L'impact du Say on Pay: étude récente

Transparence oblige, j'avoue être quelque peu sceptique face au mouvement de Say on Pay. D'un point de vue théorique, je me demande si ce mécanisme n'est pas susceptible simplement de déplacer le problème en donnant un droit de regard aux actionnnaires sur la rémunération. Si le concept est de prime abord séduisant, n'y-a-t-il pas un risque d'opportunisme de leur part? De plus, est-ce que ce transfert peut avoir pour effet de diminuer (étrangement) l'imputabilité des administrateurs? D'un point de vue empirique, au moins une étude de RiskMetrics avait démontré que dans les pays où ce mécanisme avait été mis en place, la rémunération des dirigeants avaient... continué à augmenté.

Une étude empirique récente, Say on Pay Votes and CEO Compensation, vient fournir un appui aux tenants du Say on Pay. Voici le résumé de l'étude, actuellement sur le blogue du Harvard Law School Program on Corporate Governance:

... we perform two sets of tests. First, we examine UK firms’ responses to say on pay votes by analyzing the changes to compensation policies made by firms after facing high voting dissent against their remuneration report. We document that a significant number of these firms removed or modified provisions that investors viewed as “rewards for failure” (e.g., generous severance contracts, low performance hurdles, provisions allowing the retesting of performance conditions)—often in response to institutional investors’ explicit requests—and established a formal process for proactive consultation with their major shareholders going forward. These actions paid off, in that voting dissent at the subsequent annual meeting was substantially lower. We also find evidence of similar actions taken before the vote by a subset of firms that subsequently experienced low voting dissent, suggesting that the threat of a vote induced some firms to revise CEO pay practices ahead of the annual meeting.

Second, we examine the trend in CEO pay and its sensitivity to economic determinants before and after the introduction of say on pay for a large sample of UK firms. We find no evidence of a change in the level and growth rate of CEO pay—after controlling for firm performance, size and other factors. However, we find a significant increase in the sensitivity of CEO pay to poor performance. The increase is most pronounced in (i) firms with high voting dissent, and (ii) firms with an ”excessive” level of CEO pay (relative to the level predicted by its economic determinants) before the adoption of say on pay, regardless of the voting dissent. Interestingly, we do not find a more pronounced increase in firms with higher raw levels of CEO pay. These findings confirm the insights from our small-sample evidence of explicit changes to pay contracts and suggest the following: (i) UK investors used say on pay to push for greater accountability for poor performance; (ii) firms responded to adverse shareholder votes, in spite of their non-binding nature; (iii) (at least some) firms responded to the threat rather than the realization of an adverse vote; (iv) shareholders focused on firms with controversial CEO pay packages (as captured by high voting dissent or excessive CEO pay levels) rather than firms with high (but not abnormal) CEO pay levels.

Il y aurait donc une effet disciplinaire au Royaume-Uni découlant de ce mécanisme. Intéressant et à suivre.