The story is probably apocryphal, but it is said that when Richard Nixon debated John F. Kennedy for the presidential election 1960, the people listening on the radio preferred Nixon while the people watching on TV preferred Kennedy. Facial expressions can undermine a message, though, as analysis of CEO interviews on CNBC shows.
Elizabeth Blankespoor and her collaborators from the University of Hong Kong used facial analysis software to classify the emotions displayed by CEOs in 508 CNBC interviews between 2013 and 2017. The facial analysis software measured the emotional expression of faces during the interview and classified them into seven different basic expressions. We don’t have to bother too much about those seven types of expressions because when it comes to CEO interviews on business TV, the dominant expression was ‘neutral’ with either a tilt toward ‘happy’ or ‘sad’. The rest was mostly noise. Makes sense since business TV is not designed to be ‘Downtown Abbey’ or ‘The Sopranos’ (though sometimes CEOs try their best to emulate Tony Soprano).
Facial analysis of CEO interview
Source: Blankespoor et al. (2023)
But there should be some emotion in these TV interviews because most of them are recorded in response to a quarterly or annual results report by a company. If a company just beat expectations, one would expect the CEO to be happy and upbeat in the interview while sadness would be expected after earnings miss.
But what happens if there is a mismatch between the results presented and the facial expression of the CEO in the interview? What if the CEO appears happy after an earnings miss or sad after a beat?
This dissonance influences analyst forecasts. If the CEO shows emotions that are at odds with the content of her message, the dispersion of analyst forecasts increases. The lack of consistency in the messaging confuses analysts on a subconscious level and dilutes the message.
The effect isn’t large. It is about one third in size compared to the effect of the earnings surprise, but it is still there. And it gives you an idea of why CEOs of large corporations are routinely trained by PR professionals in interview techniques and how to get their message across. But even with this PR training in place, I am still surprised by how bad many CEOs are in communicating their message. Maybe they are just not good at their job? That is something I will address tomorrow…
Somehow, this brought to my mind Klaus Kleinfeld, ex-CEO of Siemens and Alcoa, maestro at shooting himself in the foot communication-wise. He'd smile while trying to blackmail his own investors. Indeed, what works for Tony Soprano doesn't (usually) work for CEO's.
The below is from WSJ. Do you think playing a role in this analysis either by us or AI?
Similar findings hold true for every so-called universal facial expression of emotion. Frowning in sadness, smiling in happiness, widening your eyes in fear, wrinkling your nose in disgust and yes, scowling in anger, are stereotypes—common but oversimplified notions about emotional expressions.
In short, we can’t train AI on stereotypes and expect the results to work in real life, no matter how big the data set or sophisticated the algorithm. Shortly after the paper was published, Microsoft retired the emotion AI features of their facial recognition software.
https://www.wsj.com/tech/ai/think-ai-can-perceive-emotion-think-again-2b4c7d29?st=7d4ik212hwmhpu3&reflink=article_copyURL_share