THERE IS NO ACCOUNTING FOR STOCK OPTIONS

If ever there was a bad idea that would not die it is that the grant of stock options should be treated as an expense for accounting purposes. After more than a decade of trying, the FASB succeeded in imposing such a rule by adopting SFAS 123R in 2005, though the rule is still in the process of phase-in and there are many questions about what it requires.

The reformers’ seemingly simple argument is that options are a substitute for cash compensation. But an option may just as easily be viewed as a deal in which an employee works in part for a share of the profits or a partnership in the business. Is that an expense? Hardly.

This is not to say that an option has no value at the time of grant. But the value lies in the possibility that the price of the underlying stock will rise. Even the almighty Internal Revenue Code agrees that the recipient of options has income only when the option is exercised. And no income means no expense. On the other hand, the value of an option can now be calculated with relative precision. The rub is that one of the variables that must be considered is the volatility of the underlying stock. Curiously, options on riskier stocks are more valuable than options on safer stocks, because riskier stocks are more volatile and more likely to jump to the exercise price at some point. Thus, other things being equal, it costs more for a growth company to use options as compensation than it does for an established company to do so.

For example, on March 30, 2005 Intel closed at $27.43, and an option to buy Intel at $30 in January 2006 closed at $2.25 per share, for an aggregate premium of $4.82 per share over the current share price. On the same day, McDonald’s closed at $28.66, and a January 2006 option to buy McDonald’s at $30 closed at $1.85 per share, for an aggregate premium of $3.19 per share. In other words, the Intel option was worth about 50 percent more than the McDonald’s option even though the Intel option was farther out of the money. Why? Because Intel is riskier than McDonald’s. If the two were to make identical option grants, Intel would be required under the proposed rule to book an expense of $1.51 for every dollar booked by McDonald’s. To add insult to injury, options are more important for growth companies that need to conserve cash and whose employees seem to prefer working for a piece of the action. In short, expensing will affect growth companies disproportionately, not only because each option is more valuable, but also because growth companies tend to grant more options.

One of the goals of accounting should be to facilitate comparisons between companies. Treating the grant of options as an expense will do just the opposite. In order to compare the performance of Intel and McDonald’s, an investor will need to adjust for the fact that Intel’s earnings are disproportionately reduced by the recognition of more expense per option granted. Moreover, expensing will lead to a string of bizarre adjustments likely to confuse things. Currently, earnings per share are reported on a fully diluted basis -- as if all outstanding matured options have been exercised. Effectively, options are already treated as an expense because they reduce per share earnings. If we expense them at the time of grant, per share earnings will need to be reported on the basis of actual shares outstanding. And that in turn will mean that investors must fend for themselves to figure out the potential dilution if outstanding options are exercised.

In fairness, one could recognize option expense at the time of exercise and in an amount equal to the simple difference between market price and exercise price (as the tax man does). That would avoid the disproportionate impact of valuation methods and would reflect actual dilution as it happens. But that is not the proposal. And in fairness, recognition of expense at the time of exercise would give option holders a new way to manage earnings. The solution is not accounting rules but rather disclosure (of which there is plenty already). There is no reason to think that the market cannot understand and adjust for the (potential) dilution that comes with the grant of options.

In the end, it might not matter whether a company treats the grant of options as an expense. Studies show that a company’s choice of accounting convention makes no difference as to stock price. As it is, analysts can translate earnings into cash flow, while CFOs can explain away the aberrant effects of accounting rules by calculating pro forma earnings. But does it really make sense to invent yet another way by which the numbers diverge? Moreover, it will be exceedingly difficult to unwind the effects of expensing options. First, expensing options may change management behavior by eliminating the incentive to distribute cash through repurchases. Second, option pricing models are based on an options market composed of diversified investors who can use options for hedging. For the CEO who gets paid in options, they are an all-or-nothing proposition. If your stock goes up, you win. If it goes down, you get squat. It follows that options are worth a whole lot less as compensation than they are as market instruments.

Indeed, management compensation has been declining as a percentage of income as options have become the primary form of compensation. In 1985 officer compensation was more than 70 percent of corporate taxable income in the aggregate, whereas during the five years up to 2000 it averaged about 40 percent of taxable income. Thus, the perception that management compensation is out of control is mostly about the redistribution of pay from losers to winners.

