Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 1 of 14
`
`UNITED STATES DISTRICT COURT
`DISTRICT OF CONNECTICUT
`
`
`IN RE TEVA SECURITIES LITIGATION
`_______________________________________
`
`THIS DOCUMENT RELATES TO:
`
`
`
`No. 3:17-cv-00558 (SRU)
`
`
`No. 3:17-cv-00558 (SRU)
`
`January 27, 2021
`
`DEFENDANTS’ REPLY IN SUPPORT OF
`MOTION TO EXCLUDE PLAINTIFFS’ EXPERT DAVID TABAK
`
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 2 of 14
`
`TABLE OF CONTENTS
`
`
`Page
`
`
`PRELIMINARY STATEMENT ................................................................................................... 1
`ARGUMENT ................................................................................................................................. 2
`I.
`APPLYING FRE 702, COURTS RIGOROUSLY SCRUTINIZE PROPOSED
`EXPERT OPINIONS AND EXCLUDE UNRELIABLE EXPERT TESTIMONY. ........ 2
`DR. TABAK’S OPINIONS ARE UNRELIABLE. ........................................................... 2
`A.
`Dr. Tabak’s Tests Are Not Grounded On Sufficient Data. .................................... 2
`1.
`Dr. Tabak Did Not Identify Material, Unexpected Information,
`And Thus His Tests Lacked Sufficient Data.............................................. 2
`Dr. Tabak’s Opinions Rest On A Made-For-Litigation
`Methodology Rendering Them Unreliable And Subject To
`Rigorous Scrutiny. ..................................................................................... 4
`Dr. Tabak’s Opinions Are Not The Product Of Reliable Methods........................ 4
`1.
`Dr. Tabak’s FDT Test Does Not Test Market Efficiency. ......................... 4
`2.
`Dr. Tabak’s FDT Test Is Not Generally Accepted By Financial
`Economists And Lacks Meaningful Peer-Review. .................................... 6
`Dr. Tabak’s FDT Test Lacks A Known Error Rate. .................................. 8
`3.
`Dr. Tabak Did Not Apply His Tests In A Reliable Manner. ................................. 9
`1.
`Dr. Tabak’s Date Selection Methodology Is Unreliable, As It Is
`Infected With A Reverse Causality Problem. ............................................ 9
`Dr. Tabak Has Deviated From His Prior Approach In Ways That
`Bias His Results In Favor Of Plaintiffs’ Desired Outcome. .................... 10
`APPLYING DAUBERT, COURTS HAVE EXCLUDED PURPORTED
`EXPERT OPINIONS BASED ON DR. TABAK’S METHODOLOGY. ....................... 10
`CONCLUSION ............................................................................................................................ 10
`
`
`II.
`
`III.
`
`B.
`
`C.
`
`2.
`
`2.
`
`
`
`
`
`-i-
`
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 3 of 14
`
`
`
`TABLE OF AUTHORITIES
`
`Cases Page(s)
`
`In re Alstom SA Sec. Litig.,
`253 F.R.D. 266 (S.D.N.Y. 2008) ...............................................................................................7
`
`Cosby v. KPMG, LLP,
`No. 3:16-CV-121-TAV-DCP, 2020 WL 3548379 (E.D. Tenn. June 29, 2020) ..................7, 10
`
`Cosby v. KPMG, LLP,
`No. 3:16-CV-121-TAV-DCP, 2020 WL 3548653 (E.D. Tenn. June 29, 2020) ......................10
`
`Daubert v. Merrell Dow Pharm., Inc.,
`509 U.S. 579 (1993) ......................................................................................................... passim
`
`Dougherty v. Esperion Therapeutics, Inc.,
`No. 16-10089, 2020 WL 6793326 (E.D. Mich. Nov. 19, 2020) ................................................0
`
`In re Fed. Home Loan Mortg. Corp. (Freddie Mac) Sec. Litig.,
`281 F.R.D. 174 (S.D.N.Y. 2012) ...............................................................................................3
`
`Första AP-Fonden v. St. Jude Med., Inc.,
`312 F.R.D. 511 (D. Minn. 2015)................................................................................................7
