Intel (NASDAQ:INTC) CEO Craig Barrett strafed advocates of stock option expensing this past week with an article in The Wall Street Journal.

In his piece, "'Kind of Right' Isn't Good Enough," Barrett took a swipe at former Federal Reserve Chairman Paul Volcker, who stated recently that he thought "it would be better to expense stock options inaccurately than not at all." Barrett, on the other hand, believes the methodology used to value options does not allow for accuracy and is, as such, insufficient.

Barrett makes a good point, to an extent. The accounting should be as accurate as possible. And the Black-Scholes pricing model, as revolutionary as it is, was not created to value employee stock options because there's no external market for them -- employee stock options cannot be bartered or sold.

OK, let's stipulate this: The financial cost of expensing stock options is an elusive figure. But does such logic excuse us or the Financial Accounting Standards Board or companies from trying to come up with a better way? Apparently, it does for Barrett, the folks at the American Electronics Association (AeA), and for Sens. Barbara Boxer (D-Calif.) and John Ensign (R-Nev.), who have proposed legislation that would make it illegal to expense employee stock options for the next three years.

If "Kind of Right" were actually the problem, these folks would be seeking the better solution. They're not, and so, for Barrett and others, it is a convenient toehold from which they may stonewall the folks who are seeking to make accounting more accurate.

Barrett's veiled accusation in the piece, though, is what I found bizarre. Under Sarbanes-Oxley, Barrett, as CEO of Intel, must certify his company's accounting presentation. Were stock options expensing to be enacted, the argument goes, since it is inaccurate, CEOs would be left with two choices: "... comply with Sarbanes-Oxley and certify as accurate numbers that are inherently flawed. Or, support the spirit of the new law and refuse to sign off on the numbers because we don't believe they present an accurate financial picture."

You think they're accurate now?!
The funny thing is, any person with any knowledge of accounting could let Craig Barrett in on what ought to be a poorly kept secret: There is no number in the income statement or the balance sheet that doesn't already use assumptions, estimates, and generalizations. You would think, for example, that if accuracy were the issue, the treasury stock method for accounting for stock options, the one currently used by Intel, would be absolutely spot accurate. Instead, we find that this method uses some assumptions that can under- or over-estimate the dilution caused by employee stock option grants.

Rebecca McEnally, CFA, vice president for advocacy at the Association for Investment Management and Research, describes the treasury stock method thusly: "When accounting for shareholder dilution, the treasury stock method assumes that the hypothetical funds received by the company from the options exercise are used to buy back stock on the open market to reduce the dilutive effects." In other words, that diluted number of shares account at the bottom of the income statement is not an exact number. It's an approximation, not just of the options and warrants that are in-the-money at period-end, but also of the dilution actually caused by stock option exercising.

Whether or not the company actually did use these funds for repurchase does not matter -- the accounting is the same no matter what, and companies are not obligated to do so. Net effect, the amount of dilution as it stands, is not accurate.

What about other areas? Well, Intel uses as a component of its inventory valuation first-in-first out (FIFO) accounting. FIFO does a great job of valuing inventories, because most recent items in are valued. FIFO also, in markets with rising prices, can understate the cost of goods sold, thus making net income from operations seem higher. That's right, just by selecting its methodology for accounting for inventories, Intel can make its income seem higher than it would otherwise. By Barrett's definition, this is inaccurate. By my definition, actually, FIFO (or in Intel's case, a modified FIFO) accounting is more aggressive than LIFO.

And Intel was the same company that, during the late 1990s exhorted analysts to include its stock market gains in the continuing operations portion of its income statement. Never mind that these gains, by definition, are one-time in nature. As I noted in August in last year, "The bleating about earnings from investments stopped at about the same time the bleeding from investments began." I wonder, if you asked Barrett now which one he considers accurate: that investment gains be included as his company argued in 2000, or that losses be considered one-time, as they argued in 2002?

The point is not to berate Intel here, though I don't mind throwing them a little chin music for their calls for divine accounting. Accounting requires assumptions, period. There are multiple ways to calculate earnings and inventory. Accounting value for assets can differ wildly from their true value. Pension accounting amounts to an existing kitty and a pile of guesses.

Quite simply, there are large portions of the financials that you receive right now that amount to educated conjecture. And Barrett's argument is that this one element, this one thing that is based on the future price of a stock, ought to be 100% accurate? Please.

350,000 pieces of nothing
It was probably little more than delicious irony that another news report came out this weekend saying that the top brass at Intel had just been granted 2 million stock options, 350,000 of which went to Barrett. That's 350,000 stock options that cannot be used to pay for something else. Yep, it's true -- companies use non-cash considerations to pay for goods and services all the time. Had those 350,000 options just paid to Barrett been used to pay the people who supply the coffee at Intel's offices, the company would have expensed them. The exact same pieces of paper. And companies do it all the time, with options and with warrants.

Funny that you don't see too many companies putting out earnings reports today that say, "Earned 4 cents, more or less. We don't know because we paid some of our suppliers with stock options." Nope, they value them to the penny. And they certify their financial statements. Options are a form of currency, plain and simple, and their value is not zero. Neither is their cost. In fact, the way McEnally sees it, traders are already using derivative strategies to price out long-term risks, and with options the major components of variability -- price volatility and average cancellation -- are known numbers, at least on a historical basis.

I'll make one last point. Does it not strike anyone as more than a little irresponsible for companies such as Intel to use something as a major compensation element, something that they themselves cannot value?

Bill Fleckenstein proposed recently that options be expensed upon exercise. That would be better than what is in place today. Even if it would delay expenses for a period of years, it would eliminate this nonsense about "completely wrong" being better than "kind of right."

Fool on!
Bill Mann, TMFOtter on the Fool Discussion Boards

Bill Mann is the Senior Investing Editor for The Motley Fool. For Bill's best stock ideas and exclusive in-depth analysis each month, check out our newsletter, The Motley Fool Select. The Motley Fool is investors writingfor investors.