Stock Option Expense: The Sword of Damocles Revealed.

SYNOPSIS: Statement of Financial Accounting Standards (SFAS) No. 123 is one of the most controversial accounting standards ever issued by the Financial Accounting Standards Board (FASB 1995) (SFAS No. 123, para. 376). More than five years have passed since SFAS No. 123 first required firms to

This study examines the effect of stock option expense on the diluted earnings per share and return on assets of 100 firms identified by Fortune magazine as "America's Fastest-Growing Companies." We find that stock option expense has a material impact on the performance measures for a majority of our sample firms. Moreover, our analysis predicts stock option expense will grow over the next several years, potentially doubling in magnitude during that time. We also document some noncompliance with the disclosure requirements of SFAS No. 123, finding that 12 percent of the sample firms provided incomplete information during the most recent year examined.

Keywords: SFAS No. 123; stock option expense; disclosure.

Data Availability: Please contact the authors.

INTRODUCTION

"'With disclosures coming out next year, noting how badly companies could be hurt by the FASB proposal, investors may not realize how close is the sword of Damocles,' says Michael C. Bernstein, a partner of Grant Thornton, accountants." (1'2)

During the past two years a mere $1 per year was charged against Apple Computer's income for the cash compensation paid to CEO Steve Jobs. Yet Jobs is considered to be the recipient of perhaps the most generous compensation package in business history. How can this be? In early 2000, Apple's Board of Directors awarded Jobs ten million stock options valued at more than $400 million (Menn 2000). Apple did not charge the $400 million against income, consistent with existing accounting standards. Douglas Maine, CFO of MCI Communications Corp, agrees with this accounting treatment, stating that the cost of stock options granted to employees is a "fictitious expense" that should not be included in the computation of net income (Barr 1994). In contrast, analysts are routinely taught to use net income adjusted for stock option expense to compute valuation measures such as the price-earnings ratio. For example, the financial statement analysis text used as primary reading for the Certified Financial Analyst (CFA) exam ination clearly states, "measures that fail to recognize the real cost of stock options overstate corporate performance. [Stock option expense] disclosures should be used [when stock option grants are sizeable] to correct such overstatement" (White et al. 1998, 644). (3,4)

The debate on whether investors should or should not ignore stock option expense when assessing firm performance and value is moot if the impact of stock option expense on reported earnings is immaterial. More than five years have passed since Statement of Financial Accounting Standards 123 (SFAS No. 123) (FASB 1995) first required firms to either include stock option expense in the computation of net income or disclose its impact on income in their footnotes. As a result, sufficient data are now available to assess the materiality of the impact of stock option expense on measures of firm performance.

This paper documents the effect of stock option expense on the financial performance of 100 firms identified by Fortune magazine as "America's Fastest-Growing Companies." We select this group of firms because high-growth firms tend to use stock options extensively. If we fail to detect a material effect for this set of firms, then it is likely to be immaterial for all but a few firms.

The descriptive evidence presented in this paper suggests that stock option expense has a material effect on measures of firm performance for these firms. In the most recent fiscal year analyzed, we document a median reduction of 14.0 percent in earnings per share (EPS) and a median reduction of 13.6 percent in return on assets (ROA) due to stock option expense. Moreover, we find that the magnitude of stock option expense should grow larger, at least for the near future, thereby adding to the significance of the figures involved.

The magnitude of the effect makes it critical that users of financial statements who employ stock option expense figures in decision making have access to relevant data concerning firms' stock option plans. Since none of our sample firms opt to recognize stock option expense in reported income, users must rely on the firms' responses to the disclosure requirements of SFAS No. 123. We document some noncompliance with these disclosures, because 12 percent of our 100 sample firms fail to meet at least one of the disclosure requirements of SFAS No. 123 in the most recent fiscal year examined.

BACKGROUND

The extraordinary level of controversy surrounding the passage of SFAS No. 123 and the impact of this controversy on the standard-setting process call for discussion of how SFAS No. 123 evolved and a short summary of its requirements.

History of SFAS No. 123

The FASB's initial proposal to require firms to expense the total fair value of options when granted elicited the strongest show of opposition in the history of the FASB (SFAS No. 123, para. 59). The Board was inundated with comment letters throughout 1992 and the first half of 1993, even though the Exposure Draft, Accounting for Stock Compensation, was not issued until mid-1993 (SFAS No. 123, para. 374). The early onset of this outcry was unusual and a harbinger of the fractious debate to come. In the months after the Exposure Draft appeared, the FASB received over 1,700 comment letters, including more than 1,000 form letters from employees of firms expecting to be hard hit by the standard. Congressional resolutions both encouraged the FASB not to change accounting for employee stock options and stated that Congress should not get into the business of legislating accounting rules (SFAS No. 123, para. 376).

Faced with a level of opposition that some believe threatened the FASB's existence (SFAS No. 123, para. 60), the Board ultimately retreated from its initial, preferred position in several important respects. First, the final standard does not require recognition of stock option expense in income, instead allowing firms to choose between recognition and footnote disclosure (SFAS No. 123, para. 11). Second, the annual charge for stock option expense is not the total fair value of options granted during the year. Instead, firms amortize the total fair value over the period(s) in which the related services are rendered (SFAS No. 123, para. 30). These concessions lessened the perceived impact of the standard, permitting a final standard to be issued in October 1995.

