Large CEO shareholdings aren't necessarily beneficial.

WHAT'S with billionaire chief executive officers? Some, recognizing their vast wealth, pay themselves literally nothing or hardly anything. Others can be seen standing at the front of the line as the doors to their company's treasury swing open.

I examined the share ownership of 457 CEOs running U.S. companies with current market caps of at least $3 billion. Of these, I concentrated on the 26 who had the most actual shares and paper profits in unexercised option shares. In virtually every case, the size of their holdings was $1 billion or more.

Among this group, there are two genuine pay saints: Kinder Morgan Inc.'s Richard Kinder and Pixar's Steve Jobs.

Kinder earned precisely nothing over the three years that ended with 2003, while Jobs earned $53 a year. In his other day job as chief executive of Apple Computer Inc., Jobs has been decidedly less saintly, accepting, among other things, a free personal jet and, at one point, the largest stock option grant ever made on a single day.

The evil twins of Morgan and Jobs are Viacom Inc.'s Sumner Redstone ($32.9 million in pay per year) and UnitedHealth Group Inc.'s William McGuire ($27.1 million).

These and all other pay statistics being used are based on average annual total pay over the three years ended with 2003. Total pay is the sum of all elements of the package, including base salary; annual bonus; the value of free shares of stock, measured at the date of award; the estimated present value of stock options, measured at the date of grant using the Black-Scholes option pricing model; payouts under other long-term incentive plans and miscellaneous compensation. Data for this study were obtained from Aon Consulting's eComp database.

Three main findings

The study produced three principal findings.

First, the dispersion in pay among the 26 billionaires is awesome. Why Redstone, for example, feels the need to be paid far more than a professional working-stiff CEO earns, is best left to a moral theologian 

Second, the average billionaire CEO is paid neither higher nor lower than the average non-billionaire CEO. Although the average CEO among the 26 billionaires earned 7 percent less than the market rate of pay for a company the same size, that difference proved to be statistically insignificant.

Third, the notion that having a CEO who owns a ton of stock and who, as Warren Buffett is fond of observing, eats his own cooking will guarantee a superior return compared with other company investments seems unfounded.

For each company in the study, I calculated a weighted average total return in three time windows: three years, two years and one year, all ended Dec. 31, 2003.

These results were then transformed into what I call a performance IQ score. Like intelligence IQs, the average CEO had a performance IQ of 100. In a perfectly normal bell-shaped curve distribution, 95.4 percent of scores will fall between 80 and 120 (the actual percentage for this study was 94.1 percent), while 99.7 percent of the scores will fall between 70 and 130 (actual performance was 99 percent). The average performance IQ score for the 26 billionaires was 103. This difference was statistically insignificant when compared with the non-billionaires.

Ellison and Bezos

As with pay, there is a wide range of performance, with Larry Ellison's IQ score of 86 putting him in the bottom 8 percent of performers, and Jeffrey Bezos beating every billionaire and non-billionaire CEO with a score of 135. The normal probability of achieving such a high score is 20 out of 100,000.

There also wasn't a statistically significant link between the size of a CEO's shareholdings and his performance IQ score. Which brings up the value of insisting that a CEO hold a lot of company stock.

It seems intuitively correct that a person whose net worth is mainly tied up in his company's stock will work harder and smarter than one who holds hardly any shares and who dumps any option shares the moment he exercises them.

That belief has prompted many companies to adopt so-called share ownership guidelines. For example, the CEO is told that within five years, he must own shares worth at least five times his then-current base salary. And unexercised option shares don't count.

Based on this study and similar studies I have conducted in earlier years, that intuitively correct belief isn't borne out by the stats. Eating your own cooking can sometimes cause dyspepsia.

One could, of course, argue that having the CEO load up on company shares, even if it doesn't necessarily help matters, at least doesn't hurt, either.

Wrong. Because in companies with share ownership guidelines, there is often an attempt to help the CEO out by giving him more than the usual number of option shares and/or more than the usual number of free shares, so that satisfying the guidelines can be accomplished without breaking a sweat. The cost to shareholders in such cases is quite high.

Graef Crystal is a columnist with Bloomberg News.