`
`Lumen v. Anderson,
`280 F.R.D. 451 (W.D. Mo. 2012) ..............................................................................................5
`
`McIntire v. China MediaExpress Holdings, Inc.,
`38 F. Supp. 3d 415 (S.D.N.Y. 2014)......................................................................................2, 7
`
`Monroe Cty. Emps.’ Ret. Sys. v. S. Co.,
`332 F.R.D. 370 (N.D. Ga. 2019) ..............................................................................................10
`
`In re Namenda Indirect Purchaser Antitrust Litig.,
`No. 15-cv-6549, 2021 WL 100489 (S.D.N.Y. Jan. 12, 2021) ...................................................2
`
`OPERS v. Fed. Home Loan Mortg. Corp.,
`No. 08-cv-0160, 2018 WL 3861840 (N.D. Ohio Aug. 14, 2018) ........................................7, 10
`
`In re Petrobras Sec. Litig.,
`312 F.R.D. 354 (S.D.N.Y. 2016) ...............................................................................................6
`
`Pub. Emps.’ Ret. Sys. of Miss. v. TreeHouse Foods, Inc.,
`No. 16-CV-10632, 2020 WL 919249 (N.D. Ill. Feb. 26, 2020) ..............................................10
`
`Waggoner v. Barclays PLC,
`875 F.3d 79 (2d Cir. 2017).....................................................................................................2, 5
`
`
`
`
`
`-ii-
`
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 4 of 14
`
`
`
`PRELIMINARY STATEMENT
`
`Instead of addressing the serious flaws with their chosen expert’s made-for-litigation
`
`methodology, Plaintiffs’ Opposition (ECF 686) avoids Defendants’ fundamental Daubert
`
`challenges and tries to change the subject. For example, Plaintiffs backpedal from Cammer 5,
`
`arguing that “[s]tanding alone, [the indirect factors] are sufficient to demonstrate market
`
`efficiency.” Opp. 7. Dr. Tabak himself disagrees: he testified that he would not conclude a market
`
`is efficient without running direct tests, i.e., Cammer 5. Tabak Dep. 77:19-78:02. Months ago,
`
`Plaintiffs could have sought certification solely based on the indirect Cammer and Krogman
`
`factors. Instead, they chose to advance Dr. Tabak’s opinions, which depend on Cammer 5. Having
`
`made this strategic choice, they cannot escape its consequences. This Court should consider Dr.
`
`Tabak’s opinions only if they satisfy Daubert’s requirement that his opinions stem from sufficient
`
`facts, a reliable method, and a reliable application of the method. Daubert v. Merrell Dow Pharm.,
`
`Inc., 509 U.S. 579 (1993). Dr. Tabak’s opinions fail that test.
`
`Plaintiffs also conflate class certification with the Daubert question, hoping to use class
`
`certification decisional law to somehow show that Dr. Tabak’s methodology is a reliable test that
`
`is widely accepted among financial economists. It is not. Plaintiffs must find support in Dr.
`
`Tabak’s purported field of expertise to satisfy the Daubert standard. Again changing the subject,
`
`Plaintiffs argue that the parties merely dispute the weight of Dr. Tabak’s opinions. But the issue
`
`is whether Dr. Tabak’s made-for-litigation methodology passes the threshold tests of admissibility
`
`in view of the principles of the field of financial economics. It does not. It is unreliable, applied
`
`unreliably, and based on insufficient data.1 The Court should grant Defendants’ Motion.
`
`
`1 Plaintiffs also criticize Defendants’ experts for not conducting their own tests of market
`efficiency, see, e.g., Opp. 14, but that is irrelevant to whether Dr. Tabak’s opinions satisfy Daubert.
`
`
`
`

`

`
`
`I.
`
`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 5 of 14
`
`ARGUMENT
`
`Applying FRE 702, Courts Rigorously Scrutinize Proposed
`Expert Opinions And Exclude Unreliable Expert Testimony.
