How Can You Tell If A CEO Is Lying? – Kyle Stock, DealJournal, The Wall Street Journal
Conference call Q&As are a confusing and cryptic dance. Executives try to be attractive to investors, without giving away too much. In many cases, they are trying to put a good spin on bad results.
But what if an investor could read right through all of the posturing and careful prose to know if they were being strung along?
A pair of professors at Stanford recently tried to do just that. The team built a model that tries to flush out executive lies, using psychological and linguistic studies and transcripts of conference calls from companies that later restated earnings.
They fed their filter almost 30,000 earnings transcripts from 2003 to 2007 and found that it worked quite nicely. Executives who later had to revise their books displayed some very consistent clues.
For one, they seldom referred to themselves or their firms in the first person; “I” and “we” were replaced by terms like “the team” and “the company.” Deceitful executives passed up humdrum adjectives like “solid” and “respectable” in favor of gushing words like “fantastic,” and (not surprisingly) they seldom mentioned shareholder value.
They also tended buttress their points with references to general knowledge with phrases like “you know” and to make short statements with little hesitatation, presumably because they had carefully scripted the untruths in advance and had no interest in lingering on them.
Though the study doesn’t call out particular companies, chiefs across a wide-range of industries raised the censor’s red-light 14% of the time. Those in the finance business proved slightly more honest than average, tagged for lying only 10% of the time. The study didn’t specify the industry with the most dissembling.
CFOs, it appears, hold their cards a little closer to their chests. They spoke about half as much as their bosses, and, unlike CEOs, they showed no “positive emotions” via “brilliant” and “astounding” adjectives. Maybe they were busy picturing themselves in brilliant orange coveralls.
The model is not perfect. It proved accurate enough to make predictions between 50% and 65% of the time, in part because individual executives have unique ways of speaking that don’t fall neatly into a pattern of deception.
Still, big money has to like those odds. We bet that the authors of the study, David Larcker and Anastasia Zakolyukina, will be hearing from some hedge funds soon, if they haven’t already.