Accounting for Employee Stock Options

Under SFAS No. 123, firms must use the fair value method of accounting for employee stock options to estimate stock option expense. (5) The standard encourages firms to recognize this expense in reported net income, but allows them to continue using the intrinsic value method prescribed by APB No. 25 (Accounting Principles Board 1972) for recognition purposes. (6) Firms using APB No. 25 must provide footnote disclosures of pro forma net income (NI) and earnings per share (EPS) computed using the fair value method (SFAS No. 123, para. 11). A firm adopting the fair value method for recognition purposes may not reverse its decision (SFAS No. 123, para. 14). Due to this restriction, managers may perceive the cost of adopting the recognition provisions of SFAS No. 123 to be higher than would otherwise be the case.

The fair value of an option on the grant date is estimated using any accepted option-pricing model that incorporates each of the following: (1) exercise price, (2) expected life of the option, (3) current price of the underlying stock, (4) expected stock return volatility, (5) expected dividend yield, and (6) the risk-free interest rate (SFAS No. 123, para. 19). While SFAS No. 123 mentions the Black-Scholes and binomial option-pricing models, any reasonable option-pricing model that incorporates the listed variables is acceptable. Annual stock option expense is computed by amortizing the fair value of options granted during the year over the period(s) the related employee services are rendered. (7) Generally the service period is presumed to begin on the grant date and end on the vesting date (SFAS No. 123, para. 30).

The risk-free interest rate used in the computation is the rate on a zero-coupon U.S. government bond with a remaining term equal to the expected life of the option (SFAS No. 123, para. 19). The exercise price and, for public companies, the current price of the underlying stock are both readily available. The remaining values are based on past experience adjusted to reflect management's best estimate of the future (SFAS No. 123, Appendix B).

Disclosure Requirements

Firms opting to retain the intrinsic value approach must disclose in a footnote pro forma net income and earnings per share for every year that an income statement is provided. Firms must also provide a description of their plan (or plans) and the following information for each year they provide an income statement:

* the number and weighted-average exercise price of options:

* outstanding at the beginning of the year

* granted during the year

* exercised, forfeited, or expired during the year

* outstanding at the end of the year

* exercisable at the end of the year

* the weighted-average grant-date fair value of options granted during the year

* the option-pricing model employed and weighted-average values of the assumed risk-free rate, expected life, expected volatility, and expected dividend yield

Effective Date and Transition

The FASB applies SFAS No. 123 prospectively to stock options granted in fiscal years after December 15, 1994. Therefore, stock option expense does not reflect the fair value of awards granted before December 15, 1994. As a result, even if the total fair value of options granted annually remains constant, current stock option expense will increase each year until the prospective adoption of SFAS No. 123 is completely phased in. This phase-in period is often equal to the number of years in the vesting period, as most firms amortize the fair value of options granted over the vesting period.

SAMPLE SELECTION AND DATA AND SAMPLE DESCRIPTION

Sample Selection

Our sample is comprised of the companies included in Fortune's September 1999 listing of the 100 fastest-growing firms (Daniels et al. 1999) as compiled by Zacks Investment Research. This list is replicated in Exhibit 1. To be considered for inclusion in this group, a firm must have been in operation for at least three years and have revenues and market capitalization in excess of $50 million. Firms must also produce an annual growth rate both in revenues and EPS of at least 30 percent during the last three years. Firms meeting all these requirements are then ranked by growth in revenues, EPS and three-year total stock return. The 100 firms with the highest sum across these three categories of growth are included on the list.

We believe a sample of high-growth firms, many of which are high-tech, is an interesting group to examine for several reasons. First, high-growth firms often use stock options to compensate employees to conserve cash needed to support growth (Erickson 1992). In addition, as a group, high-tech firms aggressively lobbied against the FASB's original proposal. For example, high-tech industry lobbyists organized letter-writing campaigns as well as an anti-FASB rally attended by 4,000 Silicon Valley workers in March 1994 (Shao 1994). Such efforts clearly indicate that high-growth/high-tech firms expected to be hard hit by the standard. Indeed a 1995 article states, "the FASB's stock options plan would have been devastating by cutting reported earnings as much as 50 percent and discouraging investors from buying stock in high-growth companies that rely on options to attract talented personnel" (Demery 1995). A failure to detect a material effect of SFAS No. 123 for this sample of firms suggests that it is likely to be immaterial for all but a few firms.

Data Description

We obtained each firm's 10-K filed with the Securities and Exchange Commission during 1999. Data were collected from the stock option footnote for the current fiscal year (referred to subsequently as year 3) and the previous two fiscal years (years 2 and 1, respectively). For the 66 percent of our sample firms whose 1999 10-K filings pertain to their fiscal years ending in 1998, our three-year data period spans the years 1996- 1998. Because the remaining 34 percent of our sample firms' filings cover 1999 fiscal year ends, the three-year data period for those firms includes the years 1997-1999. We collected the other financial data used in our analyses from Compustat, when it was available. In several cases, we had to hand-collect the necessary data from the firm's 10-K. CRSP provides market data.

Sample Description

Sixty-nine percent of the sample firms trade on the NASDAQ or over-the-counter compared to 31 percent on the NYSE/AMEX. These data suggest that smaller, emerging companies likely populate our sample. Twenty-nine percent of the sample firms are members of the Business Services industry and the remaining 71 percent are distributed over 26 different industries. (8) The Business Services industry includes biological research, engineering services, computer programming, and prepackaged software. A 1994 editorial critical of the FASB's proposals identifies biotechnology and computer-related industries as among those to be significantly impacted by the standard (Robson 1994). Thus, the sample appears well populated by firms expected to report large amounts of stock option expense.

