Magazine

Commentary: The Artificial Sweetener in CEO Pay


For Chief Executive Officer George M.C. Fisher, 1999 was not exactly a Kodak moment. As a withering price war with Fuji cost his company market share in film, and with little to show for his much-vaunted digital strategy, Eastman Kodak Co.'s (EK) chief found himself in a financial quagmire. Profit growth was flat and shareholder returns fell by 5.4%, a fact board members found distressing. But that didn't stop them from richly rewarding him with a $2.5 million bonus, up 47% from 1998, and options valued at more than $2.9 million.

In this age of pay for performance, giving an underperforming CEO a raise seems to defy logic. But things become clearer when you read the fine print in the proxy. Fisher, who has since left Kodak, got his bonus increase in part because of the company's 5% revenue growth--and his option grant in part because of a practice that compares CEOs' pay with their peers'. Kodak spokesman Gerard Meuchner notes that Fisher's base salary hadn't changed since he joined the company in 1993, and in 2000 both his bonus and option grant declined significantly, despite profit growth of 1.4%. Still, why did Kodak use industry pay data to set Fisher's option grant? Says Meuchner: Because Kodak competes with those companies for executive talent.

The technique is called competitive benchmarking, and in the 1990s, it became widespread. By one estimate, 96% of companies in the Standard & Poor's 500-stock index use it to set pay. Company boards figure that if a CEO doesn't earn as much as his peers, he'll take a hike.

It's logical that companies would want to use some comparison in deciding on pay rates. But it may be time to reconsider the extent to which companies lean on benchmarking in pay decisions. A new study suggests that the practice was a main culprit in the skyrocketing CEO pay of the 1990s, when average compensation of chief execs at top U.S. companies grew more than 500%, from $1.9 million to $12.4 million. And that's not all. The authors of the study--John Bizjak of Oregon's Portland State University, Michael L. Lemmon of the University of Utah, and Lalitha Naveen of Arizona State University, all finance professors--found that the "underpaid" CEOs, who are benchmarking's main beneficiaries, are the worst performers of the bunch. While no one was looking, Bizjak says, corporations institutionalized a practice that rewards the least deserving. "It weakens the link between pay and performance," he says.

Pinpointing the inflationary effect of benchmarking is tricky. Companies don't say how much of an exec's raise is designed to bring him into line with industry peers. Still, Bizjak and his colleagues found that CEOs who earn less than the median for their peer group get salary and bonus raises twice as big as those granted to CEOs paid above the median. As an example, he cites wholesale trade. In five of the six years covered by the study, "underpaid" CEOs got raises in base salaries and bonuses that far exceeded those of their higher-paid brethren.

VICIOUS CIRCLE. In most industries, CEOs paid above and below the median get annual raises, with those allotted to "underpaid" execs being much bigger. When that happens, the median grows by leaps and bounds, prompting new raises for the underpaid, and the cycle continues. From 1993 to 1998, in the wholesale trade example, salaries and bonuses grew 65%, from $737,900 to $1.2 million. Bizjak won't venture a guess as to how much of the increase came from benchmarking, but others will. Fred Cook, managing director of Frederick W. Cook & Co., a New York compensation consultant, estimates that up to half the 1990s increase was from benchmarking.

What makes this research rise above the ordinary are the data on performance. Bizjak and company found that the CEOs who got the biggest raises didn't generate commensurate performance. Companies giving big raises to their below-median CEOs trailed their less generous counterparts in sales growth, return on assets, and total return.

To be sure, there are valid reasons for boards to make pay comparisons. More often than not, the reason cited by board members is retention. Without assurances of comparable pay, many feel their CEOs would head for the door. Apparel maker Russell Corp. (RML) gave John F. Ward an increase in 1999 after its stock plunged 15%, in part because he met earnings goals, and in part because of benchmarking. "We want to make our CEO happy, and the best way to make him happy is to pay him commensurate with our competitors," says Russell compensation committee member Herschel M. Bloom. "What our competitors are paying plays a role."

Many company directors, not surprisingly, deny they're paying big dollars just to keep the boss from walking. Says Jess T. Hay, compensation committee chairman for SBC Communications Inc. (SBC), which used benchmarking to boost the '99 pay of CEO Edward E. Whitacre Jr.: "We're playing for the very long haul. Ed Whitacre has been as able a chief executive as there is in the country today."

When that isn't the case--when a CEO is not performing up to par--companies do shareholders no favors by increasing their pay. The more logical course: Pay him exactly what he's worth, sending a powerful message that poor performance will not be tolerated. Some CEOs will buckle down and start earning their keep. Others will threaten to leave. Let them. Sure, a CEO search will be disruptive, but in a weakening economy, finding a replacement becomes easier every day. In either case, the company wins.

Corrections and Clarifications

"The artificial sweetener in CEO pay" (Management, Mar. 26) gave an incorrect pay comparison for Russell Corp. Chief Executive Officer John F. Ward. The article should have noted that Ward's salary and bonus in 1999 was for a full year, while his 1998 pay covered nine months. On an annualized basis, the year-to-year increase was 2%.

By Louis Lavelle

Lavelle covers executive compensation from New York.


American Apparel's Future
LIMITED-TIME OFFER SUBSCRIBE NOW

(enter your email)
(enter up to 5 email addresses, separated by commas)

Max 250 characters

 
blog comments powered by Disqus