Kip Hagopian

LETTERS TO THE EDITOR

The Point of View position paper in California Management Review’s Summer 2006 issue by Kip Hagopian [“Expensing Employee Stock Options is Improper Accounting,” Vol. 46, No. 3] has generated considerable media attention as well as public debate. We are pleased to publish four comments that we have received in response to Hagopian’s essay, along with his reply to them. We plan to publish additional letters on this important, controversial issue as space permits.

TO: The Editor of the California Management Review

FROM: George J. Staubus, Michael Chetkovich Professor Emeritus of Accounting, Haas School of Business, University of California, Berkeley

I admire the creativity in Mr. Hagopian’s arguments in support of his conclusion, but I think CMR readers should see another approach to the issue of proper accounting for employee services acquired in exchange for the issuance of stock options. The reasoning here is roughly consistent with that of the FASB in the “Basis for Conclusions” supporting its Statement of Financial Accounting Standards 123R, December 2004, but is derived independently based on generally accepted accounting theory and generally accepted accounting principles (GAAP).

Premises

▪ The objective of financial reporting is to provide information to investors and others that is useful in making investment and other decisions.

▪ To support this objective, financial reporting seeks to report on the success of an enterprise’s operations in achieving shareholders’ wealth-enhancing objective.

▪ The income statement—especially its “bottom line,” net income—is relied upon as the most valuable portion of a financial report useful to investors in evaluating the success of operations. The balance sheet is less seriously affected by the alternative methods of accounting for employee services acquired by the issuance of stock options.

▪ The measurement of the success of operations requires the measurement of the cost of using all resources consumed in the revenue-producing operations of the reporting period, except the cost of capital related to the equity interest.

▪ The cost of using resources is not always the cost of acquiring them and is not always run through an asset account (such as an inventory, prepayment, or property account) before being expensed. That is, acquisition of resources and their consumption may, for accounting convenience, be recorded as one transaction, as in the cases of officers’ services, professional services, and many other services.

▪ Stock options and warrants are valuable derivative securities. Existing shareholders bear an economic sacrifice when their enterprise issues additional valuable securities (dilution—i.e., claims to earnings and liquidation value).

▪ In GAAP, the issuance of securities for noncash assets or services is recorded at the fair value of the consideration received or the fair value of the securities issued, whichever is more readily determinable.

▪ Accountants and users of financial reports accept many imperfect measurements that are deemed usefully accurate, such as estimates of uncollectible accounts, depreciation, and pension costs and other post-employment benefit costs; in fact, all accounting measurements are imperfect.

▪ Any and all resources needed for an enterprise’s operations may be acquired in exchange for securities, including “supplier” stock options.

Accounting for the Acquisition of Services by Issuing Stock Options: An Example

Given:

▪ baseline earnings per share (EPS): $100 income ÷ 100 shares = $1

▪ market price of shares, $10

▪ current transaction: acquisition for $2 of a currently consumed resource that yields $2 of revenue—(a break-even operation)

Case A: $2 of a resource is acquired for cash.

Income Effects: Rev +$2, Exp +$2, Inc +$0.

Resulting EPS: $100 ÷ 100 = $1.

Case B: $2 of a resource is acquired for issuance of an option to buy one share of stock, option is worth $2, resource cost not recorded.

Income Effects: Rev +$2, Exp +$0, Inc +$2.

EPS: $102 income ÷ 101 shares = $1.0099 (after an incremental breakeven transaction?).

This is, I believe, the Hagopian solution.

Case C: $2 of a resource is acquired for issuance of option as above, resource cost recorded as expense.

Income Effects: Rev +$2, Exp +$2, Inc +$0.

EPS = $100 ÷ 101 = $0.99 (because proceeds of newly issued share earned nothing; dilution occurred).

Query: Which accounting—that in B or that in C—better informs shareholders of the success of operations?

Conclusion

Recognizing in the accounts the costs of using the services acquired in exchange for the issuance of a security, such as stock or stock options, yields financial statements more consistent with the objectives of financial reporting and with existing GAAP than does ignoring those costs.

TO: The Editor of the California Management Review

FROM: Benjamin E. Hermalin, Thomas & Alison Schneider Distinguished Professor in Finance, Walter A. Haas School of Business, and Professor and Chair, Department of Economics, University of California, Berkeley

I have been invited by the editor of the California Management Review to comment on the recent opinion piece by Kip Hagopian. As I have limited knowledge of accounting, it is not for me to opine on what is or isn’t good accounting practice as defined by accepted accounting principles. I can, however, offer opinions on whether the expensing of employee stock options is an important issue and how valid Mr. Hagopian’s arguments are. Basically, I suspect the expensing issue is, to a great extent, a tempest in a teapot. I further found some of Mr. Hagopian’s arguments lacking with regard to basic economics.

Tempest in a Teapot

While debates about interpretation have inspired people to do war before, these are typically debates about biblical interpretation, not about the correct view on accounting rules. Wars about accounting are motivated by perceived financial stakes, not convictions about accounting principles.

Broadly, the pro-expensing side can be characterized as concerned that investors are being mislead about the true costs of employee stock options and are, thus, led to pay too much for the stock of firms utilizing ESOs. The anti-expensing side can be characterized as concerned that investors don’t grasp the benefits of ESOs and will respond to seeing them expensed by undervaluing the stock of firms that utilize ESOs. In short, one thing both sides of the debate seem to agree on is that shareholders are stupid, insofar as they can’t properly interpret information disseminated by certain means; and, hence, each side sees its position as benevolent paternalism.

The evidence, however, suggests that both sides are wrong in adhering to the stupid-shareholder hypothesis. First, the hypothesis is at odds with the efficient-markets hypothesis—all public information becomes reflected in the stock price quickly—and the scores of empirical studies in finance that have supported it. That is, it doesn’t matter how the information is presented to the market (as long as it’s presented), the market will correctly incorporate it into the share price. In fact, the empirical evidence with respect to employee stock options has generally been consistent with the efficient-markets hypothesis. A number of studies find that market incorporates the economic costs of ESOs even if they are not expensed. Consistent with those findings, there is evidence that the stock price doesn’t react when firms voluntarily begin to expense options. In summary, then, the impact of expensing ESOs is likely to be de minimis and the battle over whether to do so a tempest in a teapot.

Gain-Sharing Instruments

The main argument that Mr. Hagopian offers against expensing stock options is that they are “gain-sharing” instruments and that: (1) consistent with accounting treatment of other gain-sharing instruments such as sales commissions and contingent legal fees, no expense is occurred until payment is made; and (2) the cost of the gain-sharing instrument must be located on the books of the entity sharing the gain. Let’s consider both claims fully.

A principal difference between a stock option and a sales commission, say, is that the stock option is paid (granted) before the employee performs, whereas a sales commission is paid after. Surely, a stock option must have value, otherwise an employee would hardly wish to accept it as compensation, and upon receiving the option the employee is wealthier than she was before receiving it. So if we accept as a principle that an expense is occurred when payment is made, then it seems hard to argue that some expense has not been incurred the moment the option changes hands.