Panel A of Table 1 provides descriptive statistics pertaining to several additional firm characteristics. Median (mean) market value of common equity at the end of the three-year period is approximately $512 million ($2.8 billion). (9) As a point of comparison, the market value of common equity of the average NASDAQ firm at the end of 1998 is similar at approximately $510 million (NASDAQ-Amex 1999), which provides further evidence that smaller, emerging companies likely populate our sample. Most of our sample firms are profitable. Diluted earnings per share (EPS) are positive at the 25th percentile, for example. Moreover, median (mean) ROA (computed by dividing reported net income by average total assets) is 10.2 percent (11.6 percent).

The data presented in Panel B of Table 1 document the high-growth nature of our sample. During the sample period, the firms experienced double- (and in some cases, triple-) digit growth in market value of equity, diluted EPS, and ROA. For example, median (mean) growth in the market value of common equity is 157.1 percent (228.6 percent) and median (mean) growth in diluted EPS amounted to 173.2 percent (569.3 percent). In contrast, median growth in ROA is more moderate at 28 percent, but the mean growth in ROA is similar to the mean growth in EPS (410.8 percent). (10)

These descriptive statistics suggest that the firms included in Fortune's list of the 100 fastest-growing firms share characteristics in common with firms expected to use stock options extensively to compensate employees. This supposition is confirmed by data provided in Panel C of Table 1. By the end of year 3, the median (mean) ratio of employee stock options outstanding divided by common shares outstanding is 12.1 percent (12.8 percent). Options exercisable at the end of year 3 alone amount to 4.7 percent (5.7 percent) of common shares outstanding. Finally, in year 3 the median (mean) number of options granted amounted to 3.5 percent (4.7 percent) of common shares outstanding.

Panel D of Table 1 reports growth in firms' stock option plans during the sample period. Over the three years, median (mean) growth in options outstanding deflated by the number of common shares is 10.4 percent (31.6 percent). Similarly, the median (mean) growth in options exercisable at year-end deflated by common shares outstanding is 7.8 percent (41.9 percent). Finally, the median (mean) growth in the number of options granted annually as a percentage of common shares outstanding is 13.9 percent (138.3 percent).

In summary, because our sample firms' use of stock options to compensate employees appears to be significant and increasing, we believe the sample includes firms most likely to report material effects under SFAS No. 123.

RESULTS OF ANALYSIS

Analysis of the Impact of Stock Option Expense on Firm Performance

A meaningful analysis of the impact of SFAS No. 123 must begin with the assumption that the data produced by the fair value approach prescribed by SFAS No. 123 are reasonable. Huddart and Lang (1996) provide evidence that early exercise of options is common due to the nontransferability of employee stock options. When fair value is estimated using the maximum term of the options, estimates derived from the Black-Scholes option-pricing model tend to overstate fair value. The FASB attempted to adjust for this effect by requiring firms to use the expected life of the option rather than its maximum term, but Huddart and Lang's (1996) evidence suggests that the expected life is fundamentally unpredictable. However, the FASB learned in discussions with compensation consultants that firms use the Black-Scholes and other option-pricing models to value employee stock options for purposes other than accounting (SFAS No. 123, para. 373). Moreover, Carpenter (1998) finds that the estimated fair values of employee stock o ptions produced by the Black-Scholes computations described in SFAS No. 123 are close to those produced by more complex models. Thus the SFAS No. 123 approach, while not perfect, appears to produce fair value estimates that are reasonable.

Our examination of the effect of stock option expense on firm performance focuses on two well-accepted measures of firm performance: diluted EPS and ROA. (11) Table 2 reveals differences between EPS and ROA based on reported income and income adjusted for stock option expense. Both the net difference and the percentage difference--computed by dividing the net difference by the reported amount--are presented for each of the three years. (12)

The median (mean) difference in diluted EPS due to stock option expense is $0.03 ($0.047) in year 1. This figure is 9.8 percent (45.4 percent) of reported diluted EPS. To gauge the materiality of the effect of stock option expense on diluted EPS, we compute the proportion of firms for which stock option expense reduces diluted EPS by 5 percent or more and 10 percent or more. (13) In year 1, stock option expense reduced reported diluted EPS by 5 percent or more for 61 percent of the firms and 10 percent or more for 48 percent of the sample firms.

By year 3 the median (mean) impact of stock option expense on diluted EPS grows to $0.12 ($0.162). Relative to reported diluted EPS, the median (mean) reduction in diluted EPS due to stock option expense is 14 percent (22.9 percent). Based on a cutoff of 5 percent, we find that 78 percent of firms suffer a material reduction in diluted EPS; even at the 10 percent cutoff, we find a material effect for 59 percent of the firms.

Based on these results, we conclude that the impact of stock option expense on performance as measured by diluted EPS is material for the majority of our sample firms. Although the impact of stock option expense is significant, it is much less than the dire predictions advanced during the height of the stock option debate. For example, while we find a significant reduction in earnings, it is much lower than the upto-SO-percent of reported earnings predicted in Demery (1995).

Panel B of Table 2 reports differences between ROA based on (1) reported net income and (2) net income adjusted for stock option expense. The impact of stock option expense on ROA is very similar to its impact on diluted EPS. In year 1, the median (mean) reduction in ROA is approximately 1.0 (1.5) percentage point(s). Relative to reported ROA, the median (mean) reduction in ROA is 9.5 percent (81.3 percent). Based on a 5 percent materiality cutoff, we find that the drop in ROA is material for 64 percent of the sample firms. At a 10 percent cutoff the proportion of firms with a material decline in ROA is 47 percent.

By year 3, the impact grows to a median (mean) drop in ROA of 1.4(2.4) percentage points. As a proportion of reported ROA, the median (mean) reduction in ROA amounts to 13.6 percent (22.8 percent) and, based on cutoffs of 5 percent and 10 percent, the reduction is material for 79 percent and 61 percent of the sample firms, respectively.

