A company’s “fantastic earnings” may not be so great after all, according to a recent study co-authored by David Larcker, finance professor at the Graduate School of Business (GSB) and M.B.A. student Anastasia Zakolyukina. The study analyzes CEO conference calls to shed light on the linguistic methods used by deceptive corporate managers.

“It’s an important thing for corporate governance to understand whether people are dealing in a truthful matter or not,” Larcker said. “When someone releases their accounting reports, you’d like to know whether you actually believe those numbers, or if they have been manipulated in some form.”

According to the study, deceptive executives tend to refer to general knowledge while speaking, using blanket phrases such as “as you know” to boost credibility with listeners. Deceivers also use significantly fewer self-references while speaking, substituting third person pronouns for the word “I” to avoid taking ownership of false statements.

Zakolyukina was surprised to find that deceiving chief executives and chief financial officers (CFO) actually use extremely positive language—“fantastic” or “great”—to describe earnings and avoid revealing mediocre results.

For example, in a March 2008 earnings conference call, former Lehman Brothers CFO Erin Callan used phrases like “very strong client flows” and “a robust trading environment.” Such examples of extreme positive language allow deceptive speakers to sound more persuasive while concealing the truth.

Furthermore, the use of swear words, usually indicative of negative emotions, only occurs minimally and is not statistically significant.

“In our results, we only found swear words in just one case, in a very specific setting,” Zakolyukina said. “But for some reason, perhaps because of the famous Enron case, the idea of swear words [as an indicator of deception] was developed and propagated.”

To verify that CEOs were being truthful during a conference call, the study analyzes subsequent changes to stated earnings from previous quarters. A correction signified that the call was deceptive.

“We have to assume that if the company makes a restatement, the executive at the time of the conference call knew about it and his language will exhibit deception,” Zakolyukina said.

Unlike most research efforts to predict company manipulation, Zakolyukina and Larcker chose to analyze the question-and-answer section of earnings conference calls for their spontaneous nature, which allows less room for language manipulation.

“Some papers try to look at the language used in financial reports,” Zakolyukina said. “But all the formal reporting is screened through the legal department of the company, so they are well-prepared pieces of text.”

Additionally, conference calls have the added benefit of a quarterly schedule, which added regularity and specificity to Larcker and Zakolyukina’s long-term study of deceptive language.

“We’ve seen that no matter how the reports were manipulated, if the manager knows about it at the time of the conference calls, the more likely the way he speaks will give out that he’s hiding something,” Zakolyukina said.

M.B.A. student Aaron Gelband MBA, who is co-president of the GSB’s Finance and Investment Club, has investments in Perry Capital and previously worked for Goldman Sachs, said that while the study provided insight into the methods of deception in corporate management, the findings do not always indicate dishonesty.

“I will caution that although the results are statistically significant, only 60 percent of instances in management statements would suggest that they were actually being dishonest,” Gelband added. “Even if they displayed signs that are typically indicative of dishonesty, you still have to treat that with a grain of salt.”

Despite its relevance in language research, Gelband said the study will not have any substantial effects on his post-graduation plans because of its limited professional and practical implications.

“I don’t think it’s going to substantially change the way I do anything after business school,” Gelband continued. “But it will make me a little more thoughtful and aware about the type of language that management uses.”

In fact, Larcker cautioned that the study only provides a preliminary view of corporate deception.

“We found some results, but there’s plenty of work to do,” Larcker said. “The idea of using language to understand behavior is a very active research area and it requires skills across various areas.”

But he is optimistic that despite the difficulties of analyzing an interdisciplinary subject, added research may eventually apply these findings to situations outside of a corporate setting.

Since deceptive-language research is new, “it’s going to be risky as well,” Larcker said. “But presumably, that’s what we’re supposed to do with our research.”