Every so often I read something and go ‘Wow!’ Reviewing the paper highlighted this week was one of those moments.
Eping Liu and Haoyuan Qin, both from the Business Schools of the Sun Yat Sen University in Guangzhou and Shenzhen, have done extraordinary work, they claim is the first of its kind, on how corporate manager’s facial expression on earnings calls is a reliable ‘tell’ on future stock price performance.
Using cognitive dissonance theory to form their working hypothesis they employ a deep learning model that worked, frame by frame, through 297-video calls from 2016 to 2022 to capture corporate manager’s emotions.
From there they followed the stock price performance of the respective manager’s companies to see if positive facial language was a reliable harbinger of good news ahead, and it was.
The work is flawed in some obvious respects. The sample size is small, the time period may not be representative, there’s no analysis of whether or not sourpuss-looking managers flag bad results, facial expression analysis is fundamentally subjective and etcetera.
However, this is an exciting development and the logical follow-on of work that’s been done for some time machine-analyzing written and verbal corporate communication for similar performance tells.
You can read the paper in full via the following link Manager’s Facial Expressions as a Predictor of Future Returns, and I’m still at wow!
Happy Sunday