Much is made of the gap between CEO and worker remuneration yet there is also a worrying pay disparity between one CEO and another
Large sectors of the US public are screaming about rising chief executive officer-to-worker pay ratios that are brushing up against the 500-to-1 mark, if they aren't already above it. But almost no one has noticed another rising pay-ratio series, that of one CEO to another.
What could possibly explain the pay gap as wide as the one between Berkshire Hathaway's Warren Buffett, who earned an average of $350,000 a year over a three-year period, and Citigroup Inc's Sandy Weill, with $127.5m a year, or 364 times as much during the same time?
The short answer is: hardly anything. That's the principal finding of my review of pay for CEOs of 180 major US companies.
The average three-year pay for all those CEOs was $13.9m. And the only explanation for the way they are paid is that board compensation committees appear to give them whatever they want. Those committee meetings aren't what they seem to be. Rather, they are game shows like Wheel of Fortune.
My study of three-year CEO pay at those companies shows you can account for only 15% of CEO pay variation based on differences in company size and precisely 0% based on differences in company performance.
I looked at companies with 2001 revenue of $8bn or more. Each CEO had held his position for three years or more. Data was furnished by Equilar, an independent provider of executive compensation information.
Pay used for the study included base salary, annual bonus, free share grants, the estimated present value at grant of stock options, payments under other long-term incentive plans and miscellaneous compensation.
What could possibly explain that awesome difference between CEOs, such as Buffett and Weill? Average revenue over the three-year period explained 15% of the pay variation, according to regression analyses I ran. But when I tried to coax my computer into improving its explanation of pay variation by taking account of both company size, as measured by revenue, and company performance, as measured by total return over the three-year period, it gave up and declared 15% was the best I was going to get. So 85% of pay variation comes under the category of: your guess is as good as mine.
Table one shows in descending order of overpayment the 10 most relatively overpaid CEOs (average total pay in millions of US dollars).
Look at the column labeled rank. The company with the best performance received a percentile rank of 100, while the company with the worst received a rank of 0. The median percentile rank was 50.
Note that Sandy Weill, for all his excess, delivered performance that was better than all but 2% of the companies. On the other end, Bill Esrey of Sprint Corp could also end up on the 15 most relatively overpaid list, with a performance ranking worse than 93% of the other companies. So much for the sinners. Now for the saints.
At first glance, it would appear high CEO pay is negatively correlated with total return performance. The average performance percentile rank for the 15 most relatively overpaid CEOs was 45, while that for the 15 most relatively underpaid CEOs was a higher 55. But my computer decided there was so much scatter around those averages that there was no significant difference in performance between the Top 10 CEOs and the Bottom 10 CEOs. Neither was there any difference between the performance of either group and the middle 150 CEOs.
Bloomberg newsroom, Las Vegas
Set up Vanguard in 1975
‘In the know’
Owen to pay £3m
Lasting power of attorney
Three risk profiles