Analysis of Future Growth in Stock Option Expense

Financial statement analysis and valuation models often use current financial information to predict the future. Projections of stock option expense can benefit from comparing the after-tax stock option expense firms disclose in their footnote to an estimate of the after-tax total fair value of options granted in a given year. When firms' current stock option expense equals the fair value of the options granted, this ratio equals 1. Deviations from 1 are due to two distinct factors.

First, as discussed earlier in the paper, the magnitude of stock option expense will grow in the short-term due to the prospective adoption of SFAS No. 123. Prior to the completion of the phase-in period, the ratio of current stock option expense to the fair value of options granted during the year will be less than 1. For example, consider a December31 year-end firm that annually grants options with a total fair value of $100,000 beginning in 1995 when the SFAS No. 123 phase-in period starts. If the options have a five-year vesting period, then stock option expense for the year ended December 31, 1996 is $20,000. However, for the year ended December 31,1996, stock option expense is $40,000, consisting of one year of amortization of the fair value of options granted in 1995 and one year of amortization of the fair value of options granted in 1996. Our hypothetical firm's option expense increases by $20,000 annually until 1999, at which point the phase-in is complete and stock option expense reaches a "steady- state" amount of $100,000. The ratio of stock option expense for 1996 to the fair value of options granted in 1996 suggests that 1996 expense is 40 percent ($40,000/$100,000) of the amount of stock option expense that will be charged in 1999 once the phase-in period is over.

The second factor affecting the ratio of stock option expense to fair value of options granted is a change in the fair value of options granted annually. An increase in the fair value of options granted leads to a predictable increase in current and future stock option expense, since the total fair value of newly granted options is amortized over the vesting period. As a result, the impact of large stock awards or significant changes in the fair value of options granted can extend years into the future.

To illustrate, consider our hypothetical firm that annually grants options with a total fair value of $100,000. For simplicity, assume the phase-in period is complete (year 1999) and the annual stock option expense is equal to the fair value of the options granted, absent changes in the option plan. For the years 2000-2005, if the company increases options granted by 30 percent to $130,000, without changing its vesting period, then we observe the following effect on its current stock option expense and the related ratio. The stock option expense reported in 2000 will be $106,000, consisting of $20,000 each for the amortization of the options granted in 1997, 1998, 1999, and 2000 plus $26,000 for the amortization of the fair value of the current year options granted. The fair value of options granted will be $130,000, resulting in a ratio of stock option expense to fair value of options granted of 81.54 percent ($106,000/130,000). This ratio will equal 1 at the end of the year 2005 if the fair value of the opt ions granted remains constant at the new amount.

The more this ratio deviates from 1 (100 percent), the greater the degree to which current stock option expense used by firms to calculate pro forma income understates estimated steady-state compensation expense. A comparison of stock option expense to the fair value of options granted during the year highlights such cases. In those instances, financial statement users interested in assessing future profitability may prefer to estimate pro forma stock option expense using the fair value of the options granted instead of the firm's reported current stock option expense.

To calculate the ratio mentioned above, we must first calculate the fair value of options granted. The fair value of the options granted equals the weighted-average fair value of an option granted in a given year multiplied by the number of options granted in that year. This figure is then adjusted to an after-tax amount by multiplying the fair value of the options granted by 1 minus the statutory tax rate. (14) We assume a statutory tax rate of 38 percent. (15) Hanlon and Shevlin (2001) provide a detailed analysis of the estimated tax benefits of stock options and the effect of this tax benefit on the calculation of effective tax rates and other tax-related variables. They demonstrate that estimating the effect of employee stock options on effective tax rates is extremely complex. Consequently, we make the simple tax adjustment described above to provide readers with a rule of thumb that can be estimated relatively easily. Our simplification assumes that firms receive the full tax benefit associated with the ir stock option grants. If this is not true, then our simplified approach understates the expected increase in future stock option expense. (16)

The disparity between stock option expense and the fair value of the options granted by firms in our sample appears to be quite large. Figure 1 compares the median value of stock option expense to the median fair value of stock options granted in each of the three years included in our sample. In year 1, the median fair value of options granted is $2.6 million, whereas year 1 median stock option expense is a much lower $0.7 million. The median difference between stock option expense and the fair value of options granted in year 1 is $1.6 million. (17) This translates into a median current charge for stock option expense amounting to less than 31 percent of the fair value of options granted in year 1. (18)

By year 3, the median fair value of options granted increases to $5.8 million, whereas the median stock option expense is $3.1 million. (19) Although the gap between stock option expense and fair value of options granted (median difference of $2.8 million) is greater in dollar terms in year 3 than year 1, the ratio of stock option expense to the fair value of the options granted is closer to 1 (100 percent). At the median, stock option expense in year 3 is 51.6 percent of the fair value of options granted, up from 31 percent in year 1. The narrowing of the gap is likely due to the phase-in of the prospective adoption of SFAS No. 123 becoming more complete.

If the fair value granted annually is held constant at the year 3 figure, then stock option expense at the end of the vesting period will equal the fair value of the options granted annually. Given that the year 3 expense is only about half of the after-tax fair value of options granted that year, the median pro forma charge to earnings for stock option expense should double over the next several years.

