image description

Role of the Event Study in Loss Causation Analysis

by Martis Alex and Michael W. Stocker
New York Law Journal |
Proof of loss causation, that is, a causal connection between material misrepresentations and investor losses, was a bone of contention in securities fraud cases long before the current downturn in the markets.

The recent havoc in the markets might make it seem a propitious time to be a securities lawyer. However, plummeting stock prices present their own set of challenges to plaintiffs pursuing fraud claims under the federal securities laws. Even when there appears to be ample evidence that executives misled investors about business prospects, market wide share price declines can make it difficult to prove that the revelation of the truth about these deceptions caused the losses suffered by shareholders. As a result, lawyers and the courts are increasingly focused on the proper methodology for determining whether investor losses are caused by fraud. This new scrutiny has highlighted the benefits-and significant limitations-of the event study, long the mainstay for loss causation analysis.

Proof of loss causation, that is, a causal connection between material misrepresentations and investor losses, was a bone of contention in securities fraud cases long before the current downturn in the markets. In Dura Pharmaceuticals, Inc. v.Broudo, 544 U.S. 336 (2005), the U.S. Supreme Court reviewed a decision by the U.S. Court of Appeals for the Ninth Circuit permitting plaintiffs to plead loss causation by alleging that the price of the security on the date of the purchase was artificially inflated because of the misrepresentation. The Supreme Court rejected this view, holding that an inflated purchase price does not, in itself, constitute the relevant economic loss. Id. at 345-46. The Court found the complaint to be legally insufficient because it failed to allege that the stock price fell significantly when the truth became known. Id. at 348.

After the Court's ruling in Dura, a dispute arose over the question of exactly when plaintiffs must come forward with evidence of loss causation. Some decisions have required plaintiffs to show evidence of loss causation as early as the class certification stage, in spite of the fact that such a requirement often results in inciting a "battle of the experts" over factual issues concerning the underlying merits of the litigation, which are traditionally considered no earlier than the summary judgment stage.

For example, in Oscar Private Equity Investments v. Allegiance Telecom, Inc., 487 F.3d 261, 269 (5th Cir. 2007), the U.S. Court of Appeals for the Fifth Circuit held that at least in cases involving simultaneous disclosures of negative information, not all of which are alleged to be related to the fraud, plaintiffs must establish loss causation by a preponderance of the evidence in order to trigger the fraud-onthe market presumption for class certification. The dissent in Oscar called the majority's decision "a breathtaking revision of securities class action procedure" in its departure from the rule that securities class action plaintiffs are entitled to a rebuttable presumption of transaction causation or reliance, so long as the plaintiffs traded in the stock at issue during the proposed class period in reliance on the integrity of the price set by an open and efficient market. Id. at 272.

Indeed, the Second Circuit in In re Salomon Analyst Metromedia Litig., 544 F.3d 474 (2d.Cir 2008) reiterated that plaintiffs in a class action securities fraud case are entitled to a presumption of reliance if they can show that a defendant publicly made a material misrepresentation about stock traded in an efficient market. However, the United States Court of Appeals for the Second Circuit concluded that, in cases brought against research analysts, defendants may attempt to rebut that presumption at the class certification stage. Id. at 486. Though Salomon cited Oscar as authority for that proposition, it nowhere held that the issue of loss causation is relevant to such a rebuttal. Id. at 485.

Cases like Oscar represent a sharp departure from the traditional view that the question of loss causation should not be determined before the summary judgment stage. In In re LDK Solar Securities Litigation, 255 F.R.D. 519 (N.D. Cal. 2009), the court observed that Oscar's loss-causation requirement for class-certification "placed the Fifth Circuit in a minority-indeed, apparently solitary-stance among the circuits; it is in no small amount of tension with the Supreme Court's decision in Basic v. Levinson; and although the Ninth Circuit has yet to address the issue specifically in the context of class certification, this circuit's precedent strongly suggests it would reject such a rule." Id. at 530.

Emergence of Event Study

The increasing attention to evidence of loss causation, whether presented at the class certification stage or at trial, has heightened the importance of techniques used to prove a causal connection between the disclosure of fraud and stock price movements. The challenge confronting plaintiffs in this area is that on any given day when fraud is alleged to have been revealed to the public, multiple factors unrelated to fraud, including market-wide price trends and geopolitical events, may have also affected share prices. Controlling for these non-fraud related explanations for stock price movement has become especially important for fraud claims arising from losses sustained during recent market downturns.