Certainly, it is not free to a firm to grant an option if only because it could have sold the same or similar financial instruments on the market. To see this, consider a firm with four shares (I’m using simple numbers to keep the math easy; if you want the analysis to be more realistic seeming, feel free to multiple the appropriate numbers by, say, a million). With probability 1/2 it will be worth $90 at the end of the relevant period and with probability 1/2 it will be worth nothing. The expected value of the firm is $45 and so each share is currently worth $11.25.

Suppose the firm announces it will create an option to buy one share with a strike price of $10. Observe that in the good state, the option would be exercised and the firm would be worth $100 ($90 plus the strike price). Divided five ways, each share would pay off $20. Since the probability of this happening is 1/2, the value of a share will be worth $10 after this announcement. Observe the option is being offered just at the money.

What is the value of this option? With probability 1/2, the option owner can exchange $10 to get something worth $20. With probability 1/2, the option owner will forgo her “opportunity” to exchange $10 for something worth $0. So profit in the good state is $10, profit in the bad state is $0, and, thus, expected profit and the value of the option is $5.

Observe that the firm could sell this option and distribute the $5 as a dividend among the existing shareholders. In this case, the initial shareholders would get, in expectation, $45 ($5 in dividends plus an expected payout of $40); hence, the value of a firm that sold such an option would be unchanged. On the other hand, if the firm grants this option to an employee the firm is worth $5 less; no different than if it gave the employee merchandise that it could have sold for $5 or even just handed the employee a five-dollar bill. The firm has incurred an expense the moment it hands over the option, just as would had it handed over the merchandise or the five-dollar bill.

Now I can imagine Mr. Hagopian’s disputing this example given that he argues in his article that the different structure of ESOs (vesting periods, non-transferability, etc.) as compared to conventional options means that there is no such opportunity cost incurred at the time of granting them. This, however, is wrong on a number of grounds.

First, nothing in the above example relied on transferability or any other such property of the options. Second, nothing would prevent a firm from selling options that have the structure of ESOs. Third, by issuing ESOs, the firm is reducing the price for other financial instruments it could sell.

To see the last point, suppose the firm in the example had wished to sell a share of stock. Absent the ESO, the share would have sold for $9. With the ESO, it would sell for $8.33.

Perhaps the best way to see why Mr. Hagopian’s “different terms imply no opportunity cost” argument is wrong is to consider a different scenario. Suppose a company was prohibited by law from leasing cars to the public, but could let its employees use its cars. Consider a car that is currently worth $40,000 and will be worth $35,000 after being driven a year by an employee. Giving the right to drive the car for a year to an employee is not costless; even though the company couldn’t lease the car for a year to the public (and perhaps reap large profits from doing so), it costs the company at least $5000, the depreciation in the car’s value plus the forgone interest (investment returns) the $40,000 would have earned the firm. In short, precisely because the firm is allocating currently valuable claims on its future profits when it makes an option grant, it incurs an opportunity cost precisely at the moment it makes that grant.

The second part of Mr. Hagopian’s “gain-sharing instrument” argument is that the cost of the instrument must be located on the books of those sharing the gains (i.e., in this case, the shareholders). If this position is taken to its logical end, then one could argue that a great many payments made by the firm need not be counted as expenses.

For instance, consider a hypothetical firm with one employee. Suppose that the employee is paid $20,000 at the end of each month and during each month the firm’s revenues are $30,000. For simplicity, assume the firm incurs no other costs than the employee’s salary and that all profits are paid out as monthly dividends (nothing about the logic of my argument depends on these simplifying assumptions; they’re made just for the sake of brevity and readability).

Observe that, at the beginning of the month, the shareholders expect to gain $30,000, but have agreed to share 2/3 of that amount with their employee. The salary is a form of gain-sharing instrument. It also has incentive effects insofar as the employee would not show up to work if not promised the salary and the firm would, thus, make nothing.

Note, as long as it is public information what the promised salary is, it will be reflected in the stock price—at the beginning of each month, the value of the outstanding shares will equal the present value of a stream of monthly payments of $10,000. In this sense, the cost of the gain-sharing (i.e., the fact that the shareholders have agreed to part with 2/3 of the gain) is reflected on their books; that is, in the price of the stock. If one followed Mr. Hagopian’s logic, then the salary should not be considered an expense of the firm. I doubt, however, that people are prepared to argue that employee salaries aren’t an expense of the firm.

Now I suppose one might feel that, because the salary is a fixed amount, it is somehow not a gain-sharing instrument (although, it should be noted, nothing in Mr. Hagopian’s definition of a gain-sharing instrument required it to be risky). Of course, a salary can be risky—the firm could fail to have enough cash on hand to pay it fully.

For instance, suppose that, on average, four out of five months the firm’s revenues are $35,000 and that, one out of five, they are $10,000 (note expected monthly revenues are still $30,000). Suppose the employee’s salary is $22,500 per month. Of course, given that all earnings are paid as dividends, in any month the firm earns only $10,000, she receives $10,000 (employees have priority for payment over shareholders). Her expected monthly salary is still $20,000. Value of the outstanding share at the beginning of each month remains the present discounted value of $10,000 monthly and, thus, continues to fully reflect the cost of having to share the gain.

I have used salary as an example, but it should be clear that the same argument would go through for a variety of other payments such as debt repayment (e.g., suppose each month that, instead of hiring an employee, the firm had to borrow $20,000 in order to operate). Although I am not an accountant, it strikes me that it would be a radical change in accounting practice were we no longer to consider salaries, debt repayment, and the like expenses of the firm.

Should Employee Stock Options Be Expensed?

Given that I believe that the practical consequences of expensing to be small, there isn’t really a right or wrong answer to the question of should ESOs be expensed. Provided sufficient information is disclosed, the market will properly determine the value of the firm even if they aren’t expensed.

While there might not be a right or wrong answer, there are certainly right or wrong arguments. Although the structure of ESOs differs from other forms of compensation, there is nothing so magical about them that paying (issuing) them to employees doesn’t represent a cost to the firm. In particular, the fact that they are gain-sharing instruments does not distinguish them from other payments that the firm makes.

As a believer in markets, I think a better line of argumentation is the following. Assuming adequate disclosure of option plans and assuming that firms are free to expense if they wish, then leave it for the market to decide. Given that I don’t think expensing matters if information is fully disclosed, I think the market will “shrug its shoulders” and firms will feel neither market pressure to expense or not to expense. But, if there is a market preference, then it will reveal itself in one set of firms trading at a premium over the other set ceteris paribus. If firms that expense are trading at a premium, then non-expensing firms will be under market pressure to expense. If firms that don’t expense are trading at a premium, then expensing firms will be under market pressure to cease expensing. Given that markets are generally seen as better at answering these questions than regulators, this is likely to be the optimal way of resolving the debate.

TO: The Editor of the California Management Review

FROM: George J. Benston, the John H. Harland Professor of Finance, Accounting, and Economics at the Goizueta Business School, Emory University

Kip Hagopian argues that employee stock options (ESOs) should not be reported as an accounting expense in an enterprise’s financial statements at any time for two reasons. First, an ESO is a “gain-sharing instrument, [which] by its nature, has no accounting cost unless and until there is a gain to be shared” and “[t]he cost of a gain-sharing instrument must be located on the books of the party that reaps the gain [emphasis in original].” Since the gain is obtained by the stockholders, he says, the cost of granting ESOs are not an expense of the enterprise. Second, “[n]either the granting nor the vesting of an ESO meets the standard definition of an expense.”

