Earnings per share (EPS) is considered by some users to be the single most important item in the financial statement. Frequently cited in the financial press, EPS is the proverbial "bottom line" indicator of financial performance. EPS numbers are also used to calculate other commonly used ratios such as price/earnings and dividend payout. The importance attached to EPS by financial statement users probably results, in part, from the fact that it is disclosed in the financial statements of public companies and subjected to the scrutiny of auditors.

At the same time, EPS is troublesome to preparers and auditors of financial statements. The complex provisions of APB 15 and a host of amending pronouncements are a prime example of standards overload. The burden of EPS reporting is worsened by SEC regulations requiring that 10-K reports include a supplementary schedule explaining the computation of EPS whenever it is not apparent in the financial statements. For many public companies, this is the usual, case.

The FASB's 1978 decision to exempt nonpublic companies from costly EPS reporting requirements is an example of the "differential GAAP" response to standards overload. Another possibility would be to simplify reporting requirements for all companies. This article proposes that EPS reporting is a good candidate for such a simplification and highlights some of the problems with current EPS reporting standards. it also proposes an alternative reporting format that would eliminate the need to provide supplementary disclosures in 10-K reports.

Problems with EPS Reporting Standards

Most financial statement readers are familiar with a simple formula for calculating EPS. Earnings available to common share-holders, generally net income minus any preferred dividend requirement, is divided by the weighted average number of common shares outstanding during the year. The resulting statistic is called basic (or simple) EPS. Basic EPS is presented on the income statement when a company has no outstanding securities that are convertible into common stock. When convertible securities are outstanding, more complex rules take effect. These rules are intended to make EPS reflect the potential of convertible securities to dilute earnings available to common shareholders. However, a growing body, of thought suggests that the intricate manipulations mandated by APB 15 do little to improve the basic EPS statistic.

Historical Versus Hypothetical Disclosures. Arguably, the most significant problem with current standards is that basic EPS is subject to replacement on the income statement by two hypothetical EPS numbers. One of those numbers, primary, EPS, is computed assuming that common stock equivalents, those dilative securities most similar in substance to common stock, were converted to common stock on the first day of the reporting period. The second number, fully-diluted EPS, is computed assuming that all dilative securities were converted. This computational approach is called the "if-converted" method because it produces an estimate of what EPS would be if holders of dilative securities exercised their rights to obtain common stock. When fully diluted EPS is less than 97% of basic EPS, primary, and fully diluted EPS are the only per share disclosures required under current accounting standards. In other words, historical per share disclosures are completely omitted when potential dilution is material.

Common Stock Equivalents ana Primary EPS. Concerns about the usefulness of dual EPS reporting are not limited to the relative merits of historical and proforma disclosure. Some of the specific rules governing the computation of primary and fully diluted EPS have also been questioned. The test used to identify common stock equivalent securities is among the most controversial of those rules. Accounting researchers have devoted considerable attention to determining the usefulness of this test and a variety of alternatives.

Originally termed "residual securities," common stock equivalents are convertible securities that derive a major portion of their value from the fact that they can be exchanged for common stock. Current standards mandate that convertible securities whose effective yield is less than two-thirds the Aa corporate bond yield be classified as common stock equivalents. Empirical studies have consistently failed to establish that the yield test, or any of its competitors, are good predictors of conversion to common stock.1 Inability to pinpoint the meaning of common stock equivalency, makes it difficult to describe the difference between primary and basic EPS.

That difficulty, is increased by the fact that potentially dilative securities are evaluated and permanently classified as of their issue date. In other words, "once a common stock equivalent, always a common stock equivalent." Changing relationships between the terms of particular securities and prevailing market conditions are not considered in the computational rules for EPS. Common sense suggests that the attractiveness of a conversion feature relative to other security characteristics will vary, in different circumstances. Systematically disregarding that possibility raises furcher questions about the meaning of common stock equivalent status and, thus, primary EPS.

Options, Warrants, and the Treasury Stock Method. Another questionable aspect of EPS reporting is the treasury stock method of accounting for option and warrant proceeds. The mechanics of the treasury stock method can be understood by comparing the financial statement effects of stock and bond conversions with those resulting from the exercise of options or warrants. Assumed conversions of preferred stock and convertible bonds have both numerator and denominator effects on diluted EPS. In an assumed conversion of preferred stock, for example, dividends that would be avoided are added back to income available to common shareholders (the numerator). Shares that would be issue are added to weighted average common shares outstanding (the denominator). The net-of-tax interest savings and shares issuable in an assumed bond conversion are treated similarly.

When options or warrants are exercised, common shares are issued when the instruments are surrendered along with a cash payment. Unlike stock or bond conversions, these transactions have no effect on income available to common shareholders, the numerator of EPS. To recognize the positive cash flow effect of an assumed option or warrant exercise, accountants compute diluted EPS assuming that the cash proceeds are used to acquire treasury shares. For primary EPS computations, treasury stock purchases are assumed to be made at the average market price for the year. For fully diluted EPS, the treasury stock method uses the higher of the average or ending market price. In both cases, the treasury shares obtainable are subtracted from those that would be issued upon exercise. Thus, only the "net issuance" is reflected in the denominator of EPS.

The treasury stock method is modified when the number of treasury shares obtainable exceeds 20% of the common shares outstanding. In such cases, funds in excess of the amount needed to acquire 20% of the outstanding common shares are applied to hypothetical retirements of debt and purchases of government securities. This practice generates additions to the numerator of EPS in the form of assumed interest savings or earnings.

The treasury stock method dampens the dilative impact of options and warrants in both the primary and fully diluted EPS calculations. How this device improves the usefulness of per share data is unclear. The accounting literature is devoid of evidence that managers actually use option and warrant proceeds for any of the purposes assumed in accounting standards. The effect of the treasury stock method on financial statement users' decisions is similarly untested. The treasury stock method is, however, partially responsible for our final criticism of EPS reporting.

Reporting the Range of potential Dilution. The dual EPS reporting format (primary and fully diluted) is an attempt to provide information about the potential dilution of earnings through common stock issuances. In fact, however, possible dilution levels span a range from no assumed dilution, the actual situation at balance sheet date, to maximum assumed dilution, the situation that would have existed had all potential diluters been converted on the first day of the year.

If dilative common stock equivalents are outstanding, however, then primary EPS is an undetermined point in the range of potential dilution. Likewise, if options or warrants are outstanding, then the treasury stock method is applied and fully diluted EPS ceases to be the maximum dilution endpoint. Because these two conditions frequently prevail, many public companies' per share disclosures consist of two discrete points within the potential dilution range. Neither the endpoints of the dilution range nor the information needed to calculate EPS assuming various dilution scenarios are required disclosures under generally accepted accounting principles.

Do these complicated EPS disclosures assist investors and creditors in decision making? The answer is unclear. Millar, et. al., report in the September-October, 1987 Financial Analysts journal that basic EPS and cash flow per share numbers are more closely related to stock returns than either primary or fully diluted EPS. Additional research is needed to ascertain the relative usefulness of basