by introducing a daily-calculated beta for First Solar stock rather than using the more general beta adopted by Dr. Feinstein and typically used in event studies. But Plaintiffs cannot dispute that the peer-reviewed article relied on by Dr. Kleidon specifically finds that higher betas occur on dates of significant public disclosures: “Using daily firm-level betas estimated from intraday prices, we find that betas increase on earnings announcement days and revert to their average levels two to five days later.” Andrew J. Patton & Michela Verardo, Does Beta Move with News? Firm-Specific Information Flows and Learning about Profitability, 25 Rev. of Fin. Stud. 2789, 2789 (2012) (“Patton & Verardo”). Nor do Plaintiffs argue that Dr. Kleidon misapplied the technique developed by Patton & Verardo for calculating a daily beta.

10 Plaintiffs discuss Dr. Kleidon’s bellwether opinion, which suggests that First Solar was an industry leader watched by analysts and others, and therefore more likely to affect solar industry stock prices than non-bellwether companies. But their primary complaint is about how this is translated into an increased daily beta as part of his event study. Instead, Plaintiffs’ argument boils down to this: The causation described in the literature on which Kleidon relies is unidirectional – from the Company spilling over to the market. Feinstein Rebuttal, ¶¶78-82. These studies observe that a company’s earnings announcements, which may or may not include information with implications for the market or industry, could influence the market in general. Id. Because of this spillover effect, some researchers have found that there is a greater correlation between a company’s stock price returns and the market index returns on earnings announcement dates. Id. But, correlation does not imply causation, and the literature Kleidon relies on does not suggest that the company is more sensitive to sector news on earnings announcement days. Id. To the contrary, the correlation that the researchers detect is, as they explain, an artifact of the sector reacting to company news, not, as Kleidon claims, an increased sensitivity of the Company to sector news. Id. Doc. 469 at 12. The sole authority Plaintiffs cite for this key argument is their own expert, Dr. Feinstein. And the portion of his report from which this language is directly quoted cites no other authorities. See Doc. 469-3, ¶¶ 80-82. Thus, it is clear that Feinstein disagrees with Kleidon’s application of an event study, just as Kleidon disagrees with Feinstein’s. This is not a basis for inadmissibility. The Court concludes that both experts are well qualified and apply a generally accepted methodology for determining loss causation. Although the Court sees problems in the conclusions of each expert, the Court does not find a basis for excluding their opinions. The Court is satisfied that the testimony of both experts is sufficiently reliable to be admitted under Rule 702. As the Advisory Committee noted when it amended Rule 702 after Daubert: “When a trial court, applying the amendment, rules that an expert’s testimony is reliable, this does not necessarily mean that contradictory expert testimony is unreliable.” Fed. R. Evid., 2000 Advisory Committee Note. B. Dr. Kleidon’s Peer Comparison. The background section of Dr. Kleidon’s report discusses macroeconomic and industry factors that adversely affected the solar industry during the Class Period. Doc. 469-1 at 11-21. Dr. Kleidon intends to testify that “the solar industry was adversely affected by a number of issues during the Class Period, and First Solar’s stock price decline was nearly identical to that of the Industry Peers.” Jd. at 93; see Doc. 622 at 20. Plaintiffs contend that Dr. Kleidon’s peer comparison is inadmissible because it is unreliable, irrelevant to loss causation and damages, and focuses on the wrong timeframe. Docs. 469 at 20-21, 640 at 13-14. Defendants counter that the peer comparison is directly responsive to Plaintiffs’ claims about First Solar’s stock performance and Dr. Feinstein’s opinions, and provides context about First Solar’s business and the environment in which the company operates. Doc. 622 at 20-21. The Court cannot conclude that the peer comparison 1s irrelevant. Plaintiffs contend that First Solar’s stock price drop was due to Defendants’ alleged fraud. Doc. 93 4] 4. And Dr. Feinstein asserts that his damages opinion is conservative because it accounts for only $29.41 of the $280.72 that First Solar declined over the Class Period. Doc. 545-1 § 21. Given these assertions, Dr. Kleidon’s explanation of the solar industry’s general stock price decline during the Class Period is clearly relevant. Plaintiffs claim that Kleidon does not use a defined or accepted methodology for this opinion, but they do not dispute that he is qualified to opine on the causes of stock price declines, and it does not require a scientific method for such an expert to identify relevant comparison stocks and identify their general trend. The Court will not exclude this portion of Dr. Kleidon’s testimony. Plaintiffs may object at trial to any testimony they view as unfounded or irrelevant. Plaintiffs’ motion to exclude Kleidon’s testimony (Docs. 469, 485) is denied. Dated this 27th day of December, 2019. . » punk © Cour petl David G. Campbell Senior United States District Judge

-21-