Brian Bell, University of Oxford and John Van Reenen, MIT Sloan School of Management report on their findings about CEO pay related to performance and to level of institutional ownership:
CEO pay is not symmetric: it responds more to increases in firm performance than to decreases. To explore whether this is because of corporate governance issues, we split the firms into two groups depending on whether a large share of the equity of the firm is held by institutional investors—since such shareholders are more likely to exert strong external control on the firm’s executives. We show that the asymmetry in pay is only observed when there is weak external control (see Figure 2 Panel B).
Strong governance is measured by a high share of Institutional Owners and weak governance by a low share. Panel A shows that in well governed firms the pay-performance relationship is symmetric whereas in badly governed firms, CEO pay does not fall, even when the firm does badly.Second, we found substantial “pay for luck”, i.e. CEO pay increased when the industry experiences a random positive shock. Surprisingly, this was still the case even when the CEO was subject to plans rewarding them for performance relative to the industry. One reason for this was because CEOs were able to rig the system. When the share price underperformed, the top boss managed to extract some particularly juicy new LTIPs to make up for his expected loss of performance pay. This undermining of the system was only in firms who were weakly governed, consistent with the findings on pay asymmetry.
Notes: Theses figures represent the implied effect of a percentage change in a firm’s performance (as measured by TSR, Total Shareholder Returns) on the percentage increase in CEO pay. The coefficients are from the specification in column (5) of Table 4 of Bell and Van Reenen (2016). 95% confidence intervals shown.