Others also argue that it is sufficient to disclose the terms of ESOs, since accountants’ definition of “net income” and its components and the numbers presented in the balance sheet are essentially irrelevant so long as the accounts comprising these statements are described and the numbers reported are audited by independent public accountings.

I consider the last argument first, since, if it is accepted, the first two arguments are moot. Based on the fact that accounting reports (which include at least an income statement and balance sheet that are audited and attested to by independent public accountants) have been presented to investors, creditors and others long before they were legally required, it is evident that such reports have been found useful, if for no other reason than that they offer an efficient starting place for estimating the economic value of enterprises and for evaluating the performance of the enterprise and its managers.

Consequently, if the cost to a corporation of ESOs represent an accounting “expense,” as this term has been used and understood by users of financial statements, they and the reporting enterprise should benefit from accountants recording ESOs much as they record in the financial statements the impact of other similar economic events that affect the enterprise.

ESOs as “Gain-Sharing” Instruments

Consider now Hagopian’s first argument, that because ESOs are “gain-sharing instruments” that transfer resources from one group of stockholders (employees to whom the options are granted) to another (other stockholders), ESOs should not be recorded on a corporation’s books.

It should be noted, first, that accounting provides a record of economic events that affect an enterprise, rather than its owners, creditors, employees, or other related parties. If several stockholders (whether they also are employees of the corporation or not) contract to share their gains from ownership of their shares in a corporation, the effects of the contract is of no interest to the corporation’s accountants, since the agreement does not affect the corporation. An exception would be if the agreement represented a conflict of interest that worked against the interests of other stockholders. This only requires disclosure as a matter of law and corporate governance, not a difference in accounting for the corporation.

Hagopian, though, argues that gain-sharing instruments do not affect the corporation that employs them; they only affect the stockholders. He first provides three examples: profit-sharing, sales commissions, and a contingent-fee arrangement with a plaintiff law firm. These examples, though, are irrelevant to his argument, since, as he correctly states, the effect of these agreements, when they are consummated, are recorded on the books of the enterprise.

The expenses related to the first two are not recorded because, at the time the agreements were entered into, no services were provided. When the services have been provided, as indicated by the profits and revenues that are shared, the payments to employees are recorded as expenses of the enterprise. Note that the benefits from profit-sharing and sales commissions are “shared” by stockholders with the employees. As with ESOs, the gain, net of expenses including payments to employees, is “reaped” by stockholders in the form of appreciation in the value of their stocks. The related expense and the revenue, though, are first recorded in the enterprises’ income statements.

With respect to the legal agreement, although the enterprise has received something of value, it does not record a net profit because of the traditional accounting practice of “conservatism” (discussed further below). Gains (but not expenses) from any activity are not recorded as such until they have occurred and can be estimated reasonably and objectively. But when this occurs, the gain is recorded on the books of the enterprise. These gain-sharing agreements, then, result in changes recorded on an enterprise’s books and, hence, its financial statements. The issue, then, is when the cost to a corporation of ESOs should be recorded and in what amount.

The Accounting Definition of Income and Expense

Economists, investors, and others might want accountants to report the economic (market) value of total and individual assets and liabilities as of the beginning and end of a period. Net income, then, would be the end-of-period assets less liabilities (net assets) less beginning-of-period net assets plus dividends declared and less new investments by equity holders.

But, this (Hicksian) definition of net income can rarely be measured unless the firm is established at the beginning and liquidated at the end of the period. If the firm is ongoing, many of its assets and liabilities cannot be measured objectively, since their economic values necessarily involve subjectively determined estimates of value in use (e.g., net cash flows, applicable discount rates, and externalities). Hence, accounting has traditionally applied a somewhat different measure of net income and has used the balance sheet as a bridge between income statements rather than as a statement of economic values.

Accountants first recognize income as it is earned. In general, this occurs when, in exchange for goods or services rendered, title to assets received passes to the enterprise or the enterprise’s liabilities are extinguished. The amount recorded as revenue depends on the value of the asset received or liability extinguished.

Usually, these amounts can be objectively determined. For example, goods may be sold in exchange for a promise to pay (Account Receivable). The amount recorded as a sale could be the present value of the Account Receivable (factoring in the probability that some accounts may not be paid in full). Or, the sale amount could be (and usually would be) recorded as the gross amount promised and the estimated non-payment at the end of the period would recorded as “Bad Debt Expense.”

If the value to the enterprise of the asset received in exchange cannot be determined objectively, the amount of the sale would be determined by the amounts received for sales of the same or similar products or the amounts for which similar assets were sold, whichever is the most reliable number. Income also is recognized when there is objective evidence that the economic value of financial assets have increased, particularly as evidenced by the market value of securities traded in a recognized market.

Expenses generally are reductions in the claims of equity holders (including debtors if the corporation is bankrupt) over the assets of the enterprise as a result of the operations of the enterprise. The basic method of measuring net income is called “matching,” wherein costs that were incurred (assets expended or liabilities assumed) to achieve the earned income are charged (matched) against the income and recorded as expenses.

For those assets that decline in value over time or with use, but whose reduction in value cannot be objectively measured (such as depreciation of buildings and equipment), a pre-determined method of allocating the cost of those assets to specific assets is employed (e.g., straight-line depreciation). When, during a period, the economic values of non-financial assets have decreased, the decreases also are recognized as expenses. But similar increases are not recognized until validated by market transactions.

Accountants have learned that entrepreneurs and managers tend to be overly optimistic or opportunistic and that investors and creditors dislike unpleasant surprises, although they often do not object to pleasant surprises. Consequently, accountants have adopted a current-period conservative bias. That is, revenue is not recorded as earned until there is objective evidence that the amounts promised have been or are expected to be realized, and expenses are recorded even when they are difficult to measure.

Matching also involves deferring some expenses. When there is objective evidence that revenue that generated expended costs will be forthcoming and recorded in a future period, expenses that do not exceed the expected revenue are deferred and are carried as assets on the balance sheet. (Alternatively, deferred expenses may be seen as assets that represent the present value of expected future revenue.)

ESOs as Expenses

ESOs clearly are payments to employees in exchange for their present and future services. In concept, they are no different than other forms of compensation, including salary and bonuses paid in cash or in kind, deferred salary, retirement payments, and medical and other benefits. (Note several forms of compensation in additional to ESOs, such as bonuses, are designed not only to reward employees for their current-period services, but to motivate them to expend productive efforts in the future.)

In effect, accountants make the following entries with respect to employee compensation. When the employee has earned compensation, the accountant debits an asset, Employee Services, and credits Liability to Employees. As and when an asset is used up or expires because it has no future value, an expense, Employee Expense is debited and the asset, Employee Services, is credited. (In practice, costs that do not result in the production of tangible assets, such an inventory, usually are recorded directly as expenses.)

