June 18, 2013
Since its implementation in 2010, the Dodd-Frank Act has brought about a host of regulations related to corporate governance, and placed executive compensation under greater scrutiny. One effect of this increased scrutiny is the reframing of the pay and performance discussion, especially as more accurate methods of analyzing compensation and performance have surfaced. Among the most common of new disclosures is a realizable pay model. Realizable pay illustrates potential take-home pay over a given performance period, and the model’s results are compared with stock price performance over that same period.
Another critical component of the pay for performance equation is the composition of a company’s peer group. The typical practice for companies is to work with an independent compensation consultant to formulate peer groups using comparative metrics such as revenue, size, and industry classification. For determining proxy voting recommendations, ISS and Glass Lewis use different peer group selection methodologies. These peer groups may not match those compiled by companies themselves, and could ultimately lead to differing perspectives and voting recommendations.
To examine the impact of the various peer group models, this article examines an alternative pay model (realizable pay), total direct compensation, and TSR across three sets of peer groups: company-disclosed peer groups, ISS Simulated Peers, and Glass Lewis Peers. For analysis, pay and performance rankings were determined based on Total Shareholder Return (TSR) results compared to both CEO realizable pay and Total Direct Compensation (TDC).
This article also reviews different disclosures of realizable pay in proxy statements filed between March 1, 2013 and April 15, 2013. The organic growth of the realizable pay model has resulted in several variations. Proxy advisors, ISS and Glass Lewis, have each recently released different definitions of realizable pay. While some companies are already conforming to these new models, many others continue to use their own alternative pay models.
The peer group data used in this article is drawn from those Russell 3000 companies that filed proxies between January 1 and April 30, 2013. With 1,548 companies analyzed, certain companies were excluded from both the peer and target company list: companies without three complete years of stock price data (for TSR calculations), with a latest fiscal year ending December 31, 2011 or older, and with less than five public peer companies.
Three peer groups were analyzed per company: the company-disclosed peer group, Glass Lewis Peers , and ISS Simulated Peers . Because the ISS peer group is simulated, the peers used in this study are not an exact representation of actual peer groups used by ISS.
The definition of realizable pay used throughout this article and particularly in the peer group calculations uses the Glass Lewis definition. This model incorporates the following components over a three-year period for the CEO:
Base salary and bonuses received
Intrinsic value of time-based equity granted during the measurement period, using the closing stock price at the end of the period
Values of performance-based equity earned during the performance period, using the closing stock price at the end of the period
All other compensation
Total Shareholder Return
The TSR analysis examined how companies performed in raising stockholder value against self-disclosed peers, Glass Lewis Peers, and ISS Simulated Peers.
The largest number of companies used to develop each peer group type was between the 25-49th percentile ranges of TSR: 22.4% (347 companies) of self-disclosed peers, 23.9% (370 companies) of Glass Lewis Peers, and 24.5% (380 companies) of ISS Simulated Peers fell into this TSR percentile range.
The variance in the number of companies in the different categories is apparent. The self-disclosed peer group has the smallest variance. The 25-49th percentile range contained the highest number of companies, while the 75-99th percentile range contained the fewest with 329 companies, a difference of 18 companies. Glass Lewis and ISS Simulated Peers showed greater disparity with variances of 49 and 67, respectively. As expected the smallest number of companies comprised groups that performed significantly above or below their peer groups. Companies that rank below their peer groups made up approximately 5% of the 1,548 companies analyzed for all three peer group methodologies. Companies that out-performed their peers in shareholder return made up 6.3%, 6.7%, and 4.5% of self-disclosed peers, Glass Lewis Peers, and ISS Simulated Peers, respectively.
Total Compensation Trends
This analysis examined the percentile ranking of CEO total direct compensation for the three different peer group types.
Glass Lewis Peers have a large number of companies in the 1-24th and 25-49th percentile ranges. However, 44.4% of companies ranked near the median of their self-disclosed peer group, with most companies falling in the 25-49th and 50-74th quartiles. When compared to ISS Simulated Peers, most companies were in these two ranges.
Unlike the TSR analysis, company-disclosed peer groups showed the greatest variance in the number of companies that fell into each quartile. There were 479 companies in the 25-49th percentile while 215 companies ranked in the upper quartile against their peers, a difference of 264 companies. ISS Simulated Peers had the least disparity with only 125 companies separating the largest and smallest number of companies. Glass Lewis Peers resulted in a difference of 195 companies.
For each company in the sample, a percentile rank based on realizable pay results was determined relative to peer companies in each of the three peer groups. The realizable pay calculation shows that companies analyzed under company-disclosed peer groups and Glass Lewis Peers were close in number in each percentile range. Companies analyzed under ISS Simulated Peers fell below the company-disclosed and Glass Lewis numbers in the lower percentile ranges but exceeded the groups in the higher ranges.
It is often the case that CEOs who are externally-hired receive higher compensation packages than those who are internally promoted. Internally-promoted chief executive officers, however, also generally receive an increase in base compensation. In this section, we compare the executive’s compensation prior to promotion to compensation after becoming CEO. As expected, internally-promoted CEOs across each index see an increase in total compensation after their promotions. The largest average pay difference is seen in compensation for executives in the S&P 500, where an executive makes $1.9 million more on average in the CEO role than in his or her previous role. Within the MidCap and SmallCap indices the average increases were $870,015 and $587,252, respectively.