`
`Arguing that “Daubert motions are uncommon at class certification,” and it is unclear
`
`“whether Daubert applies” here (Opp. 8), Plaintiffs ignore the wall of contrary authority
`
`establishing the opposite. See Mot. (ECF 678) 4. There are “countless opinions” applying Daubert
`
`at the class certification stage. In re Namenda Indirect Purchaser Antitrust Litig., No. 15-cv-6549,
`
`2021 WL 100489, at *8 (S.D.N.Y. Jan. 12, 2021) (“This Court is not aware of any decision from
`
`this district explicitly suggesting that a Daubert inquiry was not necessary to decide a motion to
`
`exclude at the class-certification stage. In fact, the opposite is true; there are countless opinions
`
`applying Daubert.”). An application of Daubert dictates exclusion here.
`
`II.
`
`Dr. Tabak’s Opinions Are Unreliable.
`
`A.
`
`Dr. Tabak’s Tests Are Not Grounded On Sufficient Data.
`
`1.
`
`Dr. Tabak Did Not Identify Material, Unexpected
`Information, And Thus His Tests Lacked Sufficient Data.
`
`Before trying to defend Dr. Tabak’s unusual date selection approach, Plaintiffs try to
`
`diminish Cammer 5 as being of “no moment” and invite the Court to certify the class based on
`
`indirect factors alone. Opp. 7. Many cases support the importance of Cammer 5 (see Mot.
`
`§ II(A)(1)), and even some of Plaintiffs’ own cases recognize it as the “most important Cammer
`
`factor.” McIntire v. China MediaExpress Holdings, Inc., 38 F. Supp. 3d 415, 433 (S.D.N.Y. 2014),
`
`cited at Opp. 18. Regardless, Plaintiffs’ position puts them at odds with Dr. Tabak, who testified
`
`he would not opine a market was efficient without the support of empirical tests, i.e., tests of
`
`Cammer 5. Tabak Dep. 77:19-78:02. Plaintiffs also cite Waggoner, but that decision instructed
`
`that a plaintiff’s presentation of “flawed” empirical evidence weighs in favor of examining
`
`Cammer 5. Waggoner v. Barclays PLC, 875 F.3d 79, 98 (2d Cir. 2017) (citing Teamsters Local
`
`2
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 6 of 14
`
`
`
`445 Freight Div. Pension, Fund v. Bombardier Inc., 546 F.3d 196, 210 (2d Cir. 2008)); see also
`
`In re Fed. Home Loan Mortg. Corp. (Freddie Mac) Sec. Litig., 281 F.R.D. 174, 181-82 (S.D.N.Y.
`
`2012). Waggoner does not require this Court to ignore Cammer 5.
`
`Turning to Defendants’ actual critique of the data Dr. Tabak tested, Plaintiffs engage in
`
`sleight of hand. Bridling at Defendants’ reference to Dr. Tabak’s event selection process as
`
`“unusual,” Plaintiffs shift topics, contending that “[u]sing earnings dates” is “standard practice,”
`
`purporting to rely on Dr. Bajaj’s and counsel’s previous work. Opp. 9. But Dr. Tabak’s use of all
`
`stories (and not just material, unexpected information) from Dow Jones Newswire (and no other
`
`source), and inconsistent consideration of those stories (treating some dates as “news” for some
`
`tests and “non-news” for others), is highly unusual, and Plaintiffs do not show otherwise.
`
`As for earnings dates, neither Dr. Bajaj nor counsel ever stated that an expert should blindly
`
`test earnings dates without considering whether there was material, unexpected news on those
`
`dates. Indeed, Dr. Bajaj observed that on five earnings dates, the price reaction was in the opposite
`
`direction of the earnings surprise. Bajaj Rep. ¶¶ 71-72; Bajaj Decl. ¶ 40. Dr. Tabak tried to
`
`rehabilitate four of those five dates and justify the price reaction by identifying other information
`
`(including news he had not previously included) and then emphasizing certain news while ignoring
`
`other news. But Dr. Tabak did not even try to analyze 17 of the 22 earnings announcement dates.