In order to ascertain what proportion of share price declines are attributable to the disclosure of fraud alone, economists use a statistical regression analysis to control for price movement caused by events other than the disclosure of fraud. This form of analysis, called an "event study," is a commonplace feature of securities fraud cases. As applied in securities fraud cases, an event study is, in essence, an empirical technique that measures the impact of new information-an "event"-on the price of a company's publicly traded stock. Plaintiffs' lawyers use event studies to determine whether the revelation of certain information about a company, such as the disclosure of alleged fraud, results in "statistically significant" changes in share prices.

Courts have recognized the key role that event studies have come to play in showing loss causation, and at least one court has observed that an event study is now virtually indispensable for that purpose. In In re Vivendi Universal, S.A. Securities Litigation, No. 02 Civ. 5571, 2009 WL 1066254, 10 (S.D.N.Y. April 21, 2009), the court described such studies as "almost obligatory."


In recent years, defendants in some securities fraud suits have sought to put event studies to a different use-using this statistical technique to argue that plaintiffs pursuing fraud claims cannot establish loss causation because they cannot prove that the making of a false statement resulted in a responsive inflation of stock prices. For example, in In re Marsh & McLennan Co. Securities Litigation, 04-cv-08144 (S.D.N.Y. April 19, 2006), in opposing a motion for class certification, defendants presented an event study that purported to show no statistically significant price increase immediately following 19 of the 20 misrepresentations alleged in the case.Defendants argued that the lack of responsive upward price movement upon the making of the challenged statements suggested "that the alleged misrepresentations were not material to investors and did not artificially inflate [the company's] stock price."

The argument that plaintiffs must demonstrate artificial inflation in response to the making of misstatements is inconsistent with Dura, which held that, in most securities fraud cases, loss causation cannot be established by showing price inflation upon the issuance of false statements, and focused instead on whether the disclosure of fraud results in statistically significant price decreases.

Using event studies to measure price reactions to misrepresentations when made also runs afoul of limitations inherent in the event study methodology. Event studies are not able to measure share price inflation resulting from some of the most common forms of false statements litigated in securities fraud cases, including those that have the effect of maintaining existing investor expectations of company performance.

In an efficient market, share prices of publicly traded corporations reflect investor expectations of performance based on publicly available information. If a misrepresentation or omission merely confirms market expectations, there will be no reactionary price impact. Frank Torchio, "Proper Event Study Analysis in Securities Litigation," presented at the 15th Annual Institute for Law and Economic Policy Conference in Scottsdale, Ariz. (April 24, 2009). If, for example, a company misleads investors by failing to reveal new, material, negative information, no share price reaction would result because the fraud served only to maintain existing market expectations about performance.

Event studies are equally incapable of determining whether there is stock price reaction to false information that has already been incorporated into share prices and thus cannot be used to measure market impact when misstatements are reiterated. When a material misstatement is disseminated into an efficient market, the news will be absorbed and reflected in the share price. Subsequent repetitions of the same misstatement that simply maintain existing investor expectations of company performance should not cause a change in share prices.

In Alaska Electrical Pension Fund v. Pharmacia Corporation, 554 F.3d 342 (3d Cir.2009), the United States Court of Appeals for the Third Circuit acknowledged that a misstatement that merely confirms market expectations will not cause a price impact. Accordingly, the Alaska Electrical Pension Fund court pointed out that a lack of price movement on the announcement of misleading information did not negate a finding of materiality of those statements. In fact, the court said that the materiality of the misstatements was self-evident in light of the nine-point stock price drop when the truth emerged. Id. at 352.

The court in In re Scientific-Atlanta, Inc.Securities Litigation, 571 F. Supp. 2d 1315 (N.D. Ga. 2007) came to the same conclusion in rejecting defendants' argument that the lack of any statistically significant positive price movement after the alleged misrepresentations foreclosed any link between the misrepresentations and plaintiffs' losses. The court held that "[c] ontrary to defendants' argument, the mere absence of a statistically significant increase in the share price in response to fraudulent information does not 'sever the link' between the material misstatements and the price of the stock. Rather, price stability may just as likely demonstrate the market consequence of fraud where the alleged fraudulent statement conveys that the company has met market expectations, when in fact it has not." Id. at 1340-41.


These recent developments highlight the importance of clarity and precision in both our understanding of the economic underpinnings of loss causation and the limits of the legal requirements imposed by Dura. While event studies are an indispensable tool for ensuring that investor losses are truly the result of fraud, plaintiffs and defendants alike must be mindful of the judicial, statistical and logical limitations on the use of these valuable statistical analyses.