DEBRA ANN LIVINGSTON, Circuit Judge:

Plaintiffs-Appellants Teachers' Retirement System of Louisiana and Christine Fleckles, acting on behalf of themselves and other similarly situated investors (collectively, "Plaintiffs"), brought suit in the United States District Court for the Southern District of New York against Pfizer, Inc. and several of its directors and officers,2 alleging violations of §§ 10(b), 20(a), and 20A of the Securities Exchange Act of 1934, 15 U.S.C. §§ 78j(b), 78t(a), 78t-1, as well as Securities and Exchange Commission Rule 10b-5 ("Rule 10b-5") promulgated thereunder, 17 C.F.R. § 240.10b-5. According to Plaintiffs, between October 31, 2000 and October 19, 2005, Pfizer made fraudulent misrepresentations and fraudulently omitted to disclose information regarding the safety of two of its drugs, Celebrex (celecoxib) and Bextra (valdecoxib). When the market eventually learned of the cardiovascular risks associated with these drugs, the value of Pfizer's shares fell, harming Plaintiffs and other shareholders in the process. A class-action lawsuit followed.

The district court (Swain, J.) denied Pfizer's motion to dismiss the complaint, certified the class, and allowed the parties to proceed through discovery. After extensive discovery and nearly a decade of litigation, the district court granted a motion in limine to exclude Plaintiffs' expert on loss causation and damages, Daniel R. Fischel ("Fischel"), from testifying at trial. The court provided two reasons for excluding the testimony. First, in an earlier summary judgment ruling, the court had determined that Pfizer is not liable for certain alleged misrepresentations G.D. Searle & Co. ("Searle") and Pharmacia Corporation ("Pharmacia") — the companies that manufactured Celebrex and Bextra before Pfizer — made when they owned the drugs. Because Fischel did not isolate the effects of Pfizer's alleged misrepresentations and omissions from the effects of certain of Searle's and Pharmacia's allegedly fraudulent statements, the court concluded that his analysis would be unhelpful to the jury in determining the losses that Pfizer caused.

Second, using an "event study," Fischel had calculated stock-price inflation caused by Pfizer's alleged fraud by identifying seven days on which Pfizer's stock price fell when the market discovered allegedly concealed information about Celebrex and Bextra and five days on which Pfizer's stock price rose in reaction to new information about the drugs. When the district court determined in its summary judgment ruling that the stock-price declines on two days in the event study could not reasonably be attributed to Pfizer's alleged fraud, Fischel removed those days from his analysis and also adjusted the amount of the stock-price increases that he included in his calculations. The court concluded that Fischel's methodology for adjusting the amount of price increases attributable to Pfizer's fraud was not "the product of reliable principles and methods reliably applied." In re Pfizer Inc. Secs. Litig., No. 04 Civ. 9866 (LTS) (HBP), 2014 WL 2136053, at *1 (S.D.N.Y. May 21, 2014) ("In re Pfizer II"). For these two reasons, the court prevented Fischel from testifying about loss causation or damages. Left with no testimony on these issues, Plaintiffs could not sustain key elements of their claims, and the district court granted Pfizer's motion for summary judgment.

On appeal, Plaintiffs argue that the district court abused its discretion in excluding Fischel's testimony, even if the district court correctly found that Pfizer could not be held liable for certain Searle and Pharmacia statements, because Fischel had no obligation to disaggregate those statements from Pfizer statements. Plaintiffs also argue that the district court was incorrect to conclude that Pfizer was not liable, as a matter of law, for the Searle and Pharmacia statements. We conclude, first, that the district court abused its discretion by excluding Fischel's testimony in its entirety. The court erred in concluding that Fischel needed to disaggregate the effects of Pfizer's allegedly fraudulent conduct from Searle's or Pharmacia's, regardless of whether Pfizer is ultimately found liable for the latters' statements. Under Plaintiffs' theory of the case, Fischel's testimony could have been helpful to the jury even without such disaggregation. As for Fischel's adjustment to the price increases, the district court did not abuse its discretion in concluding that this change was not sufficiently reliable to be presented to a jury. However, Fischel's error did not render the remainder of his testimony unreliable. The court should have prevented him from testifying about the adjustment, but otherwise allowed him to present his findings on loss causation and damages.

We further find that the district court erred in concluding, as a matter of law, that Pfizer had insufficient authority over certain Searle and Pharmacia statements as to have "made" them for the purposes of Rule 10b-5. We note, however, that our finding that the district court abused its discretion in excluding Fischel's testimony does not turn on the question of Pfizer's ultimate liability for these statements. Accordingly, we hereby vacate and remand the judgment of the district court for further proceedings consistent with this opinion.

BACKGROUND

A. Factual Background

Celebrex and Bextra are part of a broad class of medicines known as non-steroidal anti-inflammatory drugs, which are used to treat chronic pain and inflammation. Before 1999, this class of drugs had a common problem: patients who used the drugs over a long period of time often developed stomach ulcers and other gastrointestinal problems. In this deficiency, pharmaceutical manufacturers saw opportunity. Specifically, two companies — Merck & Co., Inc. and Searle — began researching a type of non-steroidal anti-inflammatory drug, known as a Cyclooxygenase 2 ("COX-2") inhibitor, which could reduce pain and inflammation without causing gastrointestinal distress. Both companies ultimately succeeded, with Merck creating a drug called Vioxx, and Searle creating Celebrex.

Pfizer, a "research-based, global pharmaceutical company that develops, manufactures and markets prescription medicines," In re Pfizer Inc. Secs. Litig., No. 04 Civ. 9866(LTS) (HBP), 936 F.Supp.2d 252, 257 (S.D.N.Y.2013) ("In re Pfizer I"), first became involved with COX-2 inhibitors through Searle. In February 1998, Pfizer signed a series of agreements with Searle in which it agreed to, among other things, help market Celebrex (collectively, the "Co-Promotion Agreement"). Searle later transferred control over Celebrex to Pharmacia through a merger in early 2000, and Pharmacia succeeded to Searle's rights under the Co-Promotion Agreement.3 Pfizer continued to fulfill its obligations under the Co-Promotion Agreement until April 16, 2003, when it obtained the exclusive rights to manufacture, promote, and sell Celebrex and Bextra by purchasing Pharmacia.

Plaintiffs contend that, while Celebrex and Bextra eliminated the gastrointestinal issues associated with non-steroidal anti-inflammatory drugs, the drugs presented a different, dangerous side effect. As early as 1998, they claim, Pfizer and Searle knew about studies linking the COX-2 inhibitors to cardiovascular problems in patients. But because Celebrex was an enormous commercial success, Searle issued press releases and other public statements denying that the drugs presented such risks. When ownership of Celebrex passed to Pharmacia, and later to Pfizer, both companies continued to tout its safety, as well as the safety of Bextra, notwithstanding the discovery of additional medical evidence tying the drugs' use to heightened cardiovascular risks. According to Plaintiffs, the press releases and public statements that Pharmacia and Pfizer issued during the class period had the effect of maintaining the public's misperception about the safety of Celebrex and Bextra. From a market perspective, this meant that during the class period, investors continued to value Pharmacia, and later Pfizer, as if the companies' products provided alternatives to other non-steroidal anti-inflammatory drugs without other side effects (namely, cardiovascular risks) that could deter consumers.

According to Plaintiffs, Pfizer's effort to conceal these risks reached the breaking point in the fall of 2004. On September 30, Merck announced that it was withdrawing Vioxx from the market due to cardiovascular safety concerns. Seeking to capitalize on the downfall of its largest competitor in the COX-2 inhibitor market, Pfizer issued a series of press releases, advertisements, and public statements assuring investors that no studies had "shown any increased cardiovascular risk [associated with] Celebrex" and that Bextra's "cardiovascular safety profile is also well established in long-term studies." J.A. 790, 795 (internal quotation marks omitted). But new scrutiny followed Merck's announcement. On October 6, an editorial in the New England Journal of Medicine "questioned the safety of ... [Pfizer's] Celebrex and Bextra." J.A. 802 (internal quotation marks omitted). Soon after, information about studies linking the drugs to cardiovascular risks reached the public eye. These revelations, Plaintiffs claim, caused Pfizer's share prices to fall as investors reassessed the value of Celebrex and Bextra in light of the newly discovered risks.