`
`Plaintiffs portray Dr. Tabak’s refusal to identify material news as showing his objectivity, but in
`
`fact, Dr. Tabak’s hands-off approach reflects his failure to test a cause-and-effect relationship
`
`between material, unexpected news and Teva’s securities prices. Bajaj Rep. ¶¶ 59-64.
`
`Remarkably, to argue that Dr. Tabak analyzed dates with “material, unexpected news,”
`
`Plaintiffs only cite the 13 alleged corrective disclosure dates. Opp. 11. But Dr. Tabak himself has
`
`stated that corrective disclosure dates are inappropriate for testing. Bajaj Rep. ¶ 75 (citing David
`
`3
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 7 of 14
`
`
`
`Tabak, Use and Misuse of Event Studies to Examine Market Efficiency, NERA Economic
`
`Consulting (2010)). That is because the tester will know the likely results beforehand. Defendants
`
`do not seek to have it both ways, as Plaintiffs suggest. Opp. 11. Rather, the scientific method
`
`requires Dr. Tabak to test whether material, unexpected information yields consistent price
`
`reactions. Dr. Tabak should not select dates that he knows in advance will have such reactions.
`
`2.
`
`Dr. Tabak’s Opinions Rest On A Made-For-Litigation Methodology
`Rendering Them Unreliable And Subject To Rigorous Scrutiny.
`
`Plaintiffs’ various arguments against the made-for-litigation factor all fail. First, they
`
`acknowledge that the Second Circuit considers this Daubert factor. Opp. 11. Second, Defendants’
`
`experts have shown Dr. Tabak’s methodology is not “established” among financial economists.
`
`Opp. 12; see infra, § II(B)(2); Mot. § II(A)(2); Bajaj Rep. ¶ 15; McConnell Rep. ¶ 14; Bajaj Decl.
`
`¶ 4. Third, Plaintiffs’ argument that the FDT Article’s roadmap for plaintiffs actually increased
`
`plaintiffs’ burden cannot survive a reading of the article itself. Opp. 12; see also Mot. § II(A)(2).
`
`In fact, Dr. Tabak proved this point in finding efficiency based on a market reaction to 16.8% of
`
`news stories and direct tests with error rates of up to 98.6%. Tabak Rep., Ex. 8a-a; Bajaj Rep.
`
`¶ 20. Fourth, Plaintiffs’ argument that it is the “bread and butter” of an expert’s work to revise the
`
`methodology is odd. Opp. 13. Perhaps such refashioning is typical for Dr. Tabak, but inventing
`
`a methodology for plaintiffs to establish market efficiency, deviating from it, and then retooling it
`
`to escape competing experts’ criticisms shows this approach is made for litigation. Mot. § II(A)(2).
`
`B.
`
`Dr. Tabak’s Opinions Are Not The Product Of Reliable Methods.
`
`Dr. Tabak’s FDT Test Does Not Test Market Efficiency.
`
`1.
`Plaintiffs contend that Defendants are “legally and empirically wrong” in arguing that
`
`security prices should consistently respond to “material” news in an efficient market. Opp. 2, 15-
`
`17. Plaintiffs are incorrect on both points. The definition of market efficiency is well established
`
`4
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 8 of 14
`
`
`
`and provides that, in its semi-strong form, “security prices at all times reflect all material, publicly
`
`available information” and “stock prices consistently react promptly to ‘material, new, unexpected
`
`information’” in the correct direction. Bajaj Rep. ¶ 25; see Bajaj Decl. ¶ 37. Dr. Tabak agrees
`
`with Defendants on this point. See Tabak Rep. ¶ 11. Dr. Tabak admits that a market that only
`
`“sometimes” responds to material, unexpected information is inefficient. Tabak Dep. 61:22-63:10.
`
`The case law—including Plaintiffs’ own cases—likewise supports this view. See, e.g.,
`
`Waggoner, 875 F.3d at 94 (“An efficient market is ‘one in which the prices of the [stock]
`
`incorporate most public information rapidly.’”) (quoting Bombardier, 546 F.3d at 204) (emphasis
`
`added). Plaintiffs incorrectly cite Lumen to claim that Defendants have “a different conception of
`
`an efficient market than is used by the law.” Lumen v. Anderson, 280 F.R.D. 451, 460 (W.D. Mo.
