HE Financial Accounting Standards Board published its long-anticipated statement about options accounting last week, and as expected it recommended that options grants to employees be counted as an expense in corporate financial statements. Barring any last-minute changes, this will probably become a requirement of the Securities and Exchange Commission in 2005.

Technology companies have doggedly lobbied against this change, but the corporate scandals of the last few years have made Congress reluctant to block moves toward more transparent accounting.

The central issues in options compensation are examined in "The Trouble with Stock Options," an article in the summer 2003 issue of The Journal of Economic Perspectives by Brian J. Hall of the Harvard Business School and Kevin J. Murphy of the Marshall School of Business at the University of Southern California. (A summary of the article is at http://www.nber.org /digest/mar04/w9784.html.)

Under current accounting standards, companies have to describe in detail the option grants that they make to employees, but they are not required to estimate or report a value for these options. Under the accounting board's proposed change, companies will be required to assign a specific value to these options and subtract that value from earnings.

Researchers like David Aboody at the University of California, Los Angeles, have found that financial markets already seem to do a pretty good job in valuing options grants, so such a change is unlikely to have a big effect on stock prices.

Despite such findings, many technology executives say they believe that shareholders will force them to cut back on options grants if they must expense them, and that this will make it harder to attract and retain employees.

Professors Hall and Murphy argue that options are an inefficient form of compensation, particularly for rank-and-file employees, who have little influence on corporate decisions and are most likely averse to taking unnecessary risks. Such employees would, it is argued, be better off with more liquid and less risky rewards.

This is certainly in accord with conventional economic theory, but then again, according to economic theory, Las Vegas should not exist.

In reality, people sometimes like to gamble, and a young engineer just out of college might well choose a job with a low salary but a chance to strike it rich. All the engineers I know are convinced that they are experts at picking winning technologies.

Silicon Valley executives are probably right that offering options has been an effective way to induce people to work for new, risky companies.

(Disclosure: I have received options grants from various start-ups as compensation for consulting and advisory services. I have also dropped a few dollars into the slot machines in Las Vegas, but don't tell any of my economist colleagues.)

However, even if employees sometimes find risky compensation attractive, there is no reason that giving employees stock instead of options would not be just as effective. Indeed, some companies - Microsoft is a notable example - have replaced their options programs with stock grants, without mass defections.

Publicly traded stocks have the attractive feature of explicit market values. The value of options, on the other hand, generally has to be estimated. One problem with the accounting board's proposal is that it does not recommend a specific way to value options, but allows companies to choose among various option-pricing models.

The traditional choice has been the Black-Scholes formula, but it applies to options that are freely traded on organized exchanges with lifetimes of a few months.

Employee options generally have a long lifetime and cannot be traded. Restrictions often apply to the exercise of the options, and typically they must be forfeited if employees leave the company.

In principle, the "binomial pricing model" allows for adjustments for some of these additional complications. It gives the same options values as the Black-Scholes formula, given the same assumptions, but is more flexible in that it shows new values as the assumptions are changed. Even so, it is a model, not magic.

For example, the binomial model allows the incorporation of changes in the risk posed by the underlying stock. But some pattern has to be specified about how that risk is assumed to change: is the stock going to fluctuate more in the future than it does now, or less?

A bad assumption will yield a bad estimate, even if the model itself is accurate.

If companies adopt radically different ways of measuring option values, it could become even more difficult to compare earnings across companies than it is now.

Investors may become suspicious of complex methods of options valuation, figuring that companies using the methods must have something to hide. If using a complex valuation for options becomes a negative signal, well-established public companies may switch to restricted stock grants or other forms of compensation.

Companies that stick with options will still be faced with a difficult valuation problem. Two Stanford economists, Jeremy I. Bulow and John Shoven, have outlined a novel way that companies could account for options that might help solve this problem.

Typical options grants require that employees who leave the company exercise their options within 90 days. Hence, long-lived option grants can be viewed as 90-day options that are extended every quarter that the employee stays with the firm.

Mr. Bulow and Mr. Shoven propose that options therefore should be treated as a continuing expense: each quarter, the company would subtract the value of extending its existing options another three months as a compensation expense, along with the three-month cost of any newly granted options.

Compared with proposals that require a once-and-for-all valuation at the time the options are issued, the Bulow-Shoven method would probably give a more accurate picture of the cost of outstanding options grants on a continuing basis.

Companies may decide that recalculating the value of past options grants every quarter is not worth the trouble, and move to straight stock grants.

But this is unlikely to have a negative impact on recruiting engineering talent: if hot-shot engineering grads are attracted by the opportunity to strike it rich, holding speculative stock should be just about as attractive as holding speculative options.

Hal R. Varian is a professor of business, economics and information management at the University of California, Berkeley.