How the liability to the employee is met is important only in that it determines the cost to the enterprise of the employee’s services. The liability could be met with cash, with a promise of future payment (such as a pension or deferred compensation), with tangible goods and services (such as inventory or the right to use a car or occupy a house), or with financial goods (such as stock or stock options).

The value to the employee of the payment is important only in that the more valuable the payment of a particular form to the employee, the less the employee has to be paid to secure his/her services. The essential issues for the accountant are: (1) what is the economic value of assets to the enterprise that is given to employees in exchange for their services, and (2) how much of this amount is allocable as an expense to the current period and how much to future periods?

ESOs clearly are valuable assets. If not granted to employees, close substitutes could be sold to investors. The fact that ESOs have restrictions not found in ordinary stock options makes them less valuable and, hence, less costly to stockholders when they are granted to employees rather than sold. But, the cost to investors of issuing them to employees is not zero (which is implicitly assumed when ESOs are not recorded as expenses).

Many expenses are estimates of their cost to stockholders, such as the cost of warranties, pensions, health insurance for retirees, and depreciation. Many of these, such as unvested pensions and health insurance, do not require a strict liability to make a future payment for their cost to stockholders to be estimated. In this regard ESOs are no different.

Indeed, the economic value to stockholders of ESOs (and, hence, their cost when they are issued to employees) could be estimated more effectively than many other items. For example, ESOs with terms similar to those given to employees (such as delayed vesting) could be offered to investors, thereby establishing their value. Or, such ESOs could be distributed to stockholders as dividends; the after-issue market price then would establish the value. Or, an investment banker or finance professor might provide an estimate of the value from experience or from the output of a variant of the Black-Scholes option-pricing model or other model, with a suitable discount for the restrictiveness of ESOs.

One way to give corporate executives an incentive to correctly value ESOs would be to amend the Internal Revenue Code to allow a deductible expense for ESOs only for the amount recorded as an expense on the taxpaying corporations’ financial statements. Thus, if they understated the expense corporate taxes would be higher and their stock options worth less. If they overstated the expense, taxes would be lower, but they would have to inform shareholders of the amounts paid for their and other employees’ services.

Given that an estimate of the cost to stockholders of ESOs can be and is estimated, the next issue is how much should be allocated to the period in which they are granted and how much to future periods.

Considering that an important benefit to the enterprise of ESOs is their incentive to employees to increase the net worth of stockholders, it would seem that, in accordance with “matching,” a portion of the cost should be allocated to future periods. However, traditionally, accountants have generally recorded expenditures for similar intangible assets as current period expenses, even though these expenditures are incurred to and are likely to generate future revenue.

For example, expenditures on advertising and business development often are incurred to generate higher income in future periods. Accountants, though, do not attempt to allocate these expenditures to those periods, essentially because the amount of the future benefits can rarely be estimated reliably, and managers who want to report lower current-period expenses are likely to overestimate the benefits. This is another manifestation of the conservative bias in accounting.

I conclude, therefore, that the value of ESOs should be recorded as an expense in the period in which they are given in exchange for employees’ services. Although this does not result in a completely accurate report of the economic situation and performance of a corporation in a particular period, it is consistent with traditional accounting practice.

As noted earlier, accounting financial statements do not (and I believe, neither should nor can) represent completely the value of an enterprise to investors and changes in that value over a period. These reports, though, are more useful to investors when they consistently follow an understood set of rules, known as Generally Accepted Accounting Practices (GAAP). Consistent with GAAP for other forms of employee compensation and expenditures that enhance the value of other intangible assets, the value to the corporation of ESOs should be recorded as expenses in the period in which they are granted.

TO: The Editor of the California Management Review

FROM: Joshua Ronen, Research Professor, Stern School of Business, New York University

On December 16, 2004, the Financial Accounting Standards Board (FASB) published FASB Statement No. 123 (revised 2004), which significantly changes the accounting for employee stock options. The standard it replaces had rekindled the debate over stock option accounting that, a decade ago, preceded the issuance of FASB Statement No. 123, Accounting for Stock-Based Compensation (“FAS 123”). The debate pivoted on the requirement to expense stock options and concerns about the reliability of the measurement methodology. Before it issued the revision, the FASB’s proposed rules came under tremendous scrutiny from Congress, the media, and preparers and users of financial statements.

Under the new standard, equity-based compensation (EBC) results in a cost to the issuing enterprise and should be measured at its fair value on the grant date, based on the estimated number of awards that are expected to vest. The compensation cost for equity awards that vest will not be reversed if such awards expire without being exercised.

The recent article by Kip Hagopian suggests that expensing employee stock options is inappropriate. I agree with Hagopian’s arguments but extend the discussion in this Note to cover all EBC instruments. Specifically, I contend that the proper treatment would reflect the value of share-based payments as an expense of the pre-existing shareholders (the shareholders of the firm before the granted options are exercised) rather than as an expense of the corporation.

Currently, under SFAS 128, diluted earnings per share are provided in the income statement, but share-based payments are not usually expensed. This diluted earnings per share computation does not fully reflect the effects of dilution on earnings, nor is the effect of dilution on the book value captured.

I suggest that it would be better to reflect the dilution cost as an expense in a separate statement, such as one labeled “Statement of Costs and Benefits to Pre-Existing Shareholders.” This will eliminate the need to present diluted earnings per share.

A combined “Statement of Enterprise Income” that sums the respective amounts in the corporation’s income statement and this new statement would reflect the total costs and benefits of engaging in the particular enterprise’s operations from the perspectives of both pre-existing and new shareholders combined, while at the same time showing the wealth transfer between the two groups.

In this Note, I discuss a benchmark example that features three consecutive annual grants of restricted stock that vests before the end of each of the three years, where the grant is the sole means of compensating the employees and incentivizing them to exert effort. An asset is manufactured and sold in the first year, whereupon all proceeds are invested along with employee labor to manufacture and sell another asset in the second year; the same process is repeated in the third year. At the end of the third year, the proceeds of the sale are distributed to the shareholders of record at that time. To simplify the analysis and the presentations, I assume, with no loss of insight, that no taxes exist.

In the full working paper version of this Note (available online), I move on from this benchmark case to introduce debt as well as options and show how my proposed multiple statements better inform creditors and shareholders about the gains and losses accruing on their respective investments. In evaluating the appropriateness of the accounting treatments that I employ in the separate statements, I rely on the criteria used by Kirschenheiter et al.

Restricted Stock: The Benchmark Case

The Table below presents a benchmark case. Employees are granted 10 shares valued at $10 per share, for a total compensation of $100 every year over the three-year horizon of our hypothetical company. These are restricted shares that vest one instant before the end of each year and mature at the end of the three years. Thus, we have three cohorts of employees (who could be the same individuals) that receive each year $100 worth of stock. As a result, they share in the company’s profits in the proportion of their ownership.

These share grants are equivalent to options with zero exercise price that vest just one instant before each year’s end. A $100 initial capital contribution by the owners (10 shares times $10 each) is invested, along with the labor inputs of the employees (valued at $100 and paid for by the stock grant) to yield a certain, inter-temporally constant 10% rate of return.