B. Procedural History

On December 15, 2004, a putative class action was filed against Pfizer in the United States District Court for the Southern District of New York (Swain, J.). Plaintiffs represent all investors who purchased Pfizer stock between October 31, 2000 and October 19, 2005 (the "Class Period"). They allege that Pfizer concealed the safety risks of Celebrex and Bextra — both through statements delivered by Searle, Pharmacia, Pfizer, and their employees when Pfizer was a party to the Co-Promotion Agreement, and also through Pfizer's own statements when Pfizer later owned the drugs — in violation of §§ 10(b), 20(a), and 20A of the Securities Exchange Act of 1934. With respect to the claims under these three sections, the district court denied Pfizer's motion to dismiss the complaint and, on April 6, 2012, certified the class.

With the class certified, Plaintiffs turned to developing their theory of Pfizer's liability. Notably, Searle and Pharmacia owned Celebrex and Bextra before Pfizer, and both Searle and Pharmacia allegedly made fraudulent misrepresentations and omitted to disclose material information about the cardiovascular risks associated with the drugs. Plaintiffs alleged that Pfizer is responsible for some of these misrepresentations — in particular, statements by Searle, Pharmacia, and their employees — because Pfizer had authority over those statements. But Pfizer, Plaintiffs contended, also engaged in its own fraudulent misrepresentations and omissions about the drugs both before, and during, the Class Period. Like the Searle and Pharmacia statements, these misrepresentations and omissions supposedly concealed the cardiovascular risks associated with Celebrex and Bextra.

This array of fraudulent misrepresentations and omissions created an issue for Plaintiffs: how to determine whether Pfizer's alleged fraud, as differentiated from alleged fraud by Pharmacia or Searle, caused Pfizer's stock price to fall. To address that issue, Plaintiffs relied on an inflation-maintenance theory of liability. All three companies, Plaintiffs claim, engaged in fraudulent misrepresentations and omissions that concealed the same information — namely, that Celebrex and Bextra increase the risk of cardiovascular problems; investors, however, typically change their assessment of a company's value based on new information, not statements that reiterate old news. Thus, by fraudulently concealing the same risks that Pharmacia and Searle hid, Pfizer perpetuated the market's misperceptions about Celebrex and Bextra, which caused the market to maintain the company's stock price at an artificially high level. Because the market would have adjusted the value of Pfizer's stock to reflect the true risks associated with Celebrex and Bextra if Pfizer had not continued to conceal those risks, Pfizer should be liable for the full amount by which its stock price fell when the market eventually discovered the truth.

If correct, this theory of liability could provide a means for establishing that Pfizer is liable for the full value of the information that its predecessors concealed.4 A theory of liability, however, does not constitute proof of liability. It does not show that the alleged misinformation about the drugs, once revealed, actually caused shareholders to lose money, nor does it identify the extent of those losses. For those issues, Plaintiffs turned to Daniel R. Fischel, Professor Emeritus of Law and Business at the University of Chicago Law School and a former dean of that institution, who issued his initial expert report on January 13, 2012 (the "Initial Report").

In this Initial Report, Fischel performed an event study analysis to determine whether, and the extent to which, Pfizer's stock price changed when the market learned about the cardiovascular risks associated with Celebrex and Bextra. Fischel contended that, assuming that Pfizer engaged in misrepresentations and omissions that kept information about the cardiovascular risks of Celebrex and Bextra from the market, his methodology isolated the extent to which the alleged fraud had artificially inflated Pfizer's stock price throughout the Class Period.

An event study is "a statistical regression analysis that examines the effect of an event on a dependent variable, such as a company's stock price." Jay W. Eisenhofer et al., Securities Fraud, Stock Price Valuation, and Loss Causation: Toward a Corporate Finance-Based Theory of Loss Causation, 59 Bus. Law. 1419, 1425 (2004). Fischel briefly described this form of analysis in his Initial Report. In many event studies, the expert begins by analyzing how the defendant company's stock price typically changes from day-to-day as compared to securities issued by similar companies and the market as a whole. This comparison makes it "possible to predict what the return of a [defendant company's] security should be on a certain date given the return for the market as a whole" — known as the "predicted return." Daniel R. Fischel, Use of Modern Finance Theory in Securities Fraud Cases Involving Actively Traded Securities, 38 Bus. Law. 1, 18 (1982); J.A. 935. The expert then identifies days on which information correcting the alleged fraud reached the market and notes the amount that the defendant company's stock price changed on that day — the "actual return." Fischel, supra at 18; J.A. 935. "[B]y comparing the predicted return with the actual return on the date of release of the supposedly correct information or immediately thereafter, the [event study] attempts to isolate the change in the return earned by investors that is attributable solely to the allegedly withheld or false information." Fischel, supra at 18. The difference between the actual return and the predicted return is known as the "residual return." J.A. 935.

Performing an event study can thus help an expert to determine at least two things. First, assuming that the defendant company fraudulently concealed information, the event study shows how much money the fraud caused shareholders to lose. Identifying residual returns on days when allegedly concealed information reached the market indicates that the supposedly withheld information caused the company's stock price to change. See Frederick C. Dunbar & Arun Sen, Counterfactual Keys to Causation and Damages in Shareholder Class-Action Lawsuits, 2009 Wis. L.Rev. 199, 228. If the release of allegedly withheld information causes a stock price decrease, shareholders who purchased the defendant company's stock after the alleged fraud but before the revelation may have paid a higher price than they would have but for the defendant's fraudulent conduct — known as an "artificial[ly] inflat[ed]" price. J.A. 939.

Second, the event study helps the expert "calculat[e] what the price of [the defendant company's] security would have been had the alleged wrongful conduct not occurred," by estimating the amount of artificial inflation in the company's stock price over time. Fischel, supra at 17. Just as the existence of a residual return on a day when the market discovers allegedly concealed information shows that the company's stock price was artificially inflated, the size of the residual return on such a day provides evidence of the amount by which concealing that particular information inflated the defendant company's stock. As a result, if concealed information reached the market through multiple corrective disclosures, the sum of the residual returns associated with those disclosures provides evidence about the amount of artificial inflation in the company's stock after the fraud but before those corrections. See Bradford Cornell & R. Gregory Morgan, Using Finance Theory to Measure Damages in Fraud on the Market Cases, 37 U.C.L.A. L.Rev. 883, 899 (1990). Thus, an expert using an event study can estimate the amount of artificial inflation in the defendant company's stock price when shareholders purchased their shares, which is equivalent to estimating the difference between what those investors should have paid for the shares but-for the alleged fraud, and what they actually paid. See Dunbar & Sen, supra at 231.

After explaining these background principles, Fischel turned to the particulars of Plaintiffs' claims. At the outset, he emphasized that Plaintiffs had not asked him to determine whether Pfizer did, in fact, fraudulently conceal information about Celebrex and Bextra, or whether it concealed the same information as Pharmacia and Searle. Rather, he "assume[d] Plaintiffs will be able to establish at trial that [Pfizer] withheld material information" about the cardiovascular risks associated with the drugs "from at least the beginning of the Class Period," J.A. 927 n. 6, and instead used his analysis to determine whether, and the extent to which, Pfizer's share price fell when investors discovered those risks. As Fischel explained, if Plaintiffs are correct that Pfizer concealed the cardiovascular risks associated with Celebrex and Bextra, measuring the amount by which the revelation of the risks caused Pfizer's share price to change identifies the "artificial inflation" in the company's stock due to the alleged fraud. J.A. 938-39.

Operating under this theory, Fischel first identified seven days on which the market learned information about the cardiovascular risks associated with Celebrex and Bextra, and Pfizer's stock price fell by a greater amount than expected. In other words, on the days these disclosures (the "Corrective Disclosures") occurred, Pfizer's stock price showed negative residual returns. But Fischel also observed that, around the same time as the Corrective Disclosures, Pfizer's stock price rose more than expected on five days when the market learned new information about the COX-2 inhibitors. That is, on the five days that these disclosures (the "Positive Disclosures") occurred, Pfizer's stock price showed positive residual returns. To account for these positive residual returns, Fischel offset the stock price movements on the Corrective Disclosure days by the movements on the Positive Disclosure days. The result revealed that, throughout the Class Period, Pfizer's alleged misrepresentations and concealments about the cardiovascular risks associated with Celebrex and Bextra artificially inflated the value of its stock. Specifically, Fischel determined that Pfizer's stock was inflated by $1.46 per share at the beginning of the Class Period and that inflation rose as high as $3.95 per share before eventually falling to $0 by October 20, 2005.

