Sam Bankman-Fried’s trial on federal fraud charges began this month. The founder of the failed crypto exchange FTX is just one recent example of high-profile executives who deceived analysts — a practice that a new study published in Strategic Management Journal may help deter, saving billions of dollars per year in corporate fraud.
“One aspect of their fraud was their buying all of these other exchanges, with money they didn’t have,” says Steven J. Hyde of Boise State University, lead researcher of the study. “One analyst pointed to their acquisitions as being evidence that this company was really successful. But it was actually evidence of their fraud.”
Hyde and his team — Eric Bachura of University of Texas at San Antonio, Jonathan Bundy of Arizona State University, Richard T. Gretz of University of Texas at San Antonio, and Wm. Gerard Sanders of the University of Nevada, Las Vegas — used the machine learning platform they developed to find that the language Bankman-Fried used with investors was often deceitful.
The researchers were curious as to how leaders at FTX, Wells-Fargo, Theranos, and other organizations were able to lie without getting caught. They designed a study that asked two core questions: Do analysts detect when CEOs lie? And in what context are they more or less likely to pick it up? To answer their questions, they looked at linguistic patterns that are found to be indicative of lying among CEOs (for example, distancing language tends to be associated with lying).
Previous studies of the topic used regression analyses, with an accuracy rate of about 65%. But Hyde’s team developed a machine learning component in which they identified instances where CEOs committed very serious fraud and created a sample of CEOs like them; their model was 85% accurate. The researchers connected the findings from their model with how analysts reacted.
The study came away with three major findings:
- Analysts by and large reward CEOs for deception.
- There is, however, a “boy who cried wolf effect,” in that analysts will catch on over time if they are repeatedly lied to.
- Lastly, they found that analysts with the highest reputations are the ones who are the slowest to pick up on deception.
“All of these effects are being driven by our general human tendency to assume that people are being honest with us,” Hyde says. “(Star analysts’) prestige exaggerates the bias. They assume even more that people won’t be lying to them; there’s a level of ego that comes in.”
The takeaways for investors? Consider looking to lesser-known analysts if you’re concerned that a company’s CEO is being dishonest. They’re likely to pick up on those cues faster than a more all-star analyst. Hyde says he does expect more analysts to incorporate linguistic analysis to improve the accuracy of their reports.
There are the ethical implications: Hyde cautions that the machine learning models are not perfect, and there could be false negatives or positives (they had a harder time catching lies from Theranos CEO Elizabeth Holmes, for example). Leaders could also learn to run their speeches through these platforms and change their language to better hide their deceit.
“Yes, we’re measuring deception, but really what we’re measuring is how similar the linguistic pattern of this CEO is to someone who’s committing deception,” Hyde says. “You still need a human component here. You still need to investigate what’s actually been said, and not just take that algorithm at face value.”
Find a full explanation of the study in the full text, available in the Strategic Management Journal.