The continuing brouhaha over a new accounting rule that would require expensing of employee stock options could amount to a tempest in a teacup, if an alternative valuation model is embraced by regulators and proves to be as accurate as advertised.
To be sure, corporate lobbyists continue to complain about the rule to legislators, the Securities and Exchange Commission (SEC), and the accounting standards-setter, the Financial Accounting Standards Board (FASB). The anti-expensing argument, if anyone needs reminding, is twofold. First, say the critics — the most vociferous of which hail from the high-tech industry — a rule requiring expensing is inappropriate because the traditional valuation method, known as Black-Scholes, produces wildly inaccurate results. And even if it didn’t, goes the gripe, imposing such a requirement would be a mistake because it would limit entrepreneurial companies’ ability to attract, motivate, and retain key employees. That limitation, add the critics, would curtail their ability of those companies to innovate, depriving the nation of a key economic edge.
Some experts contend, however, that the alternative method for valuing these employee incentives is not only far more accurate than Black-Scholes, but also could significantly reduce the added expense that companies would have to report. And that obviously could go a long way to undermine the argument against expensing options.
FASB has yet to issue its rule, but its exposure draft (anticipated in a matter of days) is expected to embrace the alternative “binomial” valuation method, to one degree or another. The binomial method is more robust than Black-Scholes, though both techniques need to be adjusted to take into account an essential fact: Employees rarely hold the options they’re granted for the entire length of their typical 10-year term.
More Flexible but Less Familiar
Unlike Black-Scholes, the binomial method divides the time from the option’s grant date to the expiration date into small increments. Since the share price may increase or decrease during any interval, the binomial model takes into account how changes in price over the term of the option would affect the employee’s exercise practice during each interval. The binomial model can also consider an option grant’s lack of transferability, its forfeiture restrictions, and its vesting restrictions — even for options with more-complicated terms such as indexed and performance-based vesting restrictions.
Consultants who favor the binomial method contend that this flexibility is critical. “You need to look at the value that employees attach to a specific set of compensation requirements,” explains Ron Rudkin, a vice president of the Boston-based Analysis Group. “You can’t do that with Black-Scholes,” says Rudkin. “It just doesn’t have the flexibility.”
As a result, the new method produces a much lower estimate of the value of the same grant, according to a study by Analysis Group. Anywhere from 28 percent to 56 percent lower, in fact, compared with the cost estimated under a Black-Scholes model modified to reflect the length of time the options are likely to be held.
To be sure, adjusting the binomial model for this purpose isn’t easy. For one thing, the new model is far less familiar than Black-Scholes, so users must spend considerable time figuring out how to use it. “Black-Scholes is so widely used that there are lots of software packages, for laptops and handheld computers, to run the model,” says Rebecca McEnally. A vice president of the Association for Investment Management and Research (AIMR) — an analyst group that advocates expensing options — she is also an advisor to the International Accounting Standards Board (IASB), FASB’s counterpart outside the U.S. The binomial method “is not quite as easily adopted,” says McEnally.
Here’s a complication: Projecting the exercise practices of employees involves assumptions based on historical figures. But if companies change the design of their plans — as many are considering, in light of pressure to adopt benchmarks tied more closely to performance — then it may take some doing to convince the SEC and auditors that the use of historical data is appropriate. What’s more, companies that have divested a large segment of their business might find that employees exercise their options much earlier than expected.
As a result, says Susan Eichen of Mercer Consulting, “it’s difficult for auditors to verify the expected life assumption for all the factors.” Rudkin of Analysis Group agrees. “I think there will be a rather steep learning curve,” he says.
If FASB embraces the binomial model, however, companies may have little choice but to climb aboard. The model is already in use at a handful of companies, including American International Group and Washington Mutual Inc.; by publication time, neither could be reached for comment. And at this point, how many others would become early adopters is anyone’s guess.
