An Empirical Test of the Effect of Outrage Costs on CEO Compensation Level

Abstract

Managerial power theory predicts that increased disclosure about CEO compensation will limit or even decrease CEO compensation through increased outrage costs. Empirical testing of this key conjecture by using data from a mandatory disclosure setting is difficult as any change in CEO compensation after mandatory disclosure can be attributed to the fact that CEO compensation did not reflect the underlying economic determinants. By using data from a voluntary disclosure environment, this study develops a cleaner test of the effect of disclosure on CEO compensation level. In line with managerial power theory, we expect a decrease in compensation levels for CEOs who are in the upper part of the compensation distribution. For CEOs who are in the lower part of the compensation distribution, we expect that increased disclosure will instigate social comparison processes, leading to an increase in compensation levels. To test our theory, we use data from Belgian listed firms as the introduction of a corporate governance code in 2005 can be considered as a shock with respect to disclosure of CEO compensation for Belgian listed firms. Consistent with our theory, we find a dual effect of disclosing CEO compensation in which the influence of disclosure on CEO compensation is driven by the difference between the compensation of the CEO and the average (or median) compensation of the reference group. First, disclosure has the intended effect, since CEOs who are in the upper part of the compensation distribution are faced with a decrease in compensation. Second, disclosure of CEO compensation also leads to an unintended effect since compensation levels of CEOs who are below the average (or median) of the peer group increase through increased social comparison. This paper thus provides a more nuanced view on the claim of managerial power theory that ‘sunlight is the best disinfectant’.