Analysis of Accounting Policy, Disclosure, and Estimate Choices

In adopting SFAS No. 123 firms must decide whether to use the fair value method or the intrinsic value method to recognize stock option expense. Given the furor surrounding the passage of SFAS No. 123, perhaps it is not surprising that the majority of firms chose the disclosure-only option. However, we find 100% of companies making this choice. Why is this the case, when under different circumstances firms choose to voluntarily record larger expenses than required by accounting standards? For example, some firms voluntarily expense all software development costs even though capitalization is allowed, and even encouraged, under SFAS No. 86 (FASB 1985 [SFAS No. 86]) (Aboody and Lev 1998).

Perhaps the fundamental difference between these two accounting choices relates to their effect on cumulative earnings. The decision to capitalize vs. expense software development costs does not alter the firm's cumulative earnings, since the capitalized software development costs of today become the amortization expenses of tomorrow. However, the decision to recognize vs. disclose stock option expense results in a permanent positive difference in cumulative earnings, since options granted at-the-money never generate a charge to income under the intrinsic value method. This reason combined with the irreversibility of the adoption of the fair value method for recognition purposes provide little incentive for firms to discontinue using the intrinsic value method.

The fact that none of our sample firms recognize stock option expense in income accentuates the need for investor access to adequate information about stock option plans. Table 3 reports the frequency of several items of required disclosure associated with SFAS No. 123 for the most current year in our sample. Data pertaining to the disclosure (or lack thereof) of prior years' stock option information that should be provided in the firms' 1999 10-Ks are not reflected in this table.

In Panel A of Table 3 we report that 28 (1) sample firms do not disclose pro forma basic (diluted) EPS even though basic (diluted) EPS is disclosed on their income statement. SFAS No. 123 (para. 45) states that "if earnings per share is presented, pro forma earnings per share (is required)." Although the standard does not explicitly state that both pro forma basic and diluted EPS are required, providing pro forma figures corresponding to each EPS figure presented on the income statement seems more in keeping with the spirit of the standard. Firms are required to disclose the number of options exercisable at the end of the year, but one sample firm does not do so.

Table 3, Panel B indicates the frequency of various disclosures required when options are granted. Under SFAS No. 123, firms granting options must disclose their estimates of volatility, risk-free rate, life expectancy, and dividend yield in point or range form. Ninety-eight of our sample firms granted options in year 3. Two of these firms do not disclose volatility estimates and two other firms do not disclose estimates of option life expectancy. Although seven firms fail to disclose their dividend yield estimates, this may be because many of our sample firms do not pay dividends.

Perhaps our most troubling finding is that seven sample firms do not report the weighted-average fair value of options granted. These data are critical to the computation of the total fair value of options granted during the year. Suppose the weighted-average fair value of options granted, or alternatively the total fair value of options granted, is not provided. The burden of estimating these data using the Black-Scholes option-pricing model, or some other model, falls on the user of the financial statements. Given the complexity of the computations involved, it is likely that most users do not have adequate information or skills to produce proper estimates. Finally, although firms are required to disclose the option-pricing model used, one sample firm does not disclose this information.

The descriptive data presented in Table 3 indicate some measure of noncompliance with the disclosure standards of SFAS No. 123. A pictorial representation of compliance/noncompliance with the requirements of the standard appears in Figure 2. We classify firms as in "full compliance" if they disclose the following information for year 3: (l) pro forma basic or fully diluted EPS, (2) the option-pricing method used, (3) estimated volatility, risk-free rate, and option life expectancy, (4) the weighted-average fair value of options granted, and (5) the number of options exercisable at year-end. By this definition, 88 of the sample firms are in "full compliance." Of the 12 firms not in "full compliance," four failed to supply one of the estimates used in their option-pricing model, six failed to disclose the weighted-average fair value of options granted, one failed to disclose both the weighted-average fair value of options granted and the option-pricing method they employ, and one failed to indicate the number o f options exercisable at year-end.

SUMMARY AND CONCLUSION

This paper examines the impact on firm performance of perhaps the most controversial accounting standard ever produced by the FASB. We examine 100 of the fastest-growing companies in the United States; if we fail to detect a material effect for this set of firms, then it is likely to be immaterial for all but a few firms.

Our primary results are as follows:

* The impact of stock option expense on firm performance is material for the majority of the sample firms.

* Stock option expense will become even more economically significant in the near future, potentially doubling over the next three to five years.

* Not all firms comply with the disclosure requirements of SFAS No. 123.

This study does not address the ongoing debate over the usefulness of stock option expense data in firm valuation. However, we do find that the magnitude of the effect is too large to be ignored by financial statement users without careful consideration of the potential implications of stock option expense for firm performance and value. It is important to remember that our sample of firms was selected to include firms expected to use stock options extensively; other types of firms might be less affected by the standard.

Christine A. Botosan and Marlene A. Plumlee are Assistant Professors at the University of Utah.

(1.) Berton (1993).

(2.) Damocles was a courtier in ancient Syracuse who, according to legend, was seated at a banquet beneath a sword suspended by a single horsehair. "The sword of Damocles," therefore, has come to represent imminent disaster.

(3.) The list of primary readings for the CFA exam is available at http://www.aimr.org/cfaprogram/readings.html.

(4.) For a thorough discussion of arguments for and against the position that stock option expense is a cost that should be recognized in the financial statements see SFAS No. 123, pp. 28-40.

(5.) SFAS No. 123 also provides guidance regarding accounting for transactions with other than employees. Such transactions are not the focus of this study.

(6.) The intrinsic value method values a stock option at the difference between the exercise price of the option and the market price of the underlying security on the date of issue. Since most firms set the exercise price of the option equal to the market price of the stock on the issue date, application of the intrinsic value method generally results in no charge to income for stock option expense.