Aside from the difficulties inherent in expensing options, it is not clear that options do have a cost. An option will be exercised only if the value of the underlying stock increases. Who is to say that the rising stock price is not the result of the added incentive created by stock options? Do options have a cost if shareholders also enjoy a gain as a result? Moreover, an option that expires unexercised presumably costs the company nothing even though it would have been booked as an expense under the new order. Will the company recognize income if its stock price remains flat and options are never exercised?

Some say that the problem with options as compensation is not accounting rules as much as options themselves. The argument is that because options are risky too many options must be granted to substitute for cash compensation. But that is the nub of the matter. Options require managers to assume more risk. The trade-off is the prospect of more return. Stockholders, on the other hand, can diversify and avoid company-specific risk. A diversified investor knows that you win some and you lose some but only the average matters.

During the 1950s and 1960s managers sought to smooth out earnings by diversifying at the company level through conglomerate mergers. With the ascendance of mutual funds in the 1970s, stockholders found a cheaper way to diversify. The proof is in the pudding. The market places a lower value on diversified conglomerate companies than it does on more focused companies. Hence the junk-bond-financed bust-up takeovers of the 1980s. Bidders discovered that they were able to sell the pieces of conglomerates for more than the price of the whole, because investors prefer focused and leveraged companies.

Managers learned the lessons of the 1980s well, and stock options became the primary form of compensation -- to the tune of 90 percent of pay. Stock options focus management attention where it belongs -- on increasing the value of the company rather than on second-best indicators like earnings or sales (or oil reserves). It is difficult to believe that so many companies in the 1990s would have been so quick to spin off underperforming divisions if management had not been induced to maximize share price by taking most of its compensation in the form of options.

In addition, options induce granting companies to buy back stock to avoid dilution. The critics cite such maneuvers as part of the problem. Does anyone else perceive a contradiction here? One of the factors that led to the takeovers of the 1980s was the reluctance of management to distribute available cash. Options are a powerful incentive to share the wealth.

Still, it has been suggested that payment in stock would make more sense because it would give management the same kind of stake as an investor. Wrong. Aside from the fact that managers cannot diversify, with an outright grant of stock, management assumes the risk that stock price will fall and not simply fail to increase. With stock, management will have some interest in undertaking conservative strategies designed to maintain stock price. On the other hand, in a bear market, creating incentives to maintain stock price may sound like a pretty good idea. Then again, if one is interested primarily in safety of principal or a reliable return, there is always the bond market. It makes no sense to invest in stock unless one seeks a higher return. So it makes no sense to create incentives for management to pursue a conservative strategy. Again, investor diversification is key. A diversified investor prefers that each individual company maximize return even if it means that a few may go bust. If one is adequately diversified, the winners will usually outperform the losers by more than enough to generate a superior return. But few CEOs would bet the farm on a promising new line of business if it were not for stock options and plenty of them.

In the 1930s, academics decried the separation of ownership from control and the rise of a managerial class with little or nothing invested in the business. Now it seems the complaint is just the opposite. The classical view of the corporation is a business owned by the shareholders with management serving at their pleasure and paid primarily by salary and bonus. But it may be more accurate to think of a corporation as a partnership between management and investors. Only a Luddite would, suggest that we should seek to discourage the use of options in favor of basing compensation on some manipulable number like earnings, particularly when earnings management seems to be a real problem. Indeed, because stock options ultimately depend on the stock market, which is exceedingly difficult to manipulate, they are essentially self-regulating. No rational CEO would want so many options that it would reduce stock price.

The ultimate question is which view best comports with what shareholders want. Do they want managers who see themselves as employees? Or do they want managers who act like partners working for piece of the action? This is not a rhetorical question. The answer may differ company to company. But we should not use accounting rules to stack the deck in favor of one model or the other.

Comments

[gundryggia]

I am less sure than you that we want managers to be long volatility but it's an interesting argument.

A few points look suspicious to me, however:

"Even the almighty Internal Revenue Code agrees that the recipient of options has income only when the option is exercised."

