As we begin to analyze the first proxy season under “say on pay” in the US, it may be useful to review the evidence from the UK experience with say on pay. In the study, Say on Pay Votes and CEO Compensation: Evidence from the UK, co-authored with David Maber of University of Southern California and forthcoming in the Review of Finance, we examine the impact of “say on pay” in the UK, the first country to adopt a mandatory, non-binding annual shareholder vote on executive pay.
We perform three sets of analyses. First, we examine the market reaction to the (largely unanticipated) announcement of say on pay regulation and find positive abnormal returns for firms with weak penalties for poor performance, e.g. firms with excess CEO pay combined with poor performance and firms with generous severance contracts, which can weaken penalties in the event of poor future performance.
Second, we compare changes to compensation contracts made by high dissent firms (i.e., firms that experienced more than 20% voting dissent) before and after the first say on pay vote, with changes made by a matched sample of low dissent firms (i.e., firms that experienced less than 5% voting dissent but with otherwise similar characteristics). Our analyses indicate that, after the vote, high dissent firms were more likely to remove provisions viewed as “rewards for failure” (e.g., generous severance contracts, provisions allowing the retesting of performance conditions), often in response to institutional investors’ explicit requests, and reaped the benefits in terms of lower or no dissent at the following year’s annual meeting. Low dissent firms were more likely to remove such provisions before the vote, presumably in a (successful) effort to avoid voting dissent. These different patterns of behavior around the vote for high and low dissent firms suggest that the observed changes were the direct result of say on pay.
Our third and final set of tests employs regression analysis to evaluate the sensitivity of CEO pay to realized performance and other economic determinants before and after say on pay regulation. Using a large sample of UK firms, we find a significant increase in the sensitivity of CEO pay to poor performance (particularly in high dissent firms and firms characterized as having excess CEO pay before the adoption of say on pay), while the relationship between pay and other economic determinants remains unchanged. It thus appears that say on pay has a moderating effect on the level of CEO compensation only conditional upon poor performance. We do not find a similar effect in a control sample of UK firms not subject to say on pay, UK firms traded on the Alternative Investment Market. Taken together with our evidence of explicit changes to compensation practices, these additional tests support a causal interpretation of our findings.
Overall, our study suggests that UK investors perceived say on pay to be a value enhancing monitoring mechanism, and were successful in using say on pay votes to pressure firms to remove controversial pay practices and increase the sensitivity of pay to poor performance. Based on these findings, we expect that say on pay in the US will be similarly used to strengthen the link between pay and performance at few firms with controversial pay practices rather than to opine on pay levels across the board. Investors will try to impose certain rules of the game (e.g. use of performance conditions in equity grants) and single out questionable provisions (tax gross-ups) rather than micromanage compensation packages. Fears that say on pay will lead to massive chaos at annual meetings and that special interest groups will hijack shareholder votes will not materialize. Early evidence from the 2011 proxy season is consistent with these predictions. Perhaps the most interesting question is whether say on pay votes will lead to a level of communication between board and institutional investors similar to the one in the UK. The net effect of say on pay in the US may very well depend on this.
The full paper is available for download here.