For example, in the 1-24th percentile range, 24.8% and 24.1% of companies using company-disclosed and Glass Lewis Peers fell within this range, respectively, while only 19% of companies using ISS peer groups fell within the same range. In the 75-99th percentile range, company-disclosed and Glass Lewis Peers again placed similar numbers of companies within this range, with 15.6% of companies analyzed under company-disclosed peer groups falling within this range and 17.6% of companies analyzed under Glass Lewis Peers falling within this range. However, companies analyzed under ISS Simulated Peers exceeded the number of companies under the company-disclosed and Glass Lewis Peers with 21.3% of companies analyzed falling within this range. The middle percentile ranges showed the most similar number of companies analyzed using all three peer groups, though companies ranked higher when compared against ISS Simulated Peers.
Though this analysis incorporated the Glass Lewis definition of realizable pay, the ISS definition was also cited by companies this proxy season. Beginning in the 2013 proxy season, ISS will consider realizable pay in its analyses of S&P 500 companies. ISS uses the following methodology for calculating the value of equity and other pay:
“For all share-based awards made during the measurement period, the value (based on stock price as of the end of the measurement period) of awards made during the period (less any shares/units forfeited due to failure to meet performance criteria based on complete and clear disclosure); or, if awards remain on-going, the target level of such awards
For stock options granted during the measurement period, the net value realized with respect to such granted options which were also exercised during the period; for options granted but not exercised during the measurement period, ISS will re-calculate the option value, using the Black-Scholes option pricing model, as of the end of the measurement period
Change in Pension Value and Nonqualified Deferred Compensation Earnings reported for all years
All Other Compensation reported for all years”
(Evaluating Pay for Performance Alignment, ISS, January 2013)
This proxy season, AK Steel was one company that disclosed a realizable pay definition closely resembling this new model. Some main differences, however, are that AK Steel excludes pension and deferred compensation, and assumes all stock awards granted will vest as opposed to using actual payout values for performance-based equity awards.
AK Steel Holding Corp (AKS)
DEF 14A filed on March 12, 2013
“The term ‘realizable pay’ has been defined to include the following compensation items:
Actual base pay;
Incentive payouts and discretionary bonuses, if any, actually paid;
All other compensation actually paid; and
The fair value of all equity grants made, measured at the end of the most recently completed fiscal year. The fair value of stock options was determined using the Black-Scholes model based on year-end assumptions and the maximum remaining life of the options. The fair value of restricted stock and performance shares was determined using the average of the high and the low stock price on the last day of the fiscal year and assumes all vesting considerations were met” (p.35).
In addition, the following companies also disclosed similar realizable pay definitions:
While many companies use realizable pay definitions similar to the Glass Lewis or ISS definitions cited above, many companies elect to use their own definitions. Some companies use Summary Compensation Table values but exclude one or two columns. Others choose to replace equity granted within a fiscal year with equity granted for performance during that fiscal year, a common practice of financial companies. The most common alternative uses the realized value of equity, calculating the taxable income going to the executive over the period. Realized equity awards consist of shares that vest, and the gain on options that are exercised.
Exxon Mobil is one company that included realized pay disclosure in its 2013 proxy:
Exxon Mobil Co. (XOM)
DEF 14A filed on April 12, 2013
“Realized Pay is compensation actually received by the CEO during the year, including salary, current bonus, payouts of previously-granted Earnings Bonus Units (EBU), net spread on stock option exercises, market value at vesting of previously-granted restricted stock, and All Other Compensation amounts realized during the year. Excludes the value of new/unvested EBU and restricted stock grants, change in pension value, and other amounts that will not actually be received until a future date” (p.31).
While this model most accurately captures the “take-home” pay of an executive, there is some pay-for-performance disconnect due to the fact that exercising options is a timing choice usually connected to an award granted many years earlier. Proxy advisors have elected to construct models that require more effort than the realized pay model, to avoid penalizing an executive for equity gains earned over a much longer time horizon than the period being measured, which is typically three years as opposed to the usual ten-year term of an option.
The following are examples of companies that disclosed realized pay in their 2013 proxies:
1 Glass Lewis uses Equilar Market Peers, which pulls market data and social analytics to create an
interconnected network of peers that will generate 15 companies with the strongest connections to serve as a
logical peer group. For more information on Glass Lewis’ use of Equilar Market Peers visit:
2 ISS Simulated Peers were simulated for this analysis using current ISS selection criteria which
includes 14-24 peers based on GICS industry classification, size constraints for revenue (assets for certain
financial companies), and market value as disclosed on its website:
Please contact Dan Marcec at firstname.lastname@example.org for more information. Dan Marcec is the Director of Content & Marketing Communications at Equilar. The contributing authors of this paper are Andrew Gordon, Project Manager, Courtney Yu, Senior Research Analyst, BJ Firmacion, Frank Gonzalez, Jason Hsieh, and Julia Lee-Rittell, Research Analysts.