Facial Trustworthiness and CEO Turnover

Abstract

This paper examines the effect of facial trustworthiness on the turnover of Chief Executive Officers (CEOs) and its moderating role in the relationship between firm performance and CEO turnover. We employ a unique dataset comprising headshots of male CEOs in US listed firms from 2010 to 2019 and develop a machine learning-based facial landmark detector to construct a composite facial trustworthiness index. Our results show that CEO’s facial trustworthiness is negatively associated with executive turnover. We also find that facial trustworthiness strongly influences the link between firm performance and CEO turnover. Well-performing executives have a lower risk of being dismissed from their positions and the dismissal risk is even lower for highly trusted CEOs. Conversely, underperforming CEOs are more likely to be dismissed and the dismissal probability is greater for highly trusted CEOs.

Keywords: facial trustworthiness, CEO turnover, firm performance, machine learning, facial landmark detector