`
`2012). In that case, however, the “different conception” was that the expert’s criticisms
`
`“implicate[d] the ‘strong’ form of efficient market, which is not the sort of market needed to
`
`support the fraud on the market described in Basic.” Here, Defendants’ experts made no such
`
`mistake; they considered the semi-strong form. Bajaj Rep. ¶ 16; McConnell Rep. ¶ 13.
`
`Plaintiffs are also incorrect that Dr. Tabak’s null hypothesis that “news doesn’t matter”
`
`was scientifically proper. Opp. 16 & n.11; Bajaj Decl. ¶ 42. Given that Dr. Tabak’s study
`
`purported to test for market efficiency, there is no justification for departing from the long-
`
`accepted null hypothesis for testing market efficiency. Nor is Dr. Tabak saved by the notion of an
`
`“alternative hypothesis.” The alternative hypothesis to Dr. Tabak’s hypothesis that prices never
`
`respond to news is not that prices always respond to news. The alternative just excludes “never,”
`
`i.e., that prices (might) sometimes (16.8% of the time) respond to news. Tabak Rep., Ex. 8a-a.
`
`But a market that only sometimes responds to news is not efficient. Tabak Dep. 61:22-63:10.
`
`5
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 9 of 14
`
`
`
`Trying to explain away Dr. Tabak’s inability to show consistent stock price reactions,
`
`Plaintiffs argue that “security prices do not respond to many earnings announcements and press
`
`releases.” Opp. 16. This simply ignores the issue of materiality. Prices are supposed to respond
`
`to material, unexpected news; indeed, that is the very basis of the Basic presumption. If earnings
`
`announcements are expected—and they often are—they are not material, unexpected news. Thus,
`
`Dr. Tabak’s observation that earnings announcements do not always result in stock price
`
`movements is consistent with the efficient market hypothesis, and it does not excuse him from
`
`testing for and proving consistent stock price reactions to material, unexpected news. Last,
`
`Plaintiffs wrongly assert that Dr. Tabak’s approach was “sustained” (Opp. 17) in In re Petrobras
`
`Sec. Litig., 312 F.R.D. 354, 369 (S.D.N.Y. 2016), aff’d in part, vacated in part, 862 F.3d 250 (2d
`
`Cir. 2017)). There was no Daubert motion in Petrobras, and the expert there used a very different
`
`news selection methodology than the methodology used here, which is infected with reverse
`
`causality and a high error-rate. Id.
`
`2.
`
`Dr. Tabak’s FDT Test Is Not Generally Accepted
`By Financial Economists And Lacks Meaningful Peer-Review.
`
`Plaintiffs have no real response to the argument that Dr. Tabak’s methodology lacks any
`
`independent academic support because it was not peer-reviewed at the time of his report and has
`
`been cited in just one peer-reviewed article—co-authored by fellow plaintiffs-side litigation
`
`consultants—since then. Mot. § II(A)(2). First, Plaintiffs cite Petrobras, but that decision did not
`
`address a Daubert challenge to Dr. Tabak’s approach. Opp. 18. Plaintiffs concede the point, but
`
`ignore its import: Because that expert’s methodology never survived a Daubert challenge, that
`
`decision hardly stands for the proposition that Plaintiffs’ expert methodology satisfies Daubert
`
`here. Beyond that major distinction, the expert methodology in Petrobras differs from Dr. Tabak’s
`
`6
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 10 of 14
`
`
`
`here. Supra, § (B)(1).2
`
`Plaintiffs also try to water down the peer-review standard, contending that various
`
`constituent parts of Dr. Tabak’s methodology have been subject to peer review, so this validates
`
`Dr. Tabak’s overall approach. Opp. 20. But these peer-reviewed articles are far removed from
`
`Dr. Tabak’s work here. See Mot. § II(B)(2). Plaintiffs weakly contend that a single October 2020
`
`article that footnotes the FDT Article shows that Dr. Tabak’s methodology is fully grounded in the
`
`peer-reviewed literature. Opp. 18, 21. That only a single peer-reviewed article cites the 17-year-
`
`old FDT Article is telling. Even more revealing is that this lone citation came from a fellow
`
`litigation consultant whose opinions using this test were excluded in OPERS v. Fed. Home Loan
`
`Mortg. Corp., No. 08-cv-0160, 2018 WL 3861840, at *5 n.4 (N.D. Ohio Aug. 14, 2018).