According to Fischel, offsetting the Positive Disclosures against the Corrective Disclosures gave a more accurate measurement of the effects of the alleged fraud because "[i]gnoring [these] price increases... would overstate [the] alleged artificial inflation." J.A. 1142. On two of the Positive Disclosure dates, he noted, Pfizer's stock price rose as the market learned about the risks associated with Vioxx and implicitly predicted greater sales for Celebrex and Bextra after learning that Vioxx was being withdrawn and Celebrex and Bextra would be kept on the market. The subsequent Corrective Disclosures therefore revealed both allegedly concealed information about Celebrex and Bextra and showed the market that its prediction of greater sales was mistaken. To identify just the artificial inflation associated with allegedly concealed information, then, Fischel determined that he needed to offset the price increases on these two Positive Disclosure dates against those on the Corrective Disclosure dates.

On the other three Positive Disclosure dates, Fischel observed that investors increased their valuation of Pfizer after learning that, despite the link between the COX-2 inhibitors and cardiovascular problems, the risks were not as severe as anticipated and Pfizer would be able to continue selling the drugs. Investors, in other words, tempered their initially dire predictions about the effect of the cardiovascular risks on Pfizer's ability to sell Celebrex and Bextra. Since he determined that the price movements on Corrective Disclosure days reflected overly pessimistic projections, Fischel offset the value of the subsequent Positive Disclosures to get a more accurate assessment of the true value of the concealed information about the cardiovascular risks.

Recognizing the possibility that the fact-finder might conclude that not all the Corrective Disclosures identified by Fischel were related to the fraud, Pfizer's counsel pressed Fischel during a deposition on how he would adjust his analysis to account for such a decision. To that end, Fischel addressed the following hypothetical:

J.A. 1135-36.

Pfizer's expert, Dr. Paul A. Gompers, responded directly to Fischel's report, rather than conducting his own analysis. Gompers did not contest the use of the event study methodology and noted that he had "no major criticism" of how Fischel obtained the "residual return[s]" in his analysis. J.A. 1645; 1652. Instead, Gompers focused his criticism on Fischel's assumptions that certain Corrective Disclosures actually revealed information about Celebrex and Bextra that had fraudulently been concealed.

After this initial round of expert discovery, Pfizer filed a motion for summary judgment, which the district court granted in part and denied in part on March 28, 2013. The court rejected Pfizer's argument that Plaintiffs could not show reliance because none of Pfizer's alleged misrepresentations caused the company's stock price to rise. "A misstatement," the court countered, "may cause inflation simply by maintaining existing market expectations." In re Pfizer I, 936 F.Supp.2d at 264 (emphasis added) (internal quotation marks omitted). The court concluded that Fischel's analysis showing that Pfizer's share price fell in response to "disclosures of adverse information about Celebrex and Bextra[]" could provide a basis for a reasonable jury to conclude that the price was "artificially inflated at the beginning of the Class Period," and that Plaintiffs relied on that inflated price. Id. at 265. The district court then rebuffed Pfizer's argument that its alleged misrepresentations did not cause the company's share prices to fall, concluding that, for five of the seven Corrective Disclosure days, Plaintiffs had raised a genuine dispute of material fact about whether Pfizer's share prices fell because the market discovered previously concealed information. Id. at 265-67.

The district court, however, also ruled against Plaintiffs in two respects: First, it found that losses on two Corrective Disclosure dates could not reasonably be attributed to disclosures exposing Pfizer's alleged misrepresentations. On one day, the disclosed information did not provide evidence of a "causal link" between the COX-2 inhibitors and cardiovascular risks. Id. at 267. On the other, Pfizer did not "reveal[ ] any new information as to Celebrex and Bextra's cardiovascular risks" and instead only disseminated business information regarding earnings. Id. at 267-68. Second, the district court addressed Plaintiffs' argument that Pfizer is liable for statements by Searle, Pharmacia and their employees when the Co-Promotion Agreement was in place. Plaintiffs identified ten such misrepresentations, but the district court decided that Pfizer could not be held liable for nine of them. Pfizer, it concluded, is liable only for Searle and Pharmacia statements over which it had "ultimate authority." Id. at 271. While the Co-Promotion Agreement provided such authority for press releases, only one of the ten alleged misrepresentations was a press release. The district court concluded, as a matter of law, that Pfizer did not have sufficient authority over the other nine statements, eight of which were made directly by Searle or Pharmacia employees and one of which was contained in a Pharmacia Form 8-K filing. Id.5 The court therefore granted Pfizer's motion for summary judgment to the extent that Plaintiffs' claims were tied to the two Corrective Disclosures that the court determined had not revealed previously undisclosed risk, and to the extent that they were premised on misstatements the court deemed not attributable to Pfizer.

Plaintiffs responded to this summary judgment decision by having Fischel file an updated expert report (the "Supplemental Report"). The Supplemental Report contained no adjustments in response to the district court's conclusion that, as a matter of law, Pfizer could not be held liable for certain statements. But the Supplemental Report did contain adjustments to account for the Corrective Disclosures that the district court had decided could not be linked to Pfizer's misrepresentations. Fischel first removed these Corrective Disclosures from his event study analysis, which resulted in the overall stock price declines caused by the Corrective Disclosures falling "by 9.7 percent." J.A. 1177. Fischel then stated that, "[b]ecause eliminating the stock price declines related to Celebrex and Bextra on these dates reduce[d] the total residual stock price decline [he] estimated ... by 9.7 percent," he "proportionally reduce[d] the residual stock price increases [he] measured... by 9.7 percent." Id. (emphasis added).

During a second deposition, Fischel explained that excluding two Corrective Disclosures created the possibility that some price changes on the Positive Disclosure days involved reactions to non-fraud-related information about Celebrex and Bextra. Despite being unable to identify "any clear one-for-one relationship between any particular [P]ositive [D]isclosure and any particular [Corrective] [D]isclosure," J.A. 1114, he decided that the exclusion of the Corrective Disclosures must have "ha[d] an effect on the [P]ositive [D]isclosures because inflation cannot come out of a stock unless it goes into the stock." J.A. 1110. Thus, he settled on a 9.7 percent proportional reduction to the amount of the Positive Disclosures attributed to Pfizer's fraud (the "Proportional Reduction").

Once again, Gompers issued a report challenging Fischel's analysis (the "Supplemental Response"). His Supplemental Response raised two objections to Fischel's Supplemental Report, both of which eventually became the basis for Pfizer's motion in limine to exclude Fischel from testifying. First, Gompers argued that Fischel's Proportional Reduction relied on an unreliable methodology. He observed that the Initial Report concluded both "that there [was] positive inflation [in Pfizer's stock] at the beginning of the class period," and that Pfizer's stock price increased during the class period on the Positive Disclosure days. J.A. 1153-54 (emphasis added). Gompers argued that, while the decision to remove two of the Corrective Disclosures did "mean[ ] less inflation entered the stock price," J.A. 1159, that did not justify reducing price increases associated with the Positive Disclosures, as opposed to the amount of artificial inflation that was in Pfizer's stock when the Class Period began. If Fischel had simply removed the rejected Corrective Disclosures from his calculations without carrying out the Proportional Reduction, Gompers explained, the amount of artificial inflation at the beginning of the Class Period would have fallen by an amount equivalent to the losses previously attributed to those Corrective Disclosures, thus achieving equilibrium between inflation entering and exiting Pfizer's stock. Fischel failed to justify his decision to instead proportionally reduce the amount of inflation attributable to the Positive Disclosures, Gompers argued, which by comparison led to a higher estimate of artificial inflation in Pfizer's stock price at the beginning of the Class Period.

