Blathnaid Heaney, StoneTurn Senior Consultant, contributed to this article.

In the realm of securities litigation, event study analyses are a commonly used tool to assess whether a particular disclosure or news event was material to investors. These studies attempt to isolate the impact of an event on a company’s stock price by controlling for market- and industry-wide factors through regression modeling. At the heart of this approach is the use of an industry index to measure broader market trends. However, if this index is not properly constructed, the regression model may misattribute price movement to the event in question, potentially leading to flawed conclusions about materiality. In Law360, Dr. Atanu Saha, Dr. Yong Xu, and Blathnaid Heaney confront a key methodological question: how should one assess whether an industry index is appropriately constructed for use in an event study?

Traditionally, some courts and practitioners have turned to the R-squared value—the proportion of variance explained by the regression—as a measure of a model’s reliability. Drawing on examples from recent 10b-5 litigation cases and extensive empirical testing using over 500 sector ETF-based event studies, the authors challenge this practice, demonstrating that a low R² does not necessarily indicate a flawed industry index or an unreliable event study. Instead, they argue that the t-statistic associated with the industry index is a far more meaningful indicator of whether the index adequately captures industry-wide effects. In doing so, it offers both a statistical and legal rationale for shifting focus away from R² and toward more appropriate metrics when evaluating the soundness of event study models in securities litigation.

Read the full article in in Law360.

If you have any questions or would like to discuss this topic please reach out to Dr. Atanu Saha or Dr. Yong Xu.

To receive StoneTurn Insights, sign up for our newsletter.