`
`Ultimately, Plaintiffs suggest this is essentially a debate among similarly credentialed
`
`economists (see Opp. 1 n.2), but nothing could be further from the truth. Dr. Tabak is a proverbial
`
`gun for hire, who left school for litigation consulting and never turned back. Dr. Tabak points to
`
`only two articles he has ever published, on any subject, in peer-reviewed journals over his entire
`
`career, and both journals were in the field of forensic economics, i.e., litigation support. See Opp.
`
`1 n.2 (citing ECF 419-5 at 35). By sharp contrast, Professor McConnell has served on the editorial
`
`
`2 Plaintiffs’ other cases offer no real response to this Daubert challenge. One court recognized
`that Dr. Tabak’s comparative analysis showed “some cause and effect relationship,” but it did not
`consider whether it is a test of market efficiency accepted by financial economists or whether stock
`prices react consistently to material, unexpected information. Första AP-Fonden v. St. Jude Med.,
`Inc., 312 F.R.D. 511, 521 (D. Minn. 2015). In another, the court devoted one sentence to the
`expert’s analysis of Cammer 5. In re Alstom SA Sec. Litig., 253 F.R.D. 266, 280 (S.D.N.Y. 2008).
`And Cosby admitted that the expert’s Cammer 5 results were “not a strong showing,” but allowed
`they “weigh in favor of finding market efficiency.” Cosby v. KPMG, LLP, No. 3:16-CV-121-
`TAV-DCP, 2020 WL 3548379, at *16 (E.D. Tenn. June 29, 2020). In doing so, Cosby expressly
`broke from more persuasive analyses of this issue cited by Defendants. Id. at *16 (citing Freddie
`Mac, 281 F.R.D. at 180). Finally, the expert’s different methodology in McIntire is of little
`relevance here, particularly since the expert screened for material news. 38 F. Supp. 3d at 429.
`7
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 11 of 14
`
`
`
`boards of 18 peer-reviewed journals and has published over 100 articles in peer-reviewed journals,
`
`which have been cited over 30,000 times. McConnell Rep. ¶ 34. He is a distinguished professor
`
`at Purdue University and has taught at other leading universities. Id. ¶¶ 1, 3. Similarly, Dr. Bajaj
`
`has also taught and published in leading academic journals in the field of financial economics,
`
`including The Journal of Finance and The Journal of Financial Economics. He has published
`
`over a dozen articles in peer-reviewed journals, which have been cited almost 200 times. Bajaj
`
`Rep. ¶ 5-6. Professor McConnell and Dr. Bajaj determined that Dr. Tabak’s method is unreliable,
`
`and this Court should credit those expert opinions.
`
`Dr. Tabak’s FDT Test Lacks A Known Error Rate.
`
`3.
`Rather than confront Defendants’ criticism that Dr. Tabak’s test lacks a known error rate,
`
`Plaintiffs advance a number of straw-man arguments. For instance, they mischaracterize
`
`Defendants as arguing only that Dr. Tabak’s test is imperfect. Opp. 23. But a false positive rate
`
`approaching 98.9% is hardly a mere imperfection. See Mot. § II(B)(3); Bajaj Rep. ¶ 20. Indeed,
`
`as Dr. Tabak admits, his results are consistent with a finding of market inefficiency—the exact
`
`opposite of his conclusion. See Tabak Dep. 163:04-164:21. Given that Dr. Tabak’s tests are
`
`consistent with either result, Dr. Tabak fails to provide an error rate. 509 U.S. at 592-94.