So what? This strikes me as a rhetorical strategy rather than an argument. If Hypothetical Holdings loses or gains a few billion on its stake in Company X, but has not yet sold the stock, isn't that something that investors would want to know about? If I'm investing in Goldman Sachs, do I just need to see quarterly income based on already closed-out trading positions, or am I going to want to see mark-to-market P&L as well?

"In order to compare the performance of Intel and McDonald’s, an investor will need to adjust for the fact that Intel’s earnings are disproportionately reduced by the recognition of more expense per option granted."

Why would an investor want to compare the quantity of options granted, of all things? She cares about the value of the options, and the Intel options are more valuable.

"...what shareholders want. Do they want managers who see themselves as employees? Or do they want managers who act like partners working for piece of the action?"

I agree that the question is ultimately what shareholders want. But how exactly does expensing stock options prevent investors from making that choice? If stock options are a good form of compensation, investors will surely reward companies that use them after they are accurately reported. Won't they?

(Please don't claim it's going to traumatize the market by screwing up analyst EPS models or some such. I am reasonably confident US equity markets can survive the transition.)

[Matt Bodie]

I'm also not sure about the upside-risk argument as to stock options (vs. stock). Do we really want a compensation mechanism that is indifferent as to whether the stock price goes down $0.01 or $100?

[Broc Romanek]

Welcome to the blogosphere! Too bad I disagree with this entry. Many companies treated options like candy since there was no expense involved - and the overhang level became very dangerous. Grant practices became routine for senior managers at many companies, without consideration for whether they could really incentive management any further. This topic has been debated to death and sanity has prevailed.

[geoffrey manne]

For some reason the trackback didn't work. Find my (brief) comment on this post at Truth on the Market at http://www.truthonthemarket.com/2006/01/31/the-costs-of-options-expensing-rules/

[Tom Bozzo]

"Aside from the difficulties inherent in expensing options, it is not clear that options do have a cost."

On the contrary, your own examples clearly demonstrate that the options grants have an opportunity cost at the time of the grant. E.g., an option granted by Intel as part of an employee's compensation package could, in principle, have been sold on the market for $X -- the amount depending on the expiration date and strike price -- instead, with no effect on (potential) dilution. So the option has some ex ante cost even if it ends up going unexercised ex post: the company could have had something ($X) instead of nothing (or, more precisely, $X versus the marginal incentive value of the option -- which, as Broc Romanek notes, is not obviously guaranteed to be greater than $X).

For fun, one can imagine a counterfactual tax system in which options grantees can pay tax on the market value of the option at the time of grant instead of on the profit at the time of exercise, and the tax treatment argument fails, too.

[Dale Wettlaufer]

Paraphrasing Warren Buffett, the second-wealthiest person on Earth, who achieved that distinction though thinking through these things effectively:

If options are not compensation, what are they? If compensation is not an expense, what is it? And if expenses don't belong on the income statement, where in the world should they go?

The idea that options issuance shouldn't be counted as an expense because it reduces EPS misses the point. You can issue equity or an equity equivalent and use that directly or indirectly to pay for something in that same accounting period. What if a corporation issued stock for cash in the first quarter and purchased with that cash in Q2 services it fully consumed in that quarter? That would reduce retained earnings and increase shares outstanding in the fiscal year. What if it purchased a new steel plant with that money and put it into operation that year? The depreciation would be spread out over a number of years, but the depreciation would still be an expense and the shares outstanding would still be higher. Why is it any different with labor? I would argue it's no different.

The corporation (1) issues stock or an equity equivanent, (2) which increases equity, (3) to purchase something, which becomes (4) an expense and (5) reduces retained earnings and increases (6) shares outstanding. Whether it reduces or increases EPS depends on whether management bought something of value, whether it's a steel mill, your labor, or services.

Presumably, the corporation is receiving something of value that generates more revenues over the course of the consumption of that good or service than the expense in question. If so, then yes, it is increasing expenses, but the higher expenses are also going to increase earnings. Bottom line, there's no reason why something can't be an equity component and an expense in the same accounting period.

[Geoffrey]

Could you help me. Well done is better than well said.
I am from Italy and learning to write in English, tell me right I wrote the following sentence: "Authorsfor someoto gain weight, athletes need to consume more calories than they expend."

With respect 8-), Geoffrey.