How Use - and Misuse - of Executive Compensation Surveys Leads to Upward Escalation of CEO Compensation*

By Frederic W. Cook, Chair of Frederic W. Cook & Co.

1. Setting target pay levels at, e.g., 75^th percentile, without justification of relative size and performance, thereby creating a ratchet effect.
    

2. Biased survey sample – selecting companies that are larger, better performing, or higher paying than your company.
    

3. Using surveys for pay comparisons without taking account of relative size and/or performance.
    

4. When doing size and performance comparisons, selecting only those metrics that show the company favorably, and rejecting those that do not.
    
5. Using survey results more frequently when pay is going up than when it is going down.
    

6. Making the case, selectively, that certain of your executive positions are more responsible/important to your company than the survey benchmarks, without giving equal consideration to those of your positions that may be less responsible or important.
    

7. Setting target bonus and LTI amounts based on actual payouts at others – particularly egregious practice when the economy is improving. 
    

8. Selectively disclosing to the Compensation Committee only those surveys that support what the company wants to do.

   – Burying the others
    

9. Misuse of statistics, for example, focusing on pay averages rather than medians, and disregarding zeros when computing averages and medians.

   – For example, if five of 10 companies grant time-based restricted stock and the average of these five is $500,000, it is not right to say that the average company grants restricted stock worth $500,000
    

10. Selectively surveying individual compensation elements without considering total compensation.
     

11. Including one-time recruitment or buy-out awards in total compensation comparisons.
     

12. Implying greater prevalence to highly questionable practices to encourage adoption.

    – For example, benefits advocates are notorious for lumping ERISA Excess Plans, which only treat executives equitably, with SERPs, which treat executives better than other salaried employees, and calling them all SERPs
     

13. Including high Black-Scholes option values in surveys and then using such values to justify large grants of less risky equity compensation, such as restricted stock.

    – All pay types are not fungible
     

14. Using new, lower accounting-based binomial stock option values to derive new stock option grant amounts based on surveys that contain stock option grant values based on high Black-Scholes values.

    – This is complicated but very important

    – In anticipation of stock option expensing, companies are adopting stock option valuation techniques that produce lower values for expensing purposes than earlier Black-Scholes option values, e.g., 20% of option price vs. 30%

    – And currently lower stock market volatility is contributing to this downward pressure on stock option grant values per share

    – Compensation surveys, however, have not adopted these newer techniques and are including in total compensation stock option grant values based on the higher Black-Scholes results from prior years

    – Using the newer formulas with lower values, to recommend stock option grants to the compensation committee that match survey amounts based on higher values would be duplicitous and could warrant termination for cause if done intentionally

    – Committees should insist that recommendations based on surveys use the same assumptions as were used in the survey
     

15. Combining several of the above abuses in such a fashion that any high-paying company could show date to its Compensation Committee that proves its executives are underpaid.

*       This Reference Paper A accompanies a speech delivered by Frederic W. Cook to the Stanford Directors' College on June 20, 2005, and webcast on June 21, 2005 to members of CompensationStandards.com