`
`Plaintiffs’ argument that Dr. Tabak’s methodology is a test of statistical significance with
`
`a 5% error rate misunderstands the question and Dr. Tabak’s own hypothesis. Opp. 23-24.
`
`Plaintiffs’ suggested 5% error rate has nothing to do with Dr. Tabak’s conclusion that the market
`
`for Teva securities is efficient. Dr. Tabak’s null hypothesis was not market efficiency, but rather
`
`that the price never responds to news. Tabak Dep. 95:2-4. So the question is how often is Dr.
`
`Tabak wrong to conclude that a stock that reacts only sometimes to news is trading in an efficient
`
`market. Dr. Tabak knows of no error rate for that question. Id. at 163:04-164:21. And Plaintiffs
`
`are incorrect that a 5% error rate applies to Dr. Tabak’s tests, which are infected with reverse
`8
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 12 of 14
`
`
`
`causality issues. When his test is applied to simulated markets that are inefficient by design, it
`
`consistently yields false positives. Bajaj Rep. ¶ 20. This error rate, arising from reverse causality
`
`in Dr. Tabak’s methodology, is up to 90.6% for Dr. Tabak’s KS test and up to 98.9% for his z-test.
`
`Id. Dr. Bajaj showed these faults using Dr. Tabak’s own data. Bajaj Rep. ¶ 106.
`
`C.
`
`Dr. Tabak Did Not Apply His Tests In A Reliable Manner.
`
`1.
`
`Dr. Tabak’s Date Selection Methodology Is
`Unreliable, As It Is Infected With A Reverse Causality Problem.
`
`Plaintiffs try to brush off the reverse causation flaw as a minor detail, even though their
`
`own expert conceded the criticism (Tabak Rebuttal Rep. ¶ 3d) and offered a new report with new
`
`tests purportedly to address it. Bajaj Decl. ¶¶ 6-32.3 First, Plaintiffs’ argument that “Defendants
`
`have not shown that this [reverse causation] in fact occurred” is simply incorrect. Opp. 25. Dr.
`
`Bajaj established the presence of reverse causation. Bajaj Rep. ¶¶ 85-108; Bajaj Decl. ¶¶ 15-32.
`
`Second, Plaintiffs’ arguments misunderstand Dr. Tabak’s revised work. Dr. Tabak did not relate
`
`stories published on Monday after close to a price reaction on Tuesday; instead, he related those
`
`stories to a time period both before and after publication. Bajaj Decl. ¶¶ 6-32. Because the stories
`
`in Dr. Tabak’s news sample straddle this temporal line, they are much more likely to be following
`
`a large preceding return rather than cause a subsequent price reaction. Dr. Bajaj proved this occurs.
`
`Id. ¶¶ 15-30. Dr. Tabak has not “refuted” this reverse causality issue, as his proposed fix ignores
`
`after-hours trading on the NYSE and the TASE. Opp. 25. While Dr. Tabak claimed to exclude
`
`news that was time-stamped during trading hours, he did not do so for Teva Notes. Just excluding
`
`
`3 Plaintiffs ask the Court to disregard Dr. Bajaj’s declaration because “Defendants refused to
`produce the ‘facts or data’ underlying” it. Opp. 29. This is yet another attempt to manufacture an
`issue; the data in question is Dr. Tabak’s own data. Similarly, Plaintiffs’ argument that
`Defendants’ Daubert motion is untimely is merely an attempt to avoid consideration of Dr.
`Tabak’s flawed approach. Opp. 6. Defendants filed their Motion shortly after Dr. Tabak’s rebuttal
`report and ensuing deposition on a briefing schedule set by this Court. ECF 652 at 3-4 & n.6.
`9
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 13 of 14
`
`
`
`news that Dr. Tabak claimed should be excluded (but was not) results in six of seven Teva notes
`
`failing Dr. Tabak’s own test of market efficiency. Bajaj Decl. ¶ 9.
`
`2.
`
`Dr. Tabak Has Deviated From His Prior Approach In Ways
`That Bias His Results In Favor Of Plaintiffs’ Desired Outcome.