Second, Gompers argued that Fischel erred in his Supplemental Report by ignoring the district court's decision that Pfizer is not responsible for many of Searle's and Pharmacia's alleged misstatements about Celebrex. Fischel's analysis, he argued, is "predicated on an assumption that ... [Pfizer is] liable for all alleged misstatements and omissions" made by Searle and Pharmacia, and fails to disaggregate the "price inflation numbers that are attributable to [Searle's and] Pharmacia's alleged misstatements and omissions." J.A. 1162 (emphasis added). As a result, his testimony would not be helpful to jurors who need to decide the amount of damages attributable to Pfizer alone.

Relying on Gompers's report, Pfizer moved, pursuant to Federal Rule of Evidence 702, to exclude Fischel from testifying about loss causation and damages. In an order issued on May 21, 2014, the district court granted the motion. The court found that Fischel's opinion was unreliable because he "proffered no explanation of the analytical basis" for his Proportional Reduction and provided neither an "explanation of the relationships among the events triggering the respective price decreases and increases" nor "research reference or peer review information ... in support of [his proportional] adjustment method." In re Pfizer II, 2014 WL 2136053, at *1. Furthermore, the court decided that "Fischel's failure to account in any way for the impact of the excluded [Searle and] Pharmacia statements renders his opinions unhelpful to the jury in making calculations of damages proximately caused by [Pfizer's] alleged misrepresentations and omissions." Id. It therefore excluded Fischel's testimony in its entirety.

Plaintiffs requested, and received, an opportunity to propose an amended Supplemental Report to clarify Fischel's position. Fischel used this opportunity to explain that it was unnecessary to disaggregate Searle's and Pharmacia's misrepresentations from Pfizer's because Plaintiffs' theory of liability was that Pfizer's misrepresentations and omissions during the Class Period maintained the market's "false expectations" about the safety of Celebrex and Bextra — false expectations already in place prior to the Class Period, but kept in place after the Class Period in part by Pfizer's misrepresentations. J.A. 2182-83. Because Fischel assumed that Plaintiffs would be able to prove at trial that Pfizer's alleged misrepresentations during the Class Period concealed the same safety information that was concealed just prior to the Class Period, Fischel averred, each of Pfizer's misrepresentations was responsible for maintaining the same amount of artificial inflation. Fischel also described the Proportional Reduction in greater detail, and noted that, if the Court disagreed with his methodology, it should exclude only that reduction, rather than all of his testimony.

The district court considered Plaintiffs' submissions and, on July 8, 2014, denied their motion to amend the Supplemental Report. The court was "not persuaded" that the proposed amendments "were sufficient to meet the standards of Federal Rule of Evidence 702." S.A. 35. And even if they were, Plaintiffs had "no justification for the untimely disclosure of the additional explanations and contextual information." S.A. 35. The court also found Fischel's proposal to exclude only his proportional reduction to the Positive Disclosures no more persuasive, explaining that his "offer to permit the Court to select which of two putatively valid inflation determination methodologies to present to the jury plainly demonstrates that the proffered testimony is not deserving of an `expert opinion' label." S.A. 36. Without a loss causation and damages expert, Plaintiffs could not hope to establish liability at trial, so the district court granted summary judgment to Pfizer. This appeal followed.

DISCUSSION

On appeal, Plaintiffs contend that the district court erred by excluding Fischel from testifying. First, they argue that Fischel's failure to disaggregate the impact of Pfizer's and other companies' misrepresentations is not a basis for excluding his testimony because, contrary to the district court's summary judgment decision, Pfizer had ultimate authority over those statements. Even if it lacked such authority, they continue, Fischel did not need to account for Searle's and Pharmacia's statements in order to be helpful to the jury because, under Plaintiffs' inflation-maintenance theory, Pfizer concealed the same information as Searle and Pharmacia and is therefore liable for the full extent of Plaintiffs' alleged losses. Second, Plaintiffs contend that Fischel's proportional reduction to the Positive Disclosures was reliable and nothing more than an update to the data in his event study and that the district court should not have excluded Fischel's testimony on this basis.

We conclude that, based on the record before the district court after Pfizer's motion in limine was fully submitted, the court abused its discretion by excluding Fischel's testimony in its entirety. In the context of Plaintiffs' inflation-maintenance theory, even assuming, arguendo, that Pfizer lacked sufficient authority over several Searle and Pharmacia statements, that would not require Fischel to disaggregate the effects of Pfizer's alleged misstatements from those of Searle and Pharmacia in order to help the jury calculate the amount of losses that Pfizer allegedly caused. Similarly, although the district court did not abuse its discretion in concluding that Fischel's Proportional Reduction was unreliable, that conclusion did not justify excluding the remainder of his testimony. Instead, the district court should have prevented Fischel from testifying about the Proportional Reduction and otherwise allowed him to present his findings on loss causation and damages.6 We also find that the district court was incorrect to conclude that, as a matter of law, Pfizer lacked sufficient authority over eight statements by Searle and Pharmacia employees to the press, as to have "made" those statements for the purposes of Rule 10b-5 liability. But we note that the district court abused its discretion in excluding Fischel's testimony, regardless of whether Pfizer ultimately is liable for those statements.

We emphasize that our decision is a narrow one. The district court granted summary judgment after excluding Fischel from testifying pursuant to Rule 702. We therefore consider only whether that decision was an abuse of discretion in light of the evidence presented to the district court and the Plaintiffs' theory of the case. We need not, and do not, decide whether Plaintiffs' inflation-maintenance theory is either legally sustainable under Rule 10b-5 or sufficiently supported by the evidence in the record.

A. Searle's and Pharmacia's Statements

Our analysis begins with the district court's determination that Pfizer can be liable only for one Pharmacia press release about Celebrex, and not for nine other allegedly fraudulent public statements made by Searle and Pharmacia employees about the drug. Because the district court resolved this issue on a motion for summary judgment, we review its conclusion "de novo, resolving all ambiguities and drawing all permissible factual inferences in favor of" Plaintiffs. Burg v. Gosselin, 591 F.3d 95, 97 (2d Cir.2010) (internal quotation marks omitted). Viewing the facts in this favorable light, we conclude that Plaintiffs have raised a genuine dispute as to whether Pfizer had sufficient authority over certain statements that individual Searle and Pharmacia employees made to various media outlets regarding the risks associated with Celebrex.7

Under Rule 10b-5, it is "unlawful for any person, directly or indirectly, ... [t]o make any untrue statement of a material fact" in connection with the purchase or sale of securities. 17 C.F.R. § 240.10b-5. "To be liable, therefore, [a defendant] must have `made' the material misstatement[ ]" at issue. Janus Capital Grp. v. First Derivative Traders, 564 U.S. 135, 131 S.Ct. 2296, 2301, 180 L.Ed.2d 166 (2011); see also Fezzani v. Bear, Stearns & Co., 716 F.3d 18, 24-25 (2d Cir.2013). As the Supreme Court explained in Janus, "the maker of a statement is the person or entity with ultimate authority over the statement, including its content and whether and how to communicate it." Janus, 131 S.Ct. at 2302. A defendant does "not `make' a statement" simply by "prepar[ing] or publish[ing][it] on behalf of another." Id. Thus, in that case, a mutual fund investment advisor who "was significantly involved in preparing [a client's] prospectuses" did not "make" the statements contained therein. Id. at 2304-05. Only the client had the "statutory obligation to file the prospectuses" and nothing in the document "indicate[d] that any statements ... came from" the defendant rather than its client. Id. A "broader reading of `make,'" the Court cautioned, would "substantially undermine" the rule from Central Bank of Denver, N.A. v. First Interstate Bank of Denver, N.A., 511 U.S. 164, 114 S.Ct. 1439, 128 L.Ed.2d 119 (1994), which prohibits private suits against aiders and abettors of Rule 10b-5 violations. Id. at 2302; see also Fezzani, 716 F.3d at 24-25.

