Sunday, August 1, 2010

Is That CEO Telling the Truth?

How do you tell if CEOs are not being truthful during quarterly earnings conference calls? Stanford Graduate School of Business researchers have developed a model to analyze the words and phrases used during these calls and found some specific speech patterns that give clues.

STANFORD GRADUATE SCHOOL OF BUSINESS—How do you tell if CEOs are not being truthful during quarterly earnings conference calls? Stanford Graduate School of Business researchers have developed a model to analyze the words and phrases used during these calls and found some specific speech patterns that give clues.

After studying Q&A sections of transcripts of hundreds of calls with CEOs and CFOs, the researchers then looked to see whether financial statements being discussed were substantially restated at some point after the call. If they were restated, Professor David Larcker and Anastasia Zakolyukina, a PhD student at the school, reasoned that the executive had been less than candid in describing their firm’s quarterly figures.

Larcker, the James Irvin Miller Professor of Accounting and senior faculty of the Stanford Rock Center for Corporate Governance, and Zakolyukina developed a model to analyze words and phrases used based on prior deception detection research conducted by psychologists and linguists. CEOs who were hiding information were less likely to say "I" and more likely to use impersonal pronouns and references to general knowledge such as "you know." They also expressed more extreme positive emotions ("fantastic" as opposed to "good"), used fewer extreme negative emotions, and fewer certainty and hesitation words. They were less likely to refer to shareholders value.

Results from their model are 4% to 6% better than a random guess. Yet the authors offered some cautions about their work. "First, we are not completely certain that the CEO and/or CFO know about the manipulation when they answer questions during the conference call," they wrote. They also cautioned the words they studied might not be "completely appropriate for capturing business communication." But they said, the results were strong enough to warrant additional research.