Print
E-MailAllen Ferrell is the Harvey Greenfield Professor of Securities Law at Harvard Law School.
My co-author, Atanu Saha, and I have recently posted three papers dealing with securities damage issues. The first paper, Forward-Casting 10b-5 Damages: A Comparison to Other Methods, discusses and critiques two commonly used methods for calculating securities fraud damages under Rule 10b-5: constant dollar back-casting and the allocation method. We also present the forward-casting method, our proposed alternative to these other methods. Not only is the forward-casting method well-grounded in academic literature, but has the advantage of incorporating market expectations when determining what the stock price would have been “but for” the alleged fraud. Neither the constant dollar back-casting nor the allocation method takes these market expectations into account. We also address other issues that can arise by virtue of using these various methods. These three methods can generate substantially different 10b-5 damage estimates. We use a real world example to demonstrate these differences in the three methodologies.
The methodology used to estimate Rule 10b-5 damages is extremely important when estimating potential liability exposure and the determination of settlement value. In addition, these estimates, because they speak to the extent to which securities prices were distorted by fraudulent misinformation, can also be useful in determining whether the misinformation was “material,” an element of a Rule 10b-5 cause of action. Moreover, our analysis has implications for other causes of action, including Sections 11 and 12(2) of the Securities Act of 1933 and common law fraud actions.
ERISA class actions constitute an important component of overall U.S. securities litigation activity, and have become prominent after Enron in 2001 and the more recent credit crisis. Unlike a Rule 10b-5 action, plaintiffs don’t need to prove that defendants acted with intent or recklessly, but only need to show a breach of a fiduciary obligation. Our second paper, Calculating Damages in ERISA Litigation, focuses on comparing and contrasting four different methods for calculating ERISA damages (which we label the “best performing fund” “portfolio redistribution,” “most similar fund” and “10b-5 style” ERISA damage methods), using the facts and data from an actual ERISA matter.
As is the case when estimating Rule 10b-5 damages, the chosen methodology can have a large effect on the final estimate in ERISA actions. In our real world example, aggregate damages ranged from less than $3 million using the “most similar” fund approach, to well over $2 billion using the “best performing fund” method. Our paper also examines several common plaintiffs’ theories for ERISA liability, and discusses how the appropriate damage method will be informed by plaintiffs’ argued basis for liability. In cases where liability is based on an inadequate disclosure claim, we argue that “10b-5 style” damages can be the most appropriate method. In other cases, the most appropriate damage method can be the “most similar fund” approach.
In securities litigation dealing with misinformation or inadequate disclosure, event study analysis is often used to examine market price reaction following the correction or disclosure of this information. If market returns are abnormal at a statistically significant level, this is sometimes used to argue that the misinformation or the omitted information is material (which speaks to liability) and estimate how much the stock price was inflated (which speaks to damages). In order to assess the impact on individual shareholders, abnormal returns are often transformed into abnormal dollar price impacts.
Our third paper Event Study Analysis: Correctly Measuring the Dollar Impact of an Event addresses the common assumption that if the abnormal return on a disclosure day is statistically significant, so is the abnormal dollar effect. We demonstrate—first analytically and then through an empirical example—that need not be the case. We derive the proper t-statistic if one wishes to determine the statistical significance of an abnormal dollar effect (in other words, if a stock price drops, for example, from $5 to $4.50 the t-statistic for determining whether this 10% drop is statistically significant will differ from the t-statistic for determining whether the 50 cent drop is statistically significant). This result of the potential difference of statistical significance of the dollar impact of disclosure has obvious implications for both materiality and damages in 10b-5 securities matters.
