The Face of Trust: CEO Appearance and Market Reactions To Financial Restatements

Abstract

The thesis explores whether the facial trustworthiness of a chief executive officer is predictive of abnormal stock returns following announcements of financial restatements. Using an event-study framework, mean-adjusted, value-weighted and risk-adjusted CARs were estimated of 150 firm restatements, using a sequence of series of event windows. Scores of trustworthiness were obtained based on various artificial-intelligence algorithms. To measure the influence of CEO trust scores, cross-sectional regressions were adopted, firm size and the severity of the restatement were controlled. Interaction models were checked on whether appearance effects are moderated by the size of the firm or by the materiality of the event. The findings indicate the absence of strong unconditional correlation between the facial trustworthiness of the CEO and the CARs, and the firm size was positively correlated with returns within the short horizon window. The diagnostics of robustness such as influence analysis, winsorisation, and multicollinearity checks showed that the conclusions were not based on outliers and model misspecification. The paper thus finds that market responses to financial restatements are dominated by the informational environment and the severity of the event, and that the appearance characteristics of CEOs plays at most a small, context-dependent role.