(7.) Under SFAS No. 123, options that expire prior to vesting do not give rise to compensation expense. Firms may estimate a forfeiture rate on the grant date and compute compensation expense on the number of options expected to vest. Alternatively, firms may compute compensation expense on all options granted, and adjust for actual forfeitures as they occur (SFAS No. 123, para. 28).

(8.) We define industry membership based on four-digit SIC codes sorted into the 48 industry classifications devised by Fama and French (1997).

(9.) Market value of equity is the market value of common equity reported on Compustat at the firm's year 3 year-end.

(10.) Beginning total assets for year 1 is unavailable for three forms. In these cases we compute ROA using ending total assets in place of average total assets in year 1.

(11.) In analyses not reported in the paper, we examined the impact of stock option expense on total net income, basic EPS, and E/P ratios. The impact of stock option expense on these additional measures of firm performance mirror the impact observed on diluted EPS and ROA. We employ diluted EPS as opposed to basic EPS because approximately one-quarter of the sample firms do not disclose basic EPS adjusted for stock option expense while almost all disclose adjusted diluted EPS.

(12.) Diluted EPS adjusted for stock option expense is not disclosed by one of the 100 sample firms in years 2 or 3. In this case, we substitute basic EPS adjusted for stock option expense for the missing diluted EPS data. Because three firms in year 1 provide neither adjusted basic nor adjusted diluted EPS figures, we have 97 observations in our analysis of the effect of stock option expense on diluted EPS in year 1.

(13.) Determining an appropriate cutoff for materiality is a subjective undertaking. However, Pattillo (1976) examines several kinds of materiality judgments and concludes that in practice a "rule of thumb" of 5-10 percent of net income is a widely used materiality criterion. Accordingly we use both 5 and 10 percent cutoffs to establish materiality.

(14.) It is necessary to do the analysis on an after-tax basis as most firms do not disclose pre-tax stock option expense.

(15.) We use a 38 percent tax rate based on the federal corporate tax rate of 35 percent plus 3 percent to reflect various state tax rates. We also perform our computations with firm-specific effective tax rates, but the resulting estimates are not reliable because our sample includes a fairly high incidence of unrealistic effective tax rates.

(16.) References to stock option expense and the fair value of options granted should be understood to refer to after-tax amounts. For expositional brevity, the adjective "after-tax" is omitted from this point forward.

(17.) The median difference is not equal to the median fair value less the median stock option expense as all three values are median amounts from the sample.

(18.) These calculations provide a conservative estimate of the ratio (i.e., potentially overstate the ratio of current option expense to the fair value of options granted and, consequently, understate the steady-state stock option expense). To provide a lower bound of the ratio, an assumption of no tax benefit associated with the stock option grant could be made. To do so, users could simply divide the current stock option expense by the fair value of the options granted. Using a 38 percent tax rate, this is analogous to multiplying the conservative ratio by (1 - 0.38). For example, for our sample the lower bound of the ratio in year 1 is 19.2 percent as compared to the 31 percent figure reported in the text.

(19.) The increase in the estimated fair value of options granted over time is consistent with the growth in plan size documented previously.

REFERENCES

Aboody, D., and B. Lev. 1998. The value relevance of intangibles: The case of software capitalization. Journal of Accounting Research (36): 161-191.

Accounting Principles Board (APB). 1972. Accounting for Stock Issued to Employees. APB Opinion No. 25. Norwalk, CT: APB.

Barr, S. 1994. FASB under siege. CFO 10(9): 34-41.

Berton, L. 1993. Firms fume at possible profit drain of deducting value of stock options. Wall Street Journal (June 3): 12.

Carpenter, J. 1998. The exercise and valuation of executive stock options. Journal of Financial Economics 48 (2): 127-158.

Daniels, C., D. Lanning, T. Maroney, and N. Tarpley. 1999. Fortune's one hundred fastest-growing companies. Fortune 140 (5): 90-98.

Demery, P. 1995. FASB caves in on stock options. The Practical Accountant 28 (2): 26-32.

Erickson, M. 1992. Stock options can benefit firms, employees. Seattle Times (April 6).

Fama, E., and K. French. 1997. Industry costs of equity. Journal of Financial Economics (43): 153-193.

Financial Accounting Standards Board (FASB). 1985. Accounting for the Costs of Computer Software to be Sold, Leased, or Otherwise Marketed. Statement of Financial Accounting Standards No. 86. Norwalk, CT: FASB.

-----. 1995. Accounting for Stock Compensation. Statement of Financial Accounting Standards No. 123. Norwalk, CT: FASB.

Hanlon, M., and T. Shevlin. 2001. Accounting for tax benefits of employee stock options and implications for research. Working paper, University of Washington.

Huddart, S., and M. Lang. 1996. Employee stock option exercises: An empirical analysis. Journal of Accounting and Economics 21 (1): 5-43.

Menn, J. 2000. Apple gives CEO major option grant. Compensation: Steve Jobs could buy 10 million shares in package valued at more than $400 million. It could be the largest ever. Los Angeles Times (January 20).

NASDAQ-Amex. 1999. NASDAQ-Amex fact book & company directory. Washington, D.C.: The NASDAQ-Amex Market Group Inc.

Pattillo, J. 1976. The Concept of Materiality in Financial Reporting. New York, NY: Financial Executives Foundation.

Robson, J. 1994. Group's stance will harm economy. San Francisco Chronicle (March 14).

Shao, M. 1994. Board backs off proposal. Plan urged deducting cost of stock options from profit. Boston Globe (December 15).

White, G., A. Sondhi, and D. Fried. 1998. The Analysis and Use of Financial Statements. Second edition. New York, NY: John Wiley & Sons, Inc.

