Computer Firms Use Irrational Standard in Giving Stock Options

On March 25, 1994, the normally staid world of accounting was shaken to its foundations. The Financial Accounting Standards Board (FASB), the rule-making body for American accounting, held an open meeting in San Jose to hear comments on its proposal to charge the present value of stock option grants against corporate earnings. For the past 40-odd years, the accounting world had been holding a sale, as it were, on stock option grants. Year after year, CEOs of major companies had exercised stock options and reaped millions of dollars in compensation, yet not one penny of that gain was charged against their profits. In its proposed ruling, the FASB sought to create a level playing field among all forms of compensation, all of which, with the single exception of stock options, were then being charged against a corporation’s earnings.

The FASB’s proposed ruling engendered a firestorm of criticism from the business community, and the center of that storm was in San Jose, the principal home of the nation’s computer and software industry. CEOs in that industry argued that FASB’s action would ruin perhaps the one viable business left in America. It would only be a matter of time, they claimed, before the Japanese crossed the Bayshore Freeway and took over Silicon Valley. They shouted to anyone who would listen their ardent belief that stock options were the sine qua non of incentives and that the prospect of being enriched by stock options was the fuel in the tanks of this most vibrant of American industries.

*

Perhaps nothing better illustrated the hostility of the computer and software industry to the FASB’s proposed action than the fact that--just three miles across town, and at the very time FASB was holding its public hearing--FASB’s opponents had hired the San Jose Convention Center, along with a marching band, and were in the process of conducting a well-rehearsed rally with 5,000 cheering supporters of the status quo.

After months of further wrangling, complete with interference by Congress and the Clinton White House, the computer and software industry won the fight. There will be no charge to earnings for stock option grants, no matter how large they are. Thus, computer and software companies can continue to report Alice-in-Wonderland profits to their shareholders, even though professional stock analysts know full well that those earnings are vastly overstated.

For its part, FASB will shortly decree that what would have been a charge to earnings will now be reported in an obscure footnote to a company’s income statement. And if the Securities and Exchange Commission has its way, that obscure footnote will be whisked out of the annual report and put in the quarterly 10-K.

Now, if any group of companies should know how to manage a stock option program to obtain maximum motivation on the part of executives and employees, it should be the computer and software industry. Wrong. I just finished a study of 61 CEOs in the computer and software industry. They were chosen from among the 900 companies comprising the Standard & Poor’s 500 and Mid-Cap 400 Index groups.

*

Now, things seem rational enough in this industry if you look only at CEO base salaries. Forty-two percent of the variation in base salaries is directly related to differences in company size, as measured by 1994 revenues. The bigger the company, the higher the base salary. Nothing wrong there.

And things seem rational enough in this industry if you look only at the sum of base salary and the annual bonus. There, 39% of the variation in pay can be accounted for by differences in company size, and another 14% can be accounted for by differences in company performance, in this case, compounded annual three-year total shareholder return to investors (i.e., stock price appreciation and dividends).

But things fall apart when the present value of stock option grants is added into pay. The value of those grants dwarfs every other element of the CEOs’ pay packages. In 1994, stock options accounted for 43% of total pay--higher even than the 42% accounted for by the sum of base salary and annual bonus.

*

Those who work with the spreadsheet programs designed by some of the very companies in my study will be acquainted with the so-called @RAND function. If you use this function, you get a random number--usually between 0 and 1. Most of us who employ this function presumably think it was put there for our use. Maybe it wasn’t. Maybe it was put there so that boards of directors of computer and software firms could determine how big an option grant to give their CEOs! (For you non-computer users, that’s a joke.)

For an illustration, consider two CEOs--Robert B. Palmer of Digital Equipment and Andrew S. Grove of Intel. During the 1992-94 period, Palmer received options on $8.8 million of stock per year, and Grove received options on only $2 million of stock per year.

If the option grants made to Palmer and Grove had been made on the basis of company size and performance, then Palmer would have received options on $2.1 million a year--not $8.8 million-- and Grove would have received options on $5.5 million, not $2 million. Though the firms are roughly the same size, Intel has been far more profitable.

Think here about a 100-yard dash. The participants are all lined up at the starting point. Then the pistol is fired, the race begins and the fastest person wins. But that’s not how the stock option race is run in the computer industry. Some of our 61 CEOs are indeed lined up at the starting point. But then we have Grove, who is starting the race five miles from the stadium because his grant is so much smaller than it should be.

Palmer, on the other hand, is starting the race a yard from the finish line because his grant is so much larger than it should be. When the starter’s pistol is fired, all Palmer has to do to win is fall forward. Grove, however, is likely to be standing five miles from the stadium even after the race is finished because he would have been unable to hear the pistol shot.

Put another way, if Grove manages to double the value of his company’s stock with five years of his option’s grant, he would make $2 million. But Palmer would make the same $2 million if he managed to increase the value of his company’s stock by only 23%.

Recall here that 39% of the variation in base salary and bonus was attributable to company size and another 15% to company performance. Well, when all those random-number-like stock option grants are thrown into the pool of CEO compensation, I found that I could account for only 20% of the variation in the new grand total pay (including stock options) on the basis of company size and precisely none of the variation on the basis of company performance. In other words, stock option grants in the one industry where you would think they would be intelligently used actually destroyed all the relationship between company performance and CEO pay and most of the relationship between company size and CEO pay.

For a contrast, consider a study I just completed on the CEOs of 54 major commercial banks. When I examined the combination of their base salaries and their bonuses, I found I could explain 78% of the variation on the basis of their companies’ size and three-year shareholder return performance. And when I expanded my definition of compensation to cover various long-term incentives, including stock options, I found I could still account for 75% of the variation in pay on the basis of size and performance. Now there’s a rational market. There’s a group of companies that know that in granting stock options, they’re playing with live ammunition.

So what can be done? Well, the CEOs of Silicon Valley were able to get the Congress to threaten to disband the FASB if it committed the folly of requiring a charge against earnings for stock option grants. So I’m going to start a letter-writing campaign to our senators and representatives to outlaw the use of the @RAND function. Of course, those boards of directors will probably turn to the low-tech solution of flipping a coin. But maybe it’ll get lost in the thick pile of the boardroom carpet.

(BEGIN TEXT OF INFOBOX / INFOGRAPHIC)

Total Disconnect

Stock options for CEOs in the computer industry often have little relation to company size and peformance. Indeed, as the following chart indicates, the sum total of the compensation for the industry’s top five performers is less than that for the industry’s bottom five performers.

Top Five Performers

*--*

Total direct Company Performance compensation and CEO percentile* in 1,000s** Micron Technology 100 $1,387 Joseph L. Parkinson Atmel 97 383 George Perlegos LSI Logic 97 1,190 Wilfred J. Corrigan Oracle Systems 96 5,236 Lawrence J. Ellison EMC/MA 93 629 Michael C. Ruettgers

*--*

Bottom Five Performers

*--*

Total direct Company Performance compensation and CEO percentile* in 1,000s** Intergraph 5 306 James W. Meadlock Cray Research 3 1,392 John F. Carlson Conner Peripherals 3 4,424 Finis F. Conner Digital Equipment 1 2,555 Robert B. Palmer Structural Dynamics Rsrch. 0 2,182 Ronald J. Friedsam

*--*

* The top five and bottom five performers and their combined percentile ranks in one-, two- and three-year shareholder return. A rank of 0 represents the lowest-performing company; a rank of 100 is the highest.

** The sum of base salary, annual bonus, restricted stock, long-term incentive payouts, the present value of stock option grants and miscellaneous compensation.

Sources: Standard & Poor’s, Crystal Report