Thus, the capital and labor inputs of $200 are converted into an asset worth $220 at the end of the first year. The asset is then sold, and the proceeds are reinvested, along with a $100 labor input in the second year, to produce an asset worth $352 at the end of the second year. Finally, in the third and last year, the $352 are reinvested with another $100 of labor to produce an asset worth $452 X 1.10 = $497. This asset is sold at the end of the third year, the company is dissolved, and shareholders (including the employees) receive a liquidating dividend of $497.

The pre-existing shareholders are rationally willing to suffer the dilution in their ownership and in earnings because in exchange they anticipate a value enhancement that exceeds their cost of dilution, driven by the more highly motivated workforce being incentivized through ownership rather than cash.

Consider, for example, year 1. Beginning with a value of $10 per share, a pre-existing shareholder suffers a dilution of five dollars (50% dilution), but he also suffers the dilution of the portion of earnings saved by issuing shares instead of paying cash (50% of $100 divided by 10 shares) plus the dilution of earnings (50% of $20 divided by 10 shares). The total cost of dilution is thus $11 per share, leaving each pre-existing shareholder with a net profit of one dollar, which constitutes a 10% return on his investment of $10.

By the same token, each employee will have invested $10 worth of services and also earned a 10% return; his gross income consists of his 50% share in the initial asset worth $10 a share (which he earns as a result of the dilution of the original shareholdings) plus his 50% share in the corporation’s profit of $120 (the sale of the asset at $220 less its original cost of $100).

TABLE — YEAR 1

Statement of Enterprise Income

Sale of Productive Asset
Pre-Existing Shareholders: $220.00
Total: $220.00

Cost of Goods Sold
Pre-Existing Shareholders: –$100.00
Total: –$100.00

Dilution Cost — Pre-Existing Shareholders
Dilution of Initial Stock Value (50% of $100): $50.00
Dilution of Earnings — portion saved by issuing shares instead of paying cash (50% of $100): $50.00
Dilution of Earnings — portion of return on labor input (50% of $20): $10.00
Total Dilution Cost: –$110.00

Dilution Benefit — New Shareholders: $110.00
Total: $0.00

Cost of Employee Services (credited to equity)
New Shareholders: –$100.00
Total: –$100.00

Net Income
Pre-Existing Shareholders: $10.00
New Shareholders: $10.00
Total: $20.00

Per Share
Pre-Existing Shareholders: $1.00
New Shareholders: $1.00
Total: $1.00

ROE or EPS Growth: 0.10 / 0.10 / 0.10

Forward Price Earnings Ratio: 10.00 / 10.00 / 10.00

Expected Net Dividend
Pre-Existing Shareholders: $133.10
New Shareholders: $133.10
Total: $266.20

Equity Market Value
Pre-Existing Shareholders: $110.00
New Shareholders: $110.00
Total: $220.00

Present Value of Expected Net Dividends
Pre-Existing Shareholders: $110.00
New Shareholders: $110.00
Total: $220.00

Restricted Shares Issued to Employees with No Consideration: 10.00
Shares Issued for Cash: 10.00
Total Number of Shares Issued: 20.00
Price Per Ending Share: $11.00
Price at Which Initial Shares Were Issued: $10.00

Corporation’s Income Statement

Sale of Productive Asset: $220.00
Less: Cost of Asset Sold: –$100.00
Corporation’s Net Income: $120.00

Statement of Costs and Benefits to Pre-Existing Shareholders

Dilution of Initial Stock Value (50% of $100): $50.00
Dilution of Earnings — portion saved by issuing shares instead of paying cash (50% of $100): $50.00
Dilution of Earnings — portion of return on labor input (50% of $20): $10.00
Total: –$110.00

Balance Sheet

Cash: $220.00
Common Stock: $200.00
Retained Earnings: $20.00

TABLE — YEAR 2

Statement of Enterprise Income

Sale of Productive Asset: $352.00
Cost of Goods Sold: –$220.00

Dilution Cost — Pre-Existing Shareholders

Dilution of Initial Stock Value (9.09/29.09 of $220): $68.75
Dilution of Earnings — portion saved by issuing shares instead of paying cash (9.09/29.09 of $100): $31.25
Dilution of Earnings — portion of return on labor input (9.09/29.09 of $32): $10.00
Total Dilution Cost: –$109.99

Dilution Benefit — New Shareholders: $109.99

Cost of Employee Services (credited to equity): –$100.00

Net Income
Pre-Existing Shareholders: $22.01
New Shareholders: $9.99
Total: $32.00

Per Share: $1.10 / $1.10 / $1.10

ROE or EPS Growth: 0.10 / 0.10 / 0.10

Forward Price Earnings Ratio: 10.00

Expected Net Dividend
Pre-Existing Shareholders: $266.20
New Shareholders: $121.00
Total: $387.20

Equity Market Value
Pre-Existing Shareholders: $242.00
New Shareholders: $110.00
Total: $352.00

Present Value of Expected Net Dividends
Pre-Existing Shareholders: $242.00
New Shareholders: $110.00
Total: $352.00

Restricted Shares Issued to Employees with No Consideration: 9.09
Total Number of Shares Issued: 29.09
Price Per Ending Share: $12.10

Corporation’s Income Statement

Sale of Productive Asset: $352.00
Less: Cost of Asset Sold: –$220.00
Corporation’s Net Income: $132.00

Statement of Costs and Benefits to Pre-Existing Shareholders

Dilution of Initial Stock Value (9.09/29.09 of $220): $68.75
Dilution of Earnings — portion saved by issuing shares instead of paying cash (9.09/29.09 of $100): $31.25
Dilution of Earnings — portion of return on labor input (9.09/29.09 of $32): $10.00
Total: –$109.99

Balance Sheet

Cash: $352.00
Common Stock: $300.00
Retained Earnings: $52.00
Total Equities: $352.00

TABLE — YEAR 3

Statement of Enterprise Income

Sale of Productive Asset: $497.20
Cost of Goods Sold: –$352.00

Dilution Cost — Pre-Existing Shareholders

Dilution of Initial Stock Value (8.26/37.36 of $352): $77.88
Dilution of Earnings — portion saved by issuing shares instead of paying cash (8.26/37.36 of $100): $22.12
Dilution of Earnings — portion of return on labor input (8.26/37.36 of $45): $10.00
Total Dilution Cost: –$110.00

Dilution Benefit — New Shareholders: $110.00

Cost of Employee Services (credited to equity): –$100.00

Net Income
Pre-Existing Shareholders: $35.20
New Shareholders: $10.00
Total: $45.20

Per Share: $1.21 / $1.21 / $1.21

ROE or EPS Growth: 0.10 / 0.10 / 0.10

Equity Market Value
Pre-Existing Shareholders: $387.20
New Shareholders: $110.00
Total: $497.20

Restricted Shares Issued to Employees with No Consideration: 8.26
Total Number of Shares Issued: 37.36
Price Per Ending Share: $13.31

Corporation’s Income Statement

Sale of Productive Asset: $497.20
Less: Cost of Asset Sold: –$352.00
Corporation’s Net Income: $145.20

Statement of Costs and Benefits to Pre-Existing Shareholders

Dilution of Initial Stock Value (8.26/37.36 of $352): $77.88
Dilution of Earnings — portion saved by issuing shares instead of paying cash (8.26/37.36 of $100): $22.12
Dilution of Earnings — portion of return on labor input (8.26/37.36 of $45): $10.00
Total: –$110.00

The above analysis pertaining to the representative shareholder is reflected explicitly and in the aggregate for each of the two groups of shareholders in the Statement of Enterprise Income. This statement has three columns: Pre-Existing Shareholders, New Shareholders, and Total.

