Shareholder Votes and Proxy Advisors: Evidence from Say on Pay
University of Colorado at Boulder - Department of Accounting
Columbia Business School - Accounting, Business Law & Taxation
University of St. Gallen - Swiss Institute of Banking and Finance
August 2, 2013
Journal of Accounting Research, Forthcoming
We investigate the economic role of proxy advisors (PA) in the context of mandatory “say on pay” votes, a novel and complex item requiring significant firm-specific analysis. PA are more likely to issue an Against recommendation at firms with poor performance and higher levels of CEO pay and do not appear to follow a “one-size-fits-all” approach. PA recommendations are the key determinant of voting outcome but the sensitivity of shareholder votes to these recommendations varies with the institutional ownership structure, and the rationale behind the recommendation, suggesting that at least some shareholders do not blindly follow these recommendations. More than half of the firms respond to the adverse shareholder vote triggered by a negative recommendation by engaging with investors and making changes to their compensation plan. However, we find no market reaction to the announcement of such changes, even when material enough to result in a favorable recommendation and vote the following year. Our findings suggest that, rather than identifying and promoting superior compensation practices, PA’s key economic role is processing a substantial amount of executive pay information on behalf of institutional investors, hence reducing their cost of making informed voting decisions. Our findings contribute to the literature on shareholder voting and the related policy debate.
Number of Pages in PDF File: 68
Keywords: Say on pay, proxy advisors, shareholder votes, CEO compensation, shareholder activism
JEL Classification: G34, G38, J33, M12working papers series
Date posted: March 11, 2012 ; Last revised: November 16, 2013
© 2013 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo5 in 0.391 seconds