EXHIBIT 1

Fortune's 100 Fastest-Growing Companies (a)

Rank  Firm Name

  1   Siebel Systems
  2   Meritage
  3   THQ
  4   Network Appliance
  5   Citrix Systems
  5   Salton
  7   Labor Ready
  8   Veritas Software
  9   Jakks Pacific
  9   Pilgrim Capital
 11   Actrade international
 11   Vitesse Semiconductor
 13   Consolidated Graphics
 14   Dell Computer
 15   Cree Research
 16   American Eagle Outfitters
 16   Qualcomm
 18   Arguss Holdings
 19   Polymedica
 20   Kellstrom Industries
 21   Railamerica
 22   Armor Holdings
 23   Extended Stay America
 24   Action Performance
 25   Tekelec
 26   Legato Systems
 27   Biomatrix
 27   Monaco Coach
 29   Aspect Development
 29   Compuware
 29   Hauppauge Digital
 29   Visx
 29   Zomax
 34   AVT
 34   Pacific Sunwear of California
 36   Mercury Interactive
 36   Suma Industries
 38   Autonation
 39   Whitman-Hart
 40   Profit Recovery Group Intl.
 41   Biogen
 42   Aavid Thermal Technologies
 43   Rainforest Cafe
 44   I2 Technologies
 45   MTR Gaming Group
 45   Osteotech
 47   Resource America
 48   Tower Automotive
 49   Cade Industries
 49   Insight Enterprises
 51   Analytical Surveys
 51   HNC Software
 53   Century Business Services
 54   Cutter & Buck
 55   Ciber
 55   Federal Agricultural Mortgage
 55   Funco
 55   Minimed
 59   Colorado Medtech
 59   TSR
 61   Dycom Industries
 61   Modtech Holdings
 63   Sapient
 63   SLI
 65   National RV Holdings
 65   Visio
 67   NBTY
 68   Dayton Superior
 68   Keane
 70   NCS Healthcare
 70   Tellabs
 72   Oshkosh Truck
 73   Quintiles Transnational
 74   Hanger Orthopedic Group
 75   Micrel
 75   Premier Bancshares
 77   Windmere Durable Holdings
 78   Medicis Pharmaceutical
 79   Sylvan Learning Systems
 79   URS
 81   Greater Bay Bancorp
 82   FYI
 82   RCM Technologies
 84   Veritas DGC
 85   Morrison Knudsen
 86   Bluegreen
 87   Schuler Homes
 88   CKE Restaurants
 88   Rexall Sundown
 90   Renal Care Group
 91   Hollywood Park
 91   Impath
 93   Tejon Ranch
 94   El Paso Energy
 94   Resmed
 96   Concord EFS
 97   Dollar Tree Stores
 98   Tech Data
 99   Stewart Information Services
100   Pomeroy Computer Res.

(a) Source: Daniels et al. (1999).
TABLE 1

Descriptive Statistics of Sample Firms (a)

Variable                               n   Mean      25%    Median

Panel A: Firm Characteristics (b)

Market value of common equity        100  2770.7   171.4   512.3
Earnings per share                   100     1.06    0.54    0.87
Return on assets                     100    11.6%    5.5%   10.2%

Panel B: Growth in Firm
Characteristics (c)

Growth in market value of equity     100   228.6%   53.9%  157.1%
Growth in earnings per share          99   569.3%  100.0%  173.2%
Growth in return on assets           100   410.8%  -16.7%   28.0%

Panel C: Option Plan Size (b)

Options outstanding as a % of        100    12.8%    2.2%   12.1%
 common stock
Options exercisable as a % of         99     5.7%    2.7%    4.7%
 common stock
Options granted as a % of common     100     4.7%    4.6%    3.5%
 stock

Panel D: Growth in Option Plan
Size (c)

Growth in options outstanding as a    97    31.6%  -12.0%   10.4%
 % of common stock
Growth in options exercisable as a    61    41.9%  -35.8%    7.8%
 % of common stock
Growth in options granted as a % of   95   138.3%  -50.4%   13.9%
 common stock

Variable                               75%     Std. Dev.

Panel A: Firm Characteristics (b)

Market value of common equity        1351.0   12980.4
Earnings per share                      1.39      0.85
Return on assets                       17.3%      7.1%

Panel B: Growth in Firm
Characteristics (c)

Growth in market value of equity      305.4%   278.0%
Growth in earnings per share          280.0%  1481.3%
Growth in return on assets            165.9%  1734.4%

Panel C: Option Plan Size (b)

Options outstanding as a % of           6.1%     7.1%
 common stock
Options exercisable as a % of           7.0%     5.0%
 common stock
Options granted as a % of common       81.8%     4.0%
 stock

Panel D: Growth in Option Plan
Size (c)

Growth in options outstanding as a     55.1%    87.0%
 % of common stock
Growth in options exercisable as a     48.7%   145.4%
 % of common stock
Growth in options granted as a % of    76.9%   757.5%
 common stock

(a)Market value of common equity is stated in millions of dollars;
earnings per share are diluted earnings per share; return on assets is
net income divided by average total assets, unless beginning total
assets is missing (affects three firms in year 1). In these instances
ROA is computed by dividing net income by total assets at the end of the
year. Options outstanding (exercisable; granted) as a percentage of
common stock outstanding is options outstanding (exercisable; granted)
divided by the number of common shares outstanding as of the end of the
fiscal year.

(b)Panels A and C variables are measured as of the end of the third year
included in the analysis. Only 99 of the 100 firms provide the number of
options exercisable at the end of year 3.

