Discussion of Analyzing Speech to Detect Financial Misreporting

Abstract

One of the most important tasks facing investors, auditors, and regulators is to identify misreporting by managers, preferably using ex ante signals. Hobson, Mayew, and Venkatachalam [2012] (henceforth, HMV) examine the use of CEOs’ voice markers of cognitive dissonance for detecting financial misreporting. Prior research on financial misreporting detection has focused mostly on financial measures, and, more recently, nonfinancial performance measures. HMV innovate by bringing a new perspective into this field. They motivate this research question with the findings in the psychology literature that nonverbal vocal markers can help discriminate deceivers from truth tellers. They focus on a specific type of vocal marker in the paper—cognitive dissonance—based on prior evidence (e.g., Mazar, Amir, and Ariely [2008]) that misreporting leads to cognitive dissonance, which is the feeling of psychological discomfort when one’s actions and beliefs are discrepant. To measure managers’ cognitive dissonance, HMV rely on a proprietary software package, Layered Voice Analysis (LVA). The first step in the empirical tests is to validate the measure as capturing cognitive dissonance. They