The Pre-Existing Shareholders column reflects the revenue, cost of goods sold, and the cost of dilution. The latter is decomposed into its components: the dilution of initial stock value of $50, the dilution of the portion of earnings saved by issuing shares instead of paying cash of $50, and, finally, the dilution of earnings (return on labor input), which is 50% of $20. The total cost of dilution is thus $110, leaving a net income to Pre-Existing Shareholders of $10.

The New Shareholders column reflects the benefits to employees who become new shareholders at the end of the year. These benefits consist of the end of year value of the shares granted to them ($110 transfer of wealth from the pre-existing shareholders who suffer exactly that same amount in dilution), less the cost to them of $100 worth of labor input (effort). They, too, end up with a net income of $10.

The Total column shows the revenues and expenses of the enterprise as a whole, and in this benchmark case (and in the benchmark case with debt analyzed in Exhibit 2 of the working paper version, but not in the cases wherein the stock or options do not vest in the same year in which they are granted as in Exhibits 3–5 in the working paper version) is identical to the statement required under FAS 123. From the proceeds of the sale of productive assets ($220), we subtract the cost of goods sold ($100), and then the costs of employee services ($100) to leave $20 of net income, or a net income of one dollar per outstanding share.

Moving to year 2, the proceeds of asset sale at the end of year 1 of $220 are now reinvested with additional labor inputs of $100 similarly paid for by the grant of 9.09 shares ($100 divided by $11, which is the price per share at the end of year 1) to yield (at 10% rate of return) an asset worth $352. The cost of asset sold is $220. The total cost of dilution is again $110; it is broken down into its components: the dilution of initial stock value (as of the beginning of year 2) is 9.09 (shares issued)/29.09 (number of outstanding shares) of the $220 beginning of year 2 stock value. Similarly, the dilution of earnings is computed as $31.25 (portion saved by issuing shares) and $10 (the dilution of $32 net income). Similar transactions and computations yields a revenue of $497, and cost of asset sold of $352 in year 3, which after subtracting the cost of employee services of $100 yields a net income of $45. The cost of dilution reflected as a transfer of wealth from pre-existing shareholders to new shareholders is again $110.

To evaluate the soundness of the accounting treatments in this example, I make use of what Kirschenheiter et al. refer to as “sensible relations” that they expect “good” accounting to hold:

(1) Return on equity equal to the cost of capital.

(2) Forward price-earnings ratio equal to the inverse of the cost of equity capital.

(3) Equity market value equal to the present value of expected net dividends.

Consider relation (1) from the perspective of pre-existing shareholders. The corporation’s profit is $120 (the value enhancement). The cost to these shareholders is $110 (50% of the original asset value of $100 plus 50% of the profits of $120). Hence, the net income from operations is $10, which is 10% (the cost of capital) of their $100 equity.

Similarly, the new shareholders’ gain from the pre-existing shareholders’ dilution is $110, and their cost of providing services is $100, leaving them a net income of $10, which again is 10% of their investment. As can be seen from Exhibit 1, the computations for year 2 and year 3 yield a similarly constant return on equity (ROE) of 10%.

To test relation (2), observe that the ending price in year 1 of $11 per share, divided by the end of year 2 earnings per share of $1.10, yields a forward price-earnings ratio of $10, which is the inverse of the cost of capital. An identical forward price-earnings ratio of 10 can be computed at the end of year 2.

Finally, we can easily ascertain that the present value of the expected liquidating dividends equals the equity market value. For example, in year 1, the equity market value is the $220 proceeds from the sale of the asset. The expected net liquidating dividend is $266.20, which is the $497 cash available at the end of year 3 multiplied by the ratio of outstanding shares at the end of year 1 (20) to the outstanding shares at the end of year 3 (37.36). The present value of $266.20 computed for two years at the 10% cost of equity capital is $220.

The separate Corporation Income Statement reflects the revenue ($220) and the cost of asset sold ($100), but it does not show the shares granted as an expense. There are compelling reasons not to reflect a share-based payment as expense. By issuing shares or options in exchange for employee services, the corporation does not part with any asset. From the standpoint of the creditors, the corporation’s income is $120.

However, the cost of dilution ($110) is explicitly and visibly shown in the separate Statement of Costs and Benefits to Pre-existing Shareholders (see the Table). In the latter, the different events that cause dilution in pre-existing shareholders’ wealth are detailed. The Statement of Enterprise Income that focuses on the organization as a whole combines the Corporation Income Statement and the Statement of Costs and Benefits to Pre-existing Shareholders. As indicated, the compensation cost is shown there under the “Total” column.

This case illustrates the distinction between the corporation and its equity shareholders. The shares granted leave the corporation’s own assets intact. No resource is given up; creditors’ interests are not harmed. Pre-existing shareholders, on the other hand, suffer the cost of dilution caused by the grant.

It is quite clear in this case that expensing options in the corporation’s income statement serves neither the informational needs of creditors nor those of equity investors. Creditors, unless they undo the expensing, would compute typical income-based solvency ratios that are erroneous. Equity investors, as well, would overestimate the cost of dilution if they incorporate in their valuation a diluted income per share with the numerator taken as the net income after the expensing of shares or options granted, or if they otherwise reckon that share-based payments have diluted their holdings.

Since share-based payments have increasingly become the norm, separate statements are needed to serve the interests of the two distinct sets of users.

Consider the benefits of having these two distinct statements. The corporation’s income statement would reflect strictly the corporate entity’s sacrifice of assets and assumption of liabilities in the process of generating earnings—uncontaminated by events that did not affect the corporation’s net assets. More important, it alone would provide the relevant numbers to be used by creditors to assess the potential returns and risk associated with their investments in the company. After all, they are not affected by dilution; only common stockholders are.

The second statement would reflect the effect of such dilution. The two distinct statements would also yield insight into management’s policy with regard to financing its acquisitions.

Yet some parties might have need to evaluate the entire enterprise. Suppose, for example, an existing shareholder wished to identify the total cost, regardless of who incurred it, that was required to generate the reported revenue. The answer would be the sum of the respective components in the two distinct statements. The combined statement would aggregate these two to show the total expense, revenue, and income accruing to both the corporation and its shareholders—the enterprise.

Although the Statement of Enterprise Income looks similar to the income statement required by the FASB, there are important differences. First, the component statements distinguish between the cost to the corporation and the cost to pre-existing shareholders. Second, the Statement of Costs and Benefits to Pre-existing Shareholders displays prominently the dilution effects and states what resources the corporation chose to acquire by diluting pre-existing shareholders’ wealth. Third, the dilution is shown not only with respect to earnings but also with respect to book value (not explicitly displayed under current GAAP).