`
`Plaintiffs concede that Dr. Tabak has deviated from the approach laid out in the FDT
`
`Article, contending merely that Dr. Tabak has “used an updated version of that methodology in
`
`numerous cases.” Opp. 26. Dr. Tabak has no good answer for departing from his own approach.
`
`The use of corrective disclosures gives rise to look-ahead bias. And his pooling decision assumes
`
`the null hypothesis, rather than trying to disprove it as the scientific method requires. Nor can
`
`Plaintiffs explain away that all of Dr. Tabak’s modifications favor Plaintiffs. See Mot. § II(C)(2).
`
`III. Applying Daubert, Courts Have Excluded
`Purported Expert Opinions Based On Dr. Tabak’s Methodology.
`
`This Court would not be the first to exclude market efficiency opinions based in whole or
`
`in part on Dr. Tabak’s methodology. Mot. § III. Plaintiffs cherry pick distinctions in those cases,
`
`but none of these change the point that other courts have found this methodology flawed. Id.
`
`Plaintiffs also attempt to diminish OPERS by misreading several cases citing it. Opp. 28-29. None
`
`of these cases distinguishing OPERS are analogous, as Plaintiffs admit in citing some of them.4
`
`CONCLUSION
`
`For the foregoing reasons, the Court should grant Defendants’ motion.
`
`
`4 Cosby, 2020 WL 3548379, at *22 (citing OPERS on price impact); Pub. Emps.’ Ret. Sys. of Miss.
`v. TreeHouse Foods, Inc., No. 16-CV-10632, 2020 WL 919249, at *4 (N.D. Ill. Feb. 26, 2020)
`(market efficiency uncontested). Even a quick review shows why Plaintiffs’ other cases have little
`application here. Monroe Cty. Emps.’ Ret. Sys. v. S. Co., 332 F.R.D. 370, 385 (N.D. Ga. 2019)
`(defense expert’s own study suggested market efficiency); Dougherty v. Esperion Therapeutics,
`Inc., No. 16-10089, 2020 WL 6793326, at *4 (E.D. Mich. Nov. 19, 2020) (speculating that OPERS
`was an “outlier” only in its emphasis on Cammer 5, not its treatment of the FDT test); Cosby v.
`KPMG, LLP, No. 3:16-CV-121-TAV-DCP, 2020 WL 3548653, at *11 (E.D. Tenn. June 29, 2020)
`(citing OPERS once, and not in connection with the various arguments Plaintiffs discuss in the
`citation).
`
`10
`
`
`

`

`Case 3:17-cv-00558-SRU Document 696 Filed 01/27/21 Page 14 of 14
`
`Respectfully submitted,
`
`/s/ Jason D. Frank
`Jordan D. Hershman (admitted pro hac vice)
`Jason D. Frank (admitted pro hac vice)
`Emily E. Renshaw (admitted pro hac vice)
`Elizabeth Geyelin Hays (admitted pro hac vice)
`Andrew M. Buttaro (ct30882)
`MORGAN, LEWIS & BOCKIUS LLP
`One Federal Street
`Boston, MA 02110-1726
`Tel.: (617) 341-7700
`Fax: (617) 341-7701
`jordan.hershman@morganlewis.com
`jason.frank@morganlewis.com
`emily.renshaw@morganlewis.com
`liza.hays@morganlewis.com
`andrew.buttaro@morganlewis.com
`
`Sheron Korpus (admitted pro hac vice)
`KASOWITZ BENSON TORRES LLP
`1633 Broadway
`New York, New York 10019
`Tel.: (212) 506-1969
`Fax: (212) 500-3469
`skorpus@kasowitz.com
`
`Counsel for Defendants
`
` –
`
` and –
`
`
`Jill M. O’Toole (ct27116)
`SHIPMAN & GOODWIN LLP
`One Constitution Plaza
`Hartford, CT 06103-1919
`Tel.: (860) 251-5000
`Fax: (860) 251-5218
`jotoole@goodwin.com
`
`Counsel for Defendants except Kåre Schultz
`
`
`
`
`
`
`
`11
`
`
`