Do Compensation Consultants Drive Up CEO Pay? Evidence from UK Public Firms

Martin J. Conyon, Lars, Helge Hass, Simon, I Peck, Graham Sadler, Zhifang Zhang

Research output: Contribution to journalArticlepeer-review

12 Citations (Scopus)
19 Downloads (Pure)


Do compensation consultants drive up CEO pay for the benefit of managers, or do they design pay packages to benefit firm owners? Using a large sample of UK firms from the FTSE All-Share Index over the 2003–2011 period, we show a positive correlation between the presence of compensation consultants and CEO pay. Importantly, isolating this effect is somewhat dependent on the endogenous selection of consultants and the statistical modelling strategy deployed. We find evidence that compensation consultants improve CEO compensation design when their expertise is of greater importance (e.g. during the post-financial crisis period, or for firms that have particularly weak compensation policies). In addition, our findings show that compensation consultants increase CEO pay–performance sensitivity. The balance of evidence supports optimal contracting theory more than managerial power theory, but the authors caution the limits to this verification. We are careful to note that the more compelling evidence for the positive effect of pay consultants on CEOs is based on advanced methods (such as propensity score matching and difference-in-differences), and that more standard approaches (such as OLS and fixed effects) are unlikely to reveal the same level of causality of consultants on CEO pay.

Original languageEnglish
Pages (from-to)10-29
Number of pages20
JournalBritish Journal of Management
Issue number1
Early online date30 Jan 2019
Publication statusPublished - 2019

ASJC Scopus subject areas

  • Business, Management and Accounting(all)
  • Strategy and Management
  • Management of Technology and Innovation


Dive into the research topics of 'Do Compensation Consultants Drive Up CEO Pay? Evidence from UK Public Firms'. Together they form a unique fingerprint.

Cite this