Here, the district court found — and Pfizer does not contest — that Pfizer's participation in the Co-Promotion Agreement with Searle (later Pharmacia) created a genuine dispute of material fact about whether the company had "ultimate authority" over one of Pharmacia's press releases about Celebrex. In re Pfizer I, 936 F.Supp.2d at 271. Plaintiffs claim that Pfizer's involvement ran deeper: Pfizer, they argue, also had "ultimate authority" over statements in a Form 8-K that Pharmacia filed with the Securities and Exchange Commission ("SEC"), as well as eight statements about the drugs that Searle and Pharmacia employees made to members of the media. Although Pfizer does not contest the district court's finding that a reasonable jury could conclude that Pfizer had the right to approve press releases, Pfizer argues that there is no evidence that Pfizer had similar authority over the Form 8-K, and that its right to approve did not extend to media statements made by Searle and Pharmacia employees. The record offers no support for Plaintiffs' argument with respect to the Form 8-K statements, but does provide sufficient evidence to create a genuine dispute as to the eight other statements.

As to the statements in Pharmacia's Form 8-K, Plaintiffs have failed to identify any evidence that raises a genuine dispute as to Pfizer's authority over these statements. The Co-Promotion Agreement, unsurprisingly, dealt with the signatories' efforts to market Celebrex. But Plaintiffs have not pointed to anything in the record that might allow a reasonable jury to conclude that Pfizer influenced, much less had any authority over, Pharmacia's Form 8-K, a regulatory disclosure required by the SEC to inform the public of specific events that may be material to investors. See 17 C.F.R. § 249.308. On the contrary, the Co-Promotion Agreement stipulated that Searle (later Pharmacia) would have "sole responsibility for communicating with ... regulatory authorities" about Celebrex, which naturally runs counter to any claim that Pfizer was responsible for regulatory filings like the Form 8-K. U.S. Agreement at 12, Ex. 33 to Pls.' Opp. to Defs.' Motion for Summ. J., No. 1:04-cv-09866, ECF 439. Summary judgment as to the Form 8-K statements was therefore wholly appropriate.

Plaintiffs' evidence of Pfizer's authority over the eight statements by Searle and Pharmacia employees to various newspaper and journal articles fares slightly better, however. In broad terms, the eight statements all conveyed that there were no increased cardiovascular risks associated with Celebrex. To be clear, there is no dispute that Searle and Pharmacia employees, not Pfizer employees, actually delivered the statements to the press. Nor is there any evidence that these employees held themselves out as representing Pfizer. And "in the ordinary case," the fact that the statements were attributed to Searle or Pharmacia employees "is... strong evidence that [the] statement[s] w[ere] made by — and only by — the party to whom [they were] attributed." Janus, 131 S.Ct. at 2302. Nevertheless, we find that there is a material question of fact whether the present case deviates from the ordinary case. Notwithstanding that the eight statements to the press were attributed to Searle and Pharmacia employees, Plaintiffs have presented sufficient evidence to permit a reasonable jury to conclude that Pfizer had "ultimate authority" over the statements' "content and whether and how to communicate" them. Id.

To start, Plaintiffs cite to a fax sent by a public relations firm, jointly employed by Pfizer and Searle, to various Pfizer and Searle employees. See ECF 439, Ex. 110. Included in the fax is a "Q & A document" containing questions the press might pose regarding cardiovascular risks associated with Celebrex, along with scripted answers, id. at 5-9, some of which convey effectively the same content as the statements at issue, compare id. at 5, 9, with, e.g., J.A. 271, ¶ 348. The "Q & A document," the fax explains, would be "reviewed and finalized during a ... conference call" between Pfizer and Searle, before being "distributed ... to the appropriate parties at Searle and Pfizer for final sign-off." See ECF 439, Ex. 110, at 1 (emphasis added). This evidence is consistent with deposition testimony by Andrew McCormick, a Pfizer senior management team member, also cited by Plaintiffs, in which McCormick avers that "senior management" at Pfizer "would... need to ... approve[ ]" of "media responses [related to COX-2 inhibitors], both to reporters and to publications in the forms of letters to the editors" in order to "get [them] out the door." ECF 439, Ex. 88 at 52-53. Although the evidence is not abundant, we are unable to conclude that no reasonable jury could find from this evidence that Pfizer had "ultimate authority" over the eight statements to the press.

Plaintiffs also rely on Pfizer's Co-Promotion Agreement with Searle (later Pharmacia), pursuant to which Pfizer and Searle agreed to coordinate a wide range of promotional activity concerning Celebrex. See ECF 439, Ex. 33. The meaning of the Co-Promotion Agreement, ambiguous or unambiguous, is not dispositive of the question before us at summary judgment: whether a genuine issue of material fact exists as to Pfizer's authority over the eight Searle and Pharmacia statements to the press. Assuming, arguendo, that the Co-Promotion Agreement unambiguously fails to provide Pfizer the power to approve or disapprove of Searle's and, later, Pharmacia's statements to the press, other evidence still creates a question of fact as to whether, notwithstanding any procedures articulated in the contract, Pfizer in fact had "ultimate authority" over these statements. The meaning of the Co-Promotion Agreement is thus but one factor in the factual determination whether Pfizer indeed had "ultimate authority" over the eight statements to the media. Because we conclude that material questions of fact preclude summary judgment regardless of the correct interpretation of the Co-Promotion Agreement, we need not interpret it in this appeal.

We therefore conclude that the district court erred in determining that, as a matter of law, Pfizer lacked sufficient authority over the eight allegedly fraudulent statements that Searle and Pharmacia employees made to the media, such that Pfizer could not have "made" the statements for the purposes of Rule 10b-5. We therefore vacate the district court's grant of summary judgment to Pfizer with respect to those statements. As stated above, we leave undisturbed the district court's ruling regarding Pharmacia's Form 8-K statement.

B. Fischel's Testimony

Plaintiffs' principal arguments focus on whether the district court's decision to prevent Fischel from testifying at trial was a proper application of Federal Rule of Evidence 702, which governs the admissibility of expert testimony. Under Rule 702, an expert with "specialized knowledge [that] will help the trier of fact" may testify so long as that testimony is "based on sufficient facts or data" and "is the product of reliable principles and methods" that the witness has "reliably applied ... to the facts of the case." The proponent of the expert testimony has the burden to establish these admissibility requirements, with the district court acting as a "gatekeeper" to ensure that the "`expert's testimony both rests on a reliable foundation and is relevant to the task at hand.'" United States v. Williams, 506 F.3d 151, 160 (2d Cir.2007) (quoting Daubert v. Merrell Dow Pharms., Inc., 509 U.S. 579, 597, 113 S.Ct. 2786, 125 L.Ed.2d 469 (1993)).

The district court has broad discretion to carry out this gatekeeping function. Its inquiry is necessarily a "flexible one," Daubert, 509 U.S. at 594, 113 S.Ct. 2786, and the types of factors that are appropriate to consider will "depend[ ] upon the particular circumstances of the particular case at issue," Kumho Tire Co. v. Carmichael, 526 U.S. 137, 150, 119 S.Ct. 1167, 143 L.Ed.2d 238 (1999). We therefore review both the district court's "ultimate reliability determination" and its decision about "how to determine reliability" for abuse of discretion. Id. at 142, 119 S.Ct. 1167 (emphasis in original); see also Amorgianos v. Nat'l R.R. Passenger Corp., 303 F.3d 256, 265 (2d Cir.2002). But "while the district court's discretion is considerable, it is not unfettered." Williams, 506 F.3d at 160. The "gatekeeping" function under Daubert is fundamentally about "ensur[ing] the reliability and relevancy of expert testimony," and district courts may not stray from those goals. Kumho Tire, 526 U.S. at 152, 119 S.Ct. 1167.

In this case, the district court provided two reasons for excluding Fischel's testimony under Rule 702, both of which stemmed from its summary judgment decision. For one, the district court noted that Fischel's Supplemental Report did not account for the district court's determination that Pfizer was not liable for many of Searle's and Pharmacia's alleged misrepresentations about Celebrex. This oversight, it concluded, "renders his opinions unhelpful to the jury in making calculations of damages proximately caused by [Pfizer's] alleged misrepresentations and omissions." In re Pfizer II, 2014 WL 2136053, at *1. In addition, the court found no "analytical basis" for Fischel's proportional reduction of the Positive Disclosures, and concluded that it had not "been shown to be the product of reliable principles and methods reliably applied." Id. We address each of these reasons in turn and conclude that neither provides a sufficient basis to exclude Fischel's testimony in its entirety. Our conclusion does not depend on Pfizer's potential liability for the eight statements to the press by Searle or Pharmacia.