Yet precedent suggests more than a few would do so. Following earlier far-reaching and controversial changes in accounting rules — including the elimination of goodwill amortization and the new treatment of intangible assets in mergers and acquisitions under FAS 141 and FAS 142 — many companies changed their accounting to conform before they were required to do so.
U.S. companies raising capital abroad already have an incentive to adopt the binomial model; the IASB has effectively endorsed the model in its own rule, already finalized, that requires options to be expensed. In essence, the international standards-setter requires that companies take their cue from what “a knowledgeable willing market participant,” such as a derivatives trader, would use to value the option. The rule also notes, “This may preclude the use of the Black-Scholes model.”
Some experts worry that investors may be confused if two companies in the same sector use different models, because investors are far less likely than CFOs to evaluate the assumptions involved. “An ordinary investor may or may not know how to do that,” says the AIMR’s McEnally, “and that’s the problem.”
And some finance executives remain deeply skeptical that either model works well with employee options. As Donna Blackman, director of accounting policy at Marriott International, puts it: “Whether FASB tells us what to use or says use your own judgment, I think Black-Scholes as well as the binomial method are imperfect.”
That said, at present Marriott values options using Black-Scholes for footnote disclosure purposes; the company will wait for guidance from FASB before deciding whether to switch to the binomial method, if option expensing is required. Yet it’s unclear just how helpful the board will be. Typically, FASB doesn’t recommend one valuation model over another, but Mercer’s Eichen believes that FASB’s exposure draft may be much more prescriptive than usual and will severely limit companies’ ability to use Black-Scholes. The rule-makers, she says, want to avoid letting companies “try both and pick the one that gives them the lowest option cost.” For that reason, she expects that “companies may have to demonstrate a good reason for using Black-Scholes.”
Eichen says the U.S. rule-maker “may be even more specific or even more limiting in its language.” She’s heard more than one FASB board member express strong reservations about Black-Scholes, and she’s seen little to indicate that FASB has adopted the more general, principles-based approach to rulemaking favored by the IASB.
In other words, CFOs had better start boning up on the binomial.
Valuation Studies at 20 Paces
Predictably enough, consultants disagree over the virtues of the binomial method for valuing employee stock options.
On one side is the Boston-based Analysis Group, whose recent study concluded that the Black-Scholes method can overestimate the value of employee options by anywhere from 28 percent to 56 percent. On the other side is Mercer Consulting. Mercer’s study of 350 major companies with broad-based stock-option plans found that in 75 percent of the cases, the two formulas produced a cost differential of less than 5 percent.
Susan Eichen of Mercer goes so far as to say that the Black-Scholes and binomial models “will give you the same results,” provided you plug in the same set of assumptions. Mercer assumed that options were granted at the money; that the exercise price was equal to the stock price at the grant date; that the company paid a moderate to low dividend; and that the exercise occurred at the end of the contractual term (an assumption also made by the current accounting rule, FAS 123).
Yet Eichen acknowledges that most employees do not hold on to their options for the full term. “The approach we’ve taken is simply for comparative purposes,” she says, “to get a sense of the full value of the options.”
Analysis Group, for its part, concedes that the two models can produce the same result for options that are easily traded. “If you take the Black-Scholes model and compare it straight away against the binomial model designed for valuing exchange-traded options, given the same inputs, those two models will produce similar results,” says Ron Rudkin, a vice president of the firm. Rudkin adds, however, that employee option grants are completely different from other options because of their vesting schedules, lack of transferability, and forfeiture requirements, which the Mercer study doesn’t take into account.
What’s more, contends Rudkin, a dividend payment of any size will produce a greater value for employee options under the binomial model than under Black-Scholes. Yet it may be inappropriate to assume that a company with a broad-based option program pays any dividends, since the typical such company is an entrepreneurial enterprise that must reinvest most if not all of its cash.
Who’s right? Hard to say. But the answer may become a bit clearer if, as expected, the FASB throws its weight behind the binomial model. —C.S.