Fourth, as indicated, if there were creditors in this example (as reflected in Exhibit 2 of the working paper version of this note), the Corporation Income Statement would reveal that the value added available to them is in fact the full $120, and not the enterprise net income of $20. Fifth, the wealth transfer from pre-existing shareholders to new shareholders is visibly reflected: the cost of dilution to pre-existing shareholders is a net transfer to the new shareholders.

The return to both groups of shareholders is 10% ($10/100). Finally, as is evident from the Statement of Costs and Benefits to Pre-existing Shareholders, the cost to pre-existing shareholders (in this example) of $110 is $10 more than their cost were the corporation to pay cash for employee services, $100; the employees now also share in the return rather than merely being compensated for their labor. In essence, they have become partners, assuming the risk of ownership. The statements proposed clearly reflect this change in risk-sharing, an element that is missing in the FASB’s method.

Kip Hagopian Responds

I wish to thank Professors Benston, Hermalin, and Staubus for taking the time to critique the article, “Expensing Employee Stock Options is Improper Accounting.” Many thoughtful comments were made in these letters to the editor. In both sum and substance, the letters covered a lot of ground. Since some of the most important points were raised in more than one letter, I have chosen to combine my rebuttal arguments into one reply.

The short time period between receipt of these letters and the deadline for publication of the Fall edition of the CMR has precluded me from circulating my remarks to all of the signatories to the article; consequently the views expressed below are mine alone and may not be shared by every member of our group.

The letter from Professor Joshua Ronen arrived just prior to the deadline for my reply. While I have read this letter, I have not had time to give it the attention it deserves. For now, suffice it to say I am gratified by the Professor’s following statement, which appears near the beginning of his letter: “I agree with Hagopian’s arguments.......”

Pay for Services or Pay for Performance?

All of the respondents contend that an ESO grant is a payment for services. As we state in the article, we believe ESOs are not granted for services per se (although some services are necessary); rather, they are granted to achieve a particular result.

In the three examples of gain-sharing instruments cited in the article—sales commission agreements, profit sharing plans, and contingent-fee lawsuits—payment is contingent upon performance, not on the amount of services rendered (indeed, a salesperson on commission could work hard all year and earn no income).

In the case of an ESO, performance is defined as a rise in the market value of the granting entity; if there is no increase in this value there is no profit to the ESO holder. In all three of the examples cited, GAAP requires that expenses be booked when the associated gain is booked and not before. But the accounting treatment of ESOs is substantially different. Under FAS 123R, the ESO contract is valued at the time of grant and expensed over the vesting period regardless of whether the conditions of the contract (a rise in market value) are met, and regardless of whether the ESO is ultimately forfeited.

When Is Payment Made?

In his letter, Professor Hermalin writes (in comparing ESOs to sales commissions), a “stock option is paid (granted) before that employee performs, whereas a sales commission is paid after.” He goes on to say: “if we accept as a principle that an expense is incurred when payment is made, then it seems hard to argue that some expense has not been incurred the moment the option changes hands.”

With due respect to Professor Hermalin, I believe he is confusing an ESO with a transferable option. If a transferable option were conveyed to an employee, payment would be made (and an accounting charge would be appropriate) on the day of issuance because a transferable option is convertible into cash.

An ESO, however, does not have a realizable value at grant, so from an accounting perspective (see below for a comparison of the accounting and economic perspectives in measuring the cost of ESOs), payment is not made when it is granted.

In all pay-for-performance contracts, payment is made when it is earned; that is, when the performance conditions of the contract are met. In the case of an ESO these conditions are not met and payment is not made until the ESO vests and goes into the money. In this respect, an ESO is identical to all of the gain-sharing examples cited above. Yet under FAS 123R, ESOs are accounted for entirely differently.

Value vs. Cost

I believe the respondents have mistakenly equated an ESO’s economic value to the recipient with its accounting cost to the granting entity. One does not necessarily correspond to the other—at least not from an accounting perspective.

Consider the example of a contingency lawsuit. When a company enters into a contract with the plaintiff’s law firm, the law firm clearly has received something of economic value. If the contract were transferable, there are any number of other law firms that would buy it. But does the company incur an expense when it engages the law firm? Not according to GAAP and, I would argue, not according to common sense either.

Or more on point, consider a hedge fund partnership in which the manager receives 20% of the profits. When the partnership is formed, the hedge fund manager receives something of significant value. But has the partnership incurred an expense when this contract is consummated? Again, not according to GAAP. No expense is charged on the books of the partnership when the managers enter into their profit-sharing contracts.

Accounting Disparity Between Corporations and Partnerships

In Note 1 of the article, we point out that the profit-sharing contracts granted to hedge fund managers are conceptually identical to ESOs. Quoting from the article: “The allocation of 20% of the profits of a hedge fund or venture capital partnership is the exact economic equivalent of granting the investment managers a non-transferable option on 20% of the partnership’s assets at cost. The limited partners of a partnership are conceptually identical to the shareholders of a corporation and the assets of an investment partnership are economically the same as the assets of a corporation.”

Despite the economic equivalence of these two transactions, the accounting for ESOs (under FAS 123R) is substantially different from the accounting for hedge fund profit-sharing contracts. I see no reason why the accounting for two conceptually identical transactions should differ simply because of a difference in the legal form of the organization.

If a manufacturing company were organized in partnership form, and gave a profit participation to its employees, would this participation be expensed? If so, how would this disparity in accounting treatment (relative to hedge fund contracts) be justified?

Opportunity Cost

Professors Benston and Hermalin base much of their case for expensing on their contention that the grant of an ESO is an opportunity cost because, according to Professor Benston, “close substitutes [for ESOs] could be sold to investors.”

I believe the Professors are right to focus on the opportunity cost argument because, in my opinion, that is the only possible basis on which to justify a charge to the entity (remember, no asset or liability account is affected by the grant of an ESO, thus, the grant does not per se meet the standard definition of an expense).

We go to considerable lengths in the article to explain why we do not believe an opportunity cost is incurred when an ESO is granted. I will reiterate just part of our analysis, which I think should be dispositive.

Professor Hermalin states: “nothing would prevent a firm from selling options that have the structure of ESOs.” I disagree. First, for obvious reasons, ESOs are issued exclusively to persons that have the capacity to positively impact the market value of the enterprise. Thus, issuing “similar options to third parties” (as the FASB puts it) would be contrary to the ESO’s purpose and is proscribed. I believe that charging an opportunity cost in this circumstance would be a misapplication of what is otherwise a perfectly valid economic principle.

Second, if a contract with terms identical to an ESO were offered on the market, there would be no “willing buyers.” It is not an ESO’s vesting restrictions or even its non-transferability that make it unsaleable on the open market; rather, it is the fact that all ESOs are cancelable at the will of the entity that issues them (by terminating the employee).

A firm acting in its economic best interests, would always, therefore, cancel any outstanding ESO-like option that is held by a non-employee prior to its going into the money. Indeed, to not do so would be a breech of the fiduciary duty of the officers and directors of the firm. This means that no one other than an employee would purchase an option with the terms of an ESO, no matter how low the price. But charging employees for ESOs is uneconomic to the firm, since to do so would erode and ultimately nullify their value as an incentive compensation tool.

