Benchmarking and CEO Compensation

Between 1980 and 2004, the adjusted average compensation of a CEO at a large public corporation in the U.S. rose by 12.2% per year, from $625,000 to $9,840,000. Although the rise in CEO pay has flattened in the last few years, this dramatic rise has received extensive media, policy and academic attention. So far, scholarship on the rise in CEO pay has focused on two explanations: market forces and managerial power. These explanations treat firms as independent actors, and ignore the prevailing practice of benchmarking executives' pay against groups of 'peer' organizations in order to determine compensation packages. In 2006 however, the SEC began requiring firms to publicly disclose their executive compensation peer groups, and these groups now provide an additional perspective through which to understand the dramatic rise in CEO compensation. In order to better understand executive compensation peer groups and how they impact CEO compensation, Tom DiPrete and colleagues have compiled a dataset that indexes the compensation benchmarking disclosures of approximately 4,500 companies first made public in 2006. This dataset was created using software algorithms that analyze electronic versions of the SEC proxy filings in which companies disclose their peer groups. This data is then merged with data from Morningstar on CEO compensation, board memberships and company financial information, to form a complete picture of executive compensation peer groups. DiPrete's earlier work with this data indicates that when firms select peers in their compensation groups, they systematically choose firms that are larger and have better compensated CEOs, thus justifying higher pay for their own CEO. DiPrete will build on this earlier work to explore how SEC publication of compensation peer groups will impact the way benchmark companies are chosen. It is possible that public accountability will force companies to gradually transition away from aspirational peer groups towards more obviously normative groups, lowering CEO pay. However, it is also possible that comparable companies in a given industry will increasingly name the same high compensation companies as a part of their compensation peer groups, making the selected peer groups of these comparable companies more closely resemble each other, but not lowering CEO compensation. In order to study this impact, DiPrete will use four years of data on published compensation peer groups in order to examine how these groups have changed over time.