(c)The growth variables in Panels B and D are computed over the
three-year period included in the sample by taking ([x.sub.3] -
[x.sub.1])/absolute value ([x.sub.1]). Growth in earnings per share
cannot be computed for one firm, because the denominator of the ratio is
0. Growth in options granted is missing for five firms because these
firms did not grant options in both the first and third years of the
sample period. Growth in options outstanding is missing for three firms
and growth in options exercisable is missing for 39 firms because these
firms did not disclose this information for year 1 of the sample period.
TABLE 2

Effect of Stock Option Expense on Diluted EPS and ROA Differences
between Reported and Pro Forma Amounts

                                   n    Mean    25%  Median     75%

Panel A: Earnings Per Share (a)

Year 1

  Difference                      97   0.047  0.010   0.030   0.070
  Difference as a % of reported   97  45.38%  3.48%   9.82%  22.65%
   earnings per share

Year 2

  Difference                     100   0.099  0.040   0.070   0.130
  Difference as a % of reported  100  40.17%  4.30%  11.55%  23.36%
   earnings per share

Year 3

  Difference                     100   0.162  0.060   0.120   0.200
  Difference as a % of reported  100  22.87%  5.53%  14.01%  27.01%
   earnings per share

Panel B: Return on Assets (b)

Year 1

  Difference                      97   0.015  0.002   0.009   0.017
  Difference as a % of reported   97  81.30%  3.93%   9.50%  25.25%
   return on assets

Year 2

  Difference                     100   0.022  0.004   0.012   0.024
  Difference as a % of reported  100  35.32%  4.71%  11.57%  23.41%
   return on assets

Year 3

  Difference                     100   0.024  0.005   0.014   0.034
  Difference as a % of reported  100  22.82%  5.72%  13.55%  26.51%
   return on assets

                                 Std. Dev.

Panel A: Earnings Per Share (a)

Year 1

  Difference                         0.049
  Difference as a % of reported    160.58%
   earnings per share

Year 2

  Difference                         0.097
  Difference as a % of reported    164.92%
   earnings per share

Year 3

  Difference                         0.175
  Difference as a % of reported     38.17%
   earnings per share

Panel B: Return on Assets (b)

Year 1

  Difference                         0.022
  Difference as a % of reported    360.93%
   return on assets

Year 2

  Difference                         0.030
  Difference as a % of reported    116.70%
   return on assets

Year 3

  Difference                         0.023
  Difference as a % of reported     36.97%
   return on assets

(a)Earnings per share are diluted earnings per share. When diluted
earnings per share adjusted for stock option expense (i.e., pro forma
earnings per share) is not disclosed (one firm in year 2 and one firm in
year 3), we substitute pro forma basic earnings per share. In year 1,
the number of firms for which we have both reported and pro forma
earnings per share is 97, as three firms did not report either basic or
diluted pro forma earnings per share.

(b)Return on assets is net income divided by average total assets,
unless beginning total assets is missing (affects three firms in year
1). In these instances ROA is computed by dividing net income by total
assets at the end of the year. In year 1, three firms did not report pro
forma net income, which reduces the number of firms for which we can
calculate a difference in return on assets to 97.
FIGURE 1

Comparison of After-Tax Charge to After-Tax Fair Value (a)

Millions of Dollars

                                 Year 1  Year 2  Year 3

Estimated fair value of options   2.6     3.7     5.8
granted
Reported stock option expense     0.7     1.6     3.1
Difference                        1.6     2.1     2.8

(a)Value presented are the median amounts for each year, in millions of
dollars. The reported after-tax charge is the difference between
reported net income and pro forma net income. The estimated after-tax
fair value is the product of (1) the per-option weighted-average fair
value of options granted during the year; (2) the number of options
granted during the year; and (3) 1 minus the assumed 38 percent tax
rate.

Note: Table made from bar graph
TABLE 3

Statistics on Disclosure Policy and Estimate Choice (a)

Panel A: All Firms (100 firms)

                                                      Disclosed
                                                     Yes      No


Pro forma basic earnings per share provided           72      28
Pro forma diluted earnings per share provided         99       1
Number options exercisable at end of year disclosed   99       1
Panel B: Firm with Options Granted in Third Year (98 firms)

                                Range  Point  Not Disclosed

Volatility estimate               7     89          2
Risk-free rate estimate          17     81          0
Life expectancy estimate         10     86          2
Dividend yield estimate           1     90          7
Weighted-average fair value of    1     90          7
 options granted
                                                         No
                                 Black-                options
                       Binomial  Scholes  Unspecified  granted

Option-pricing method     1        96          1          2

(a)This table reports firms' disclosures of year 3 data only.
FIGURE 2

Firm Disclosure Policies (a)


Full compliance                               88%
Missing fair value and option pricing method   1%
Missing estimates                              4%
Missing number of exercisable options          1%
Missing fair values                            6%

(a)The reported statistics are for the third year for the 100 firms
included in our sample. Firms in full compliance disclosed the following
items for the third sample year: (1) pro forma basic or diluted earnings
per share: (2) option-pricing method used; (3) estimates of volatility,
risk-free rate, and life of the options; (4) fair value of options
granted; (5) number of options exercisable at the end of the year. Firms
that fail to disclose estimates or volatility, risk-free rate, or life
of the options are labeled "missing estimates." Firms that fail to
report either the fair value of the option granted or the number of
options exercisable at the end of the year are labeled "missing fair
values" and "missing exercisable," respectively. The firm that failed to
disclose both the fair value of the options granted and the
option-pricing method is labeled "missing fair value and options-pricing
method."

Note: Table made from pie chart