1. Disaggregation of Statements

The district court's first rationale for excluding Fischel's testimony — that he failed to account for the district court's prior determination that Pfizer could not be liable for certain statements by Searle or Pharmacia — rests on a misapprehension of Fischel's role in Plaintiffs' claim. As Plaintiffs explained in their opposition to Pfizer's motions for summary judgment (and again in opposition to Pfizer's motion in limine), their theory of liability is that Pfizer's misrepresentations and omissions regarding the cardiovascular risks associated with Celebrex and Bextra maintained Pfizer's inflated share price by keeping those risks hidden from the market. Under that theory, they argue that there is no need to separately account for Searle's and Pharmacia's misrepresentations: so long as Pfizer's own fraudulent conduct kept the same information concealed from the public, it is liable for the full value of that information to investors. Fischel's testimony, then, is important because it validates that the market reacted to information about the risks associated with Celebrex and Bextra and calculates the amount of artificial inflation in Pfizer's stock assuming that the company had been concealing those risks from the beginning of the Class Period. Fischel did not purport to analyze how inflation entered Pfizer's stock prior to the Class Period because, under Plaintiffs' theory, Pfizer may be liable for the full value of that inflation regardless of how it got there.

A closer look at Plaintiffs' inflation-maintenance theory makes Fischel's role clear. Since Basic Inc. v. Levinson, 485 U.S. 224, 108 S.Ct. 978, 99 L.Ed.2d 194 (1988), securities class actions under Rule 10b-5 have relied on the idea that, "`in an open and developed securities market, the price of a company's stock is determined by the available material information regarding the company and its business.'" Id. at 241, 108 S.Ct. 978. According to Plaintiffs, a corollary to this idea is that fraudulent statements containing the same falsehoods as earlier misrepresentations "will not cause a change in the stock price... because the market has already digested that information and incorporated it into the price." FindWhat Investor Grp. v. FindWhat.com, 658 F.3d 1282, 1310 (11th Cir.2011). "[T]he inflation level need not change," however, "for new investors to be injured by a false statement." Id. at 1314. If misrepresentations or omissions "prevent a stock price from falling," they can cause harm "by prolonging the period during which the stock is traded at inflated prices." Id. (first emphasis omitted); see also In re Vivendi Universal, S.A. Secs. Litig., 765 F.Supp.2d 512, 562 (S.D.N.Y. 2011). Thus, Plaintiffs argue, Pfizer's misrepresentations about Celebrex and Bextra need not be distinguished from any statements by Searle or Pharmacia, because Pfizer prolonged the period during which the same information about cardiovascular safety was concealed from the market.

Fischel's Initial and Supplemental Reports must be evaluated within the context of this inflation-maintenance theory. See Daubert, 509 U.S. at 591, 113 S.Ct. 2786 (discussing the requirement that the expert testimony "`fit[s]'" the plaintiff's theory of the case). As Fischel stated in his Initial Report, his task was to "analyze the amount of alleged artificial inflation in Pfizer's stock price during the Class Period" under the assumption that Plaintiffs "will be able to establish at trial that Defendants withheld material information about the cardiovascular safety of Celebrex and Bextra." J.A. 927 & n. 6; see also 1111. To that end, he opined that, "[u]ntil late 2004, investors believed that any cardiovascular safety concerns regarding Celebrex and Bextra were limited," J.A. 929, identified disclosures related to those risks that caused Pfizer's stock price to change, and then calculated the artificial inflation in Pfizer's stock price assuming that Pfizer had, in fact, concealed those risks. This approach does not "directly measure inflation caused by false statements"; instead, it "measure[s] the value of the truth" that the market eventually discovered. Glickenhaus & Co. v. Household Int'l, Inc., 787 F.3d 408, 416-17 (7th Cir.2015). And under Plaintiffs' inflation-maintenance theory, the inflation caused by Pfizer's misrepresentations and omissions is "equal to the value of the truth ... because had [its] statement[s] been truthful, the stock price would have done what it did do once the truth was revealed." Id. at 417; cf. Lentell v. Merrill Lynch & Co., 396 F.3d 161, 173 (2d Cir.2005) ("[T]o establish loss causation, a plaintiff must allege that the subject of the fraudulent statement or omission was the cause of the actual loss suffered ...." (emphasis, ellipses, and internal quotation marks omitted)); In re Omnicom Grp., Inc. Sec. Litig., 597 F.3d 501, 513 (2d Cir.2010) ("A misrepresentation is the proximate cause of an investment loss if the risk that caused the loss was within the zone of risk concealed by the misrepresentations." (emphasis and internal quotation mark omitted)). Fischel was not asked to provide, and did not offer, a way to identify how inflation entered the stock or to disaggregate the effect of specific statements made before and during the Class Period.

In denying Pfizer's summary judgment motion, the district court stated that "a misstatement may cause inflation simply by maintaining existing market expectations." In re Pfizer I, 936 F.Supp.2d at 264 (citing FindWhat.com, 658 F.3d at 1310) (emphasis added) (internal quotation marks omitted). As a result, Fischel did not need to address the district court's decision that Pfizer is not responsible for nine of Searle's and Pharmacia's misstatements in his Supplemental Report. He explained as much during his second deposition, stating that this portion of the district court's decision "didn't affect [his] analysis of inflation," and that his analysis "assum[es] that [P]laintiffs will be able to prove their liability allegations" but that he does not offer any opinion as to who was responsible for the inflation in Pfizer's stock price at the start of the Class Period.8 J.A. 1120-21. Plaintiffs then reiterated this point in opposition to Pfizer's motion in limine: "Fischel's model ... fits Plaintiffs' allegations that ... [Pfizer] made dozens of misrepresentations which effectively repeated the same false message to the market ... and served to maintain Pfizer's stock price at a constant, inflated level." J.A. 1524. As a result, he did not "assign damages on a statement-by-statement basis." J.A. 1522.

Given the scope of Fischel's opinion, it was an abuse of discretion to prevent him from testifying on the grounds that he did not disaggregate the stock price inflation caused or maintained by Pfizer's own statements from that caused or maintained by Searle's and Pharmacia's statements. At bottom, the district court's decision rests on the idea that, if Plaintiffs succeed, the jury will have to attribute specific amounts of inflation to Pfizer in order to calculate "damages proximately caused by [Pfizer's] alleged misrepresentations and omissions." In re Pfizer II, 2014 WL 2136053, at *1. But as discussed above, Plaintiffs' theory is directly contrary to this idea: they argue that Pfizer is liable for all of the artificial inflation related to Celebrex and Bextra because, through its own fraudulent conduct, Pfizer concealed the same information as its predecessors. In the context of that theory, Fischel's testimony can be helpful to the jury without disaggregating the effects of Pfizer's specific misrepresentations because it shows that the discovery of information Pfizer allegedly concealed caused shareholders to lose money and calculates the amount of money they lost.

To be clear, we do not decide in this appeal that Plaintiffs' theory is either legally or factually sustainable. It might be that Plaintiffs' inflation-maintenance theory is deficient under Rule 10b-5 or that, even under such a theory, Plaintiffs cannot establish, as they contend, that Pfizer concealed the same information as Searle and Pharmacia hid. But these are reasons why Plaintiffs' claim may be legally or factually deficient, not justifications for concluding that, in the context of Plaintiffs' theory, Fischel's testimony is unreliable or unhelpful to the jury.9 See Ambrosini v. Labarraque, 101 F.3d 129, 135 (D.C.Cir. 1996) ("The dispositive question [under Rule 702] is whether the testimony will assist the trier of fact ... not whether the testimony satisfies the plaintiff's burden on the ultimate issue at trial." (internal quotation marks omitted)). Daubert and Rule 702 merely authorize the court to ensure that the "`expert's testimony both rests on a reliable foundation and is relevant to the task at hand.'" Williams, 506 F.3d at 160 (quoting Daubert, 509 U.S. at 597, 113 S.Ct. 2786). Put differently, Fischel's failure to disaggregate the effects of Pfizer's, Searle's, and Pharmacia's statements renders his testimony unhelpful only if Plaintiffs need to disaggregate those statements to succeed at trial. Accordingly, we conclude that Fischel's failure to distinguish inflation associated with Pfizer's misrepresentations from inflation associated with Searle's and Pharmacia's statements was, in context, an inadequate basis on which to exclude his testimony.

