Several companies recently announced intentions to count the value of stock options granted to employees as a compensation expense in calculating corporate earnings. Among the leaders of the movement are the Washington Post and Coca Cola, both companies heavily influenced by director Warren Buffet, who has for years opposed employee options, and who does not use them in Berkshire Hathaway. <?xml:namespace prefix = o ns = "urn:schemas-microsoft-com:office:office" />
The expense-'em train is picking up speed. The International Accounting Standards Board is with the expensing program, and Harvey Pitt, chairman of the SEC, declares the change "inevitable," even though he opposes it.
Leaving aside philosophical debates about whether treating options as compensation expenses is a good idea — I am firmly on the "no" side — these proposals raise difficult practical questions. In the end, the numbers produced will be quite slippery, and one waits with anticipatory schadenfreude the discovery that companies that start expensing the options are misleading investors. Politicians, the press, and the trial lawyers, all of whom specialize in amnesia about what they demanded be done only yesterday, will stand ready to flay them for it.
The first question is how to value the options granted to employees. Under current accounting rules, a company discloses option grants in the Form 10K it files with the SEC, and assigns them a value calculated according to the standard Black-Scholes option pricing model. The model uses six terms: current market price of the stock, volatility of the price, exercise price of the option, current market rate of interest, term of the option, and dividends.
This is a fine model, for which its designers won a Nobel Prize, but it was constructed for options with exact lifetimes, histories that provide a market estimate of volatility, and an underlying market in the security that allows an options trader to hedge by buying or selling the underlying stock. Black-Scholes is inexact for employee options — these may not vest for years, may remain open for more years, can be heavily influenced by the activities of the employees themselves, and have other bells and whistles not found in conventional options.
Writing in the Washington Post, Buffett was contemptuous of these concerns, saying Black-Scholes was perfectly adequate. But when Coca Cola announced its shift to treating options as expenses, it waffled on the valuation question. It said it would "adopt" Black-Scholes, but that the actual valuation would be the average of "firm quotations" received from independent financial institutions to buy or sell Coke shares "under the identical terms of the stock options granted."
This is pretty artful. Obviously, Coke does not agree that the mechanical application of Black-Scholes is possible; investment-company judgment must be brought into play. But, are "firm quotations" actual bids? How will a financial institution deal with the fact that an employee option vests only over five years and is forfeited if the employee leaves, and with other special provisions? And if Black-Scholes is not routine for a company like Coca Cola — with a familiar and straightforward business, little high-tech input or uncertainty risk, and a market history going back to 1886 — how can investment firms appraise options in the latest high-tech company? Their recent record does not exactly inspire confidence.
The next question is timing: When should the expense be recognized? There are three options: 1) In the year of the grant; 2) Over the vesting period of the option (commonly five years); and 3) Over the economic life of the option (commonly five years of vesting plus a five-year window for exercise).
Almost everyone rules out the first choice. We are talking about income statements here, and the most elementary proposition of accrual accounting is that expenses should be matched with the income they produce. Options produce value over time, and cannot be taken as an expense in the first year, because this would be misleading. (Sort of like recognizing revenues that will only be received in later years.)
So the charge must be spread out, which is where the fun begins. Under the Coca Cola plan, the option is given a value in year 0. This will be charged off over the vesting period. One can argue that it should be charged off over the entire period that the option is open, since the company will continue to receive economic value as long as the employee stays. But leave that aside. Another puzzle is how to charge: straight-line, declining balance, or another depreciation exotica? Again, leave it aside.
The real problem is that the value of the option will change over the years. Year 2000 Washington Post options exercisable over seven years at a price of $544 per share carried a Black-Scholes value of $161 per share at the time of grant. Year 2001 options exercisable at $523 were worth only $108 when granted. The Post will, apparently, charge them off over five years. (Its 10K is a bit murky on this point.) So, in 2002 it will expense the year 2000 options at $32 per share and the 2001 options at $22.
But clearly the later options are worth more to the employees holding them than are the earlier ones, since they have a longer life and lower exercise price, which creates a discontinuity between the accounting numbers and the rewards as perceived by the employees. In 2005, the Post will still be charging off the same $32 and $22, even though the options from both years will have values ranging from zero (if they are under water) to hundreds of dollars (if stock has boomed). Hence, the actual rewards being reaped by Post employees will be misrepresented by the accounting numbers, which means that the depiction of the outlays necessary to produce the Post's revenues will be inaccurate.
This does not matter much to the Post, which grants few options and appears to be phasing them out anyway, but it matters a lot to other companies with different business characteristics, especially high-tech companies that are trying to use options to combine capital in the form of money with capital in the form of intellectual firepower.
Finally, consider the revenue side. The expense-'em crowd argues that options represent a real cost to the company because of opportunity cost. In this view, the company could sell the options on the open market and collect cash, but it is foregoing this opportunity and giving the stock to employees. Clearly, they say, if the company sold options and transferred the cash we would recognize it as an expense.
This view elides an important consideration: the financial result of an open-sale-plus-cash transfer would be a wash. Money in would equal money out. So if you treat a grant of options as the equivalent of this, it would be strange to recognize the pay-out without recognizing the income. But is it operating revenue? No way. You could call it a capital infusion, but that would logically lead to the conclusion that the payment to employees is also a capital transaction, not an expense item for the income statement.
So what will Coca Cola and its confreres do about the revenue side? Everyone is silent about this; it is not clear what companies that already expense options do. But the companies must, somewhere, recognize the in-flow dimension, because the books have to be in balance and you cannot have an outlay without showing a compensating inflow. So, are we now to count lost opportunity cost as "revenue"? If the expense-'em advocates have a clear answer, they are not sharing it.
The current rule is that companies disclose their options and provide a Black-Scholes estimate of value. Earnings per share are calculated on a diluted basis, which means that they reflect options that are in the money. Investors can then make of all this what they will. The disclosures could and should be expanded, but the basic system is simple, clear, and cheap to implement. The proposed changes will make the system complex, murky, and expensive, with a degradation rather than an improvement in the quality of the information available to investors.This is not progress.