Again, consider the above three gain-sharing examples that are settled by the entity: In any of those contracts, does a company incur an opportunity cost by not charging the other party to the contract the expected value (the discounted present value) of the party’s projected gains?

Alternatively, does a hedge fund partnership incur an opportunity cost by not charging the fund managers the expected value of their projected gains? (Remember, under well accepted financial economic theory, expected value is the economic equivalent of projected future value, so if the employees or the fund managers paid this amount for their contracts, they would effectively be working for nothing.)

In all cases, the answer must be “no” because to do so would defeat the purpose of the contract and consequently would not be in the company’s economic best interests.

The FASB apparently agrees because GAAP does not require that an expense be charged when any of the above-described contracts are consummated.

What Is (and Is Not) a Gain-Sharing Instrument?

The critical distinguishing characteristic of a gain-sharing instrument is that its accounting cost to the issuer is a fixed fraction of, varies with, and is wholly dependent upon the existence of the gain to which it is linked—if there is no gain, there is no cost; if there is a gain, the cost is a fraction of that gain.

In his challenge to my position that ESOs are gain-sharing instruments, Professor Hermalin writes, “If this position is taken to its logical end, then one could argue that a great many payments made by the firm need not be counted as expenses.”

He supports this view by use of a hypothetical which assumes the following facts (I am paraphrasing): A company has revenue of $30,000 per month and total expenses of $20,000 per month, which is composed solely of the salary the company pays to its only employee.

Based on these assumed facts, he argues that, “The salary is a form of gain-sharing instrument.” He then contends that “the shareholders have agreed to part with 2/3 of the gain.”

Perhaps I am missing something, but I believe that by any reasonable interpretation of the hypothetical facts set forth above, both of the professor’s statements are in error. Taking them in reverse order:

(1) The shareholders have not agreed to “part with 2/3 of the gain.” They have agreed to pay $20,000 in salary regardless of what the gain (if any) turns out to be. If the revenue turns out to be $10,000, the salary will be 200% of the gain; if it is $50,000 it will be 40% of the gain.

(2) Salary does not meet any of the three conditions established above and, thus, is not a gain-sharing instrument. Salary is a semi-fixed cost that must be paid whether revenue materializes or not.

I believe it is an irrefutable fact that an ESO is a gain-sharing instrument. Specifically, it is a contract in which the stockholders’ share their gains (stock appreciation) with employees. If the employee is successful in creating value for the shareholders, he or she will earn a profit that will be a dollar for dollar reduction in the shareholders’ gains. Thus, the cost will be located where the gain is.

Accounting Cost vs. Economic Cost

Accountants and economists often differ on the proper measurement and timing of financial transactions. The following is my perspective on both the accounting cost and the economic cost of ESOs. Here I assume that an at-the-money ESO is granted to an employee and is fully vested on the day of grant.

The Accounting Perspective

On the day of grant, the employee cannot realize any value from an ESO (it cannot be transferred and it has no intrinsic value if exercised). The grant itself does not result in an outlay, a using up of an asset or the creation of a liability; therefore, it does not meet the standard definition of an expense. Finally, the granting entity does not incur an opportunity cost when the ESO is granted.

Accordingly, no accounting cost should be recorded on the books of either the granting entity or the entity’s shareholders. But as the stock price appreciates, the ESO will gain realizable intrinsic value that will accumulate on the employee’s books as profit.

But where is the cost located that matches this profit? We know that as the stock of the entity appreciates, the shareholders’ ownership will be diluted, resulting in a reduction in the appreciation of their holdings in an amount exactly equal to the employee’s profit (ignoring taxes in both cases). Thus, the shareholders effectively transfer a share of their gains from their balance sheets to the employee’s balance sheet.

Consistent with the way almost all gain-sharing instruments are accounted for (all but ESOs) both the value (profit) and the corresponding cost (reduction of shareholder gains) are reflected only if and when they actually occur.

The cost of ESOs to shareholders is a mathematical inevitability when the stock price rises. As a result, accountants do not have a choice between locating the cost on the shareholders books or on the entity’s—it simply materializes on the books of the shareholders as dilution occurs.

Under FAS 123R, therefore, the cost of an ESO is charged twice: once to the shareholders and once to the entity owned by those shareholders. Moreover, the charge to the entity (the discounted present value of the projected spread) is the economic equivalent of the charge to the shareholders. I believe this is double counting and is, therefore, improper accounting.

The Economic Perspective

When an ESO is granted it has economic value to the recipient on the day of grant. This value is not realizable but is implicit. This implicit value is the discounted present value of the projected gain in the ESO at the end of its expected term.

The economic cost to shareholders of the ESO on the day of grant is the reciprocal. It is the discounted present value of the projected reduction in their projected future gain. In other words, whatever value the ESO’s implicit call premium is to the employee is an implicit drag on the shareholders’ future stock appreciation. Thus, from an economic perspective, there is a symmetrical transaction at the date of grant.

The Real Economic Cost (or Gain) of an ESO

There is another economic aspect to this issue that we touch on only briefly in the article. The economic costs to shareholders that are described above are really only “nominal” costs. They do not take into account the incentive effects of the ESO. Here’s what we say in the article:

“The objective of an ESO plan, like any other incentive compensation program, is to increase employee productivity, and in so doing, raise the market value of the enterprise to a level in excess of the value that might be attained if the plan did not exist. If this objective is achieved, the use of an ESO plan will result in a net economic gain to shareholders, not a cost. Nonetheless, for the purposes of this paper, shareholder value transfer is treated as a cost because it reduces the preexisting shareholders’ nominal gain.”

The point is, if ESOs are not issued in such abundance that their dilutive effects outweigh their incentive effects, there will be no economic cost—there will be an economic gain.

Conclusion

In conclusion, I again would like to express my appreciation to the four professors for taking the time to write letters to the editor regarding our article. I believe they have all made useful contributions to this debate. I regret, due to the shortage of time, I was unable to respond to Professor Ronen.

With due respect, I do not believe Professors Hermalin, Benston, and Staubus have successfully refuted the basic logic of the article. In my rebuttal to the Professors, I have asserted the following:

(1) An ESO grant is not a payment for services per se; it is a contract that provides for a payment that is conditioned upon performance.

(2) An ESO’s value to the recipient does not equate to a cost to the granting entity.

(3) An ESO grant does not meet the standard definition of an expense and is not an opportunity cost.

(4) An ESO is not akin to a transferable option; rather, it is a contract that uses an option to create a gain-sharing instrument.

(5) Under GAAP, all gain-sharing instruments (except ESOs) have no accounting cost unless and until there is a gain to be shared.

(6) In the case of an ESO, it is the shareholders that reap the gain; accordingly that is where the cost is located.

(7) This cost is already fully accounted for using the treasury stock method of measuring dilution.

(8) Charging the entity an additional cost that is the economic equivalent of the cost to shareholders is improper accounting.

I believe these assertions are logically sound and comport with accepted principles of both accounting and economics. Of course, I welcome any additional comments from the respondents or from other CMR readers.