2. The Proportional Reduction

The district court's second basis for excluding Fischel's testimony relates to his proportional reduction of the Positive Disclosures. As already explained, in his Supplemental Report, Fischel removed the stock-price declines associated with two Corrective Disclosures after the district court decided that Pfizer's alleged fraud could not reasonably be found to cause the losses on those days. Because removing those disclosures caused the overall value of the Corrective Disclosures to fall by 9.7 percent, Fischel also reduced the amount of the stock price changes associated with Positive Disclosures that he attributed to the fraud by 9.7 percent. He described his rationale for this Proportional Reduction in a single sentence contained in the Supplemental Report. The district court decided that this Proportional Reduction rendered Fischel's opinion unreliable because he "proffered no explanation of [its] analytical basis" and failed to show that it was "the product of reliable principles and methods reliably applied." In re Pfizer II, 2014 WL 2136053, at *1.

The district court's conclusion that Fischel did not adequately explain his proportional adjustment to the Positive Disclosures was not an abuse of discretion. But Fischel's failure adequately to explain the Proportional Reduction was not a basis for excluding his testimony in its entirety. Instead, the court should have prevented Fischel from testifying about the Proportional Reduction and otherwise permitted him to offer his opinion about loss causation and damages.

At the outset, the district court's decision to scrutinize Fischel's Supplemental Report when assessing whether his testimony satisfies Rule 702's requirements was consistent with its gatekeeping function under Daubert. As we explained in Amorgianos, Rule 702 directs the district court to "focus on the principles and methodology employed by the expert, without regard to the conclusions the expert has reached or the district court's belief as to the correctness of those conclusions." 303 F.3d at 266. Part of that inquiry, however, requires assessing whether the expert "appl[ied] his stated methodology reliably to the facts of the case." Id. at 269 (internal quotation marks omitted). If the opinion "is based on data, a methodology, or studies that are simply inadequate to support the conclusions reached, Daubert and Rule 702 mandate the exclusion of that unreliable opinion testimony." Id. at 266; see also General Elec. Co. v. Joiner, 522 U.S. 136, 146, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997).

Fischel's Supplemental Report involves precisely the application of methodology to facts that district courts must scrutinize under Rule 702. In the Initial Report, Fischel presented his event study methodology, identified both Corrective and the Positive Disclosures, and used the price movement on the days of those disclosures to calculate the artificial inflation in Pfizer's stock. When the district court concluded that Pfizer was not responsible for the stock-price changes associated with two of those Corrective Disclosures, Fischel had to adjust his calculations. According to Plaintiffs, Fischel's Supplemental Report made this adjustment simply by updating the data he used in his event study analysis — a type of analysis that Pfizer's expert conceded was a valid way to calculate artificial inflation. We agree that, because the district court had determined that the stock-price movements on the two excluded Corrective Disclosure dates did not occur because of Pfizer's fraud, Fischel clearly needed to remove the price changes associated with those disclosures from his calculation of artificial inflation. But this narrative tells only half the story.

The issue of the Positive Disclosures was a more complicated matter. Fischel's Initial Report identified Positive Disclosures as days on which Pfizer's stock price rose because of new information about Celebrex and Bextra. The district court's summary judgment decision left this portion of the Initial Report untouched. Nonetheless, Fischel independently decided that the Positive Disclosures needed revisiting in the Supplemental Report. According to Fischel, even though the two excluded Corrective Disclosures were not actually "corrective" because they did not reveal concealed information about cardiovascular risks associated with Celebrex and Bextra, the price movements on those days might nonetheless be related to information about the "cardiac issues" linked to those drugs. J.A. 1109. If that were the case, some of the price changes on Positive Disclosure days could reflect the market recalibrating its evaluation of that non-fraud-related cardiac information. See J.A. 1110. Such recalibration would not be related to the value of the fraudulently concealed information, and therefore should not, in effect, reduce the calculation of artificial inflation in Pfizer's stock. See J.A. 1112 (noting that removing the Corrective Disclosures "changes my analysis about what the appropriate way to ... allocate the residual price change between what is fraud related and what's not fraud related...."). On the other hand, it might be the case that the Positive Disclosures related only to fraudulently concealed information, or that the price changes associated with the two excluded Corrective Disclosures had nothing to do with new information about Celebrex and Bextra, and were thus entirely unrelated to the Positive Disclosures. In either of those situations "it would be perfectly appropriate to eliminate [the price declines associated with the excluded Corrective Disclosures] ... and not make any other adjustments." J.A. 1109; see also J.A. 1115.

A hypothetical helps illustrate the point. Suppose Pfizer did, in fact, conceal information about the cardiovascular risks associated with Celebrex and Bextra. On Monday, some of that concealed information leaks, causing Pfizer shares to fall by $2. The next day, the market learns of a new study about such risks from Celebrex and Bextra — one that had not been concealed by Pfizer and could not be attributed to its fraud — and the company's shares fall by another $2. Finally, on Wednesday, the market discovers that the cardiovascular risks from the drugs are not as severe as first expected, and Pfizer's stock price rises $1 on the news.

Certainly, according to Fischel's principles, the Monday price decline should be included when assessing the amount by which Pfizer's fraud inflated its share price and the Tuesday price decline should not. But what should be made of the $1 price increase? It could be that the market overestimated the significance of the concealed information, in which case the entire price increase should be included when calculating the inflation caused by that fraud. Or it could be that the market overestimated the significance of both the concealed information and the new study, in which case only a portion of the $1 increase would be related to the fraudulently concealed information. Under Fischel's mode of analysis, then, the expert needs to determine the relationship between the disclosures and then apportion the proper amount of the price increase, if any, to his evaluation of the fraud.

What is clear, then, is that the Proportional Reduction contained in Fischel's Supplemental Report was based on an implicit conclusion regarding the relationship between the Corrective and Positive Disclosures. Indeed, during the second set of depositions Fischel repeatedly described his process for arriving at the Proportional Reduction as based on a "methodology" for evaluating the "relationship between" the Corrective and Positive Disclosures. J.A. 1113-14, 1110. Thus, Fischel drew just the sort of link between the facts and his conclusion that district courts must evaluate for compliance with Rule 702, see Amorgianos, 303 F.3d at 267; Ruggiero v. Warner-Lambert Co., 424 F.3d 249, 254-55 (2d Cir.2005), and we agree with the court's decision to scrutinize his analysis under its gatekeeping function.

In performing that duty, the district court did not abuse its discretion by deciding that Fischel's 9.7 percent Proportional Reduction was not "the product of reliable principles and methods reliably applied." In re Pfizer II, 2014 WL 2136053, at *1. To begin, Fischel's decision to reevaluate the Positive Disclosures is in some tension with his earlier testimony. During his first deposition, Pfizer's attorneys asked him to evaluate a hypothetical situation in which the district court concluded that one Corrective Disclosure did not actually correct a prior fraudulent statement or omission. J.A. 1135. Fischel's response suggested that he would simply remove the offending disclosure — a conservative approach that assumes no link between the Positive Disclosures and the removed Corrective Disclosures. Fischel never mentioned that he might reconsider the Positive Disclosures as well.

Moreover, even overlooking this inconsistency, it was not an abuse of discretion to decide that Fischel's Proportional Reduction was not a sufficiently reliable application of his stated methodology to pass muster under Rule 702. As explained above, the general methodology that Fischel outlined calls for evaluating the "relationship between" the excluded Corrective Disclosures and the Positive Disclosures to determine how to "allocate the" price changes on Positive Disclosure days "between what is fraud related and what[ ] [is] not fraud related." J.A. 1112. Assuming this methodology is a reliable way to assess the effect of the Positive Disclosures on the value of Pfizer's alleged fraud,10 Fischel did not reliably apply it in this case. Quite simply, Fischel never "offered" an "explanation of the relationship[ ]" between the Positive Disclosures and the excluded Corrective Disclosures. In re Pfizer II, 2014 WL 2136053, at *1. Neither his Supplemental Report nor his second deposition included an analysis of whether, in light of the district court's decision, Pfizer's stock price changed on the excluded Corrective Disclosure days because of information about Celebrex and Bextra. Similarly, Fischel never explained whether, or how, the information in the Positive Disclosures related to those excluded Disclosures. Indeed, in the one comment he made on the issue, Fischel said he could not find "any clear one-for-one relationship between any particular [P]ositive [D]isclosure and any particular [Corrective] [D]isclosure." J.A. 1114.

Instead, Fischel asserted that the exclusion of Corrective Disclosures must have "ha[d] an effect on the [P]ositive [D]isclosures because inflation cannot come out of a stock unless it goes into the stock." J.A. 1110. But in addition to stock price increases on the Positive Disclosure days, Fischel quantified artificial inflation in Pfizer's stock price at the beginning of the Class Period on the assumption that fraudulent statements and omissions prior to the Class Period's beginning had already inflated Pfizer's stock price. Fischel calculated the amount of this artificial inflation by observing how Pfizer's stock price moved as the concealed information became public — a process that involved summing the price changes associated with the Corrective and Positive Disclosures. Excluding two Corrective Disclosures from that calculation automatically reduces the amount of artificial inflation present at the beginning of the Class Period, and for good reason: if two Corrective Disclosures did not involve concealed information, the price changes associated with them do not reflect the value of that information. Thus, the mere fact that the district court excluded two Corrective Disclosures does not necessarily require an adjustment to the Positive Disclosures, much less an adjustment that is proportional to the overall decline in value of the Corrective Disclosures. Instead, Fischel's stated methodology required evaluating the relationship between Corrective and Positive Disclosures precisely because it might have been "perfectly appropriate to eliminate [the excluded Corrective Disclosures] ... and not make any other adjustments." J.A. 1109.

If an opinion is based on "a methodology" that is "simply inadequate to support the conclusions reached, Daubert and Rule 702 mandate the exclusion of that unreliable opinion testimony." Amorgianos, 303 F.3d at 266. After all, "nothing in either Daubert or the Federal Rules of Evidence requires a district court to admit opinion evidence that is connected to existing data only by the ipse dixit of the expert." Joiner, 522 U.S. at 146, 118 S.Ct. 512. In light of the "analytical gap" between Fischel's stated methodology and the manner in which he assessed the relationship between the excluded Corrective Disclosures and the Positive Disclosures, Amorgianos, 303 F.3d at 266 (internal quotation marks omitted), the district court did not abuse its discretion by deciding that the Proportional Reduction was unreliable and therefore inadmissible under Rule 702.

This brings us to the final issue we consider: whether the district court's decision that Fischel's Proportional Reduction was unreliable justified excluding his testimony in its entirety. When faced with expert testimony that contains both reliable and unreliable opinions, district courts often exclude only the unreliable testimony. See, e.g., Laumann v. Nat'l Hockey League, 117 F.Supp.3d 299, 303-04, 328-29 (S.D.N.Y.2015); Fed. Hous. Fin. Agency v. Nomura Holding Am., Inc., 2015 WL 640900, at *4-*5 (S.D.N.Y. Feb. 16, 2015); Vazquez v. City of N.Y., 2014 WL 4388497, at *13 (S.D.N.Y. Sept. 5, 2014). This process of parsing expert testimony is consistent with Rule 702's "liberal admissibility standards," Amorgianos, 303 F.3d at 267, which favor allowing the jury to hear testimony that "both rests on a reliable foundation and is relevant to the task at hand," Daubert, 509 U.S. at 597, 113 S.Ct. 2786. Of course, district courts are "not obligated to prune away all of the problematic" elements of an expert's proposed testimony "to save the remaining portions, however small." Bricklayers & Trowel Trades Int'l Pension Fund v. Credit Suisse Sec. (USA) LLC, 752 F.3d 82, 96 (1st Cir.2014). But when the unreliable portion of an opinion can easily be distinguished from testimony that could help the jury, it may be an abuse of discretion to throw the good out with the bad. See City of Tuscaloosa v. Harcros Chems., Inc., 158 F.3d 548, 564 (11th Cir.1998).

Here, we conclude that, although the district court did not abuse its discretion by finding the Proportional Reduction unreliable, it went astray when it excluded all of Fischel's testimony on this basis. Fischel's Proportional Reduction was but one small part of an extensive economic analysis. In the Initial Report, Fischel performed an event study to identify days on which allegedly concealed information about Celebrex and Bextra caused Pfizer's stock price to change and to calculate the value of the information that the company had supposedly been concealing. Pfizer's expert, Dr. Paul A. Gompers, critiqued particular elements of that analysis, but "overall" did not have "any major criticism of [Fischel's] event study." J.A. 1645.

Fischel's Proportional Reduction was a response to a new issue created by the district court's summary judgment decision. As explained above, excluding two Corrective Disclosures created the possibility that some price changes on the Positive Disclosure days involved reactions to non-fraud-related information about the safety of Celebrex and Bextra. Thus, whereas the event study constituted a broad statistical analysis aimed at identifying whether and how the market responded to new information about the cardiovascular risks from Celebrex and Bextra, the summary judgment decision prompted a narrow inquiry into the particular relationship between the Positive Disclosures and the excluded Corrective Disclosures.

Significantly, both Gompers and Pfizer challenged only the Proportional Reduction and identified a specific, alternative course of action that they believed Fischel should have taken. For instance, in his Supplemental Response, Gompers focused exclusively on the lack of academic support or "an economically-sound justification for [the] 9.7% adjustment," J.A. 1154, but did not offer a broader criticism of the event study. Gompers contrasted the Proportional Reduction with simply removing the Corrective Disclosures at issue — a process that he labeled Fischel's "original methodology" for handling excluded disclosures because of Fischel's testimony in his first deposition. J.A. 1151, 1156. Pfizer made a nearly identical argument in its memorandum of law in support of its motion in limine. There, the company criticized the 9.7 percent "adjustment" and explained that, according to Fischel's first deposition, "the Court's rejection of [the two Corrective Disclosure dates] simply required him to remove them from his estimate of the inflation." J.A. 1475. Pfizer added that "removing the [Corrective Disclosure dates] ... was by itself sufficient to achieve equilibrium" between inflation entering and leaving the stock. J.A. 1476. Plaintiffs pointed out the specificity of Pfizer's critique in their opposition to the motion in limine, noting that "Defendants contend Fischel should have ... reduc[ed inflation at the beginning of the Class Period] only" and should not have reduced price changes associated with the Positive Disclosures. J.A. 1508.

On these facts, the district court abused its discretion by excluding all of Fischel's testimony. Although the Proportional Reduction may have been unreliable, it was unreliable because Fischel failed to show that the 9.7 percent of the price changes associated with Positive Disclosures was, in fact, related to the excluded Corrective Disclosures. Fischel's error on this issue, however, does not cast a pall over his opinion — arrived at through the event study — that Pfizer's stock price fell during the Class Period when the market learned of information about the cardiovascular risks associated with Celebrex and Bextra. Nor does it call into question his methodology for identifying, and calculating the residual returns associated with, the Positive and remaining Corrective Disclosures. All it means is that Fischel could not reliably associate some of the residual returns on the Positive Disclosure days with information revealed through the excluded Corrective Disclosures, rather than with Pfizer's alleged fraud. As the parties' motion papers reveal, eliminating Fischel's Proportional Reduction does not render the remainder of his analysis useless; instead, it merely ensures the adoption of the most conservative estimate of the losses Pfizer allegedly caused. Thus, rather than excluding all of Fischel's testimony, the district court should simply have prevented him from making the Proportional Reduction. The remainder of his testimony "both rests on a reliable foundation and is relevant to the task at hand." Daubert, 509 U.S. at 597, 113 S.Ct. 2786. Excluding it was an abuse of discretion.11

CONCLUSION

We conclude that the district court erred in granting summary judgment to Pfizer on the issue of Pfizer's liability for certain Searle and Pharmacia statements. We also conclude that the district court erred in granting summary judgment to Pfizer based on its exclusion of Fischel's testimony, which we find was an abuse of discretion. The judgment of the district court is therefore VACATED and the case is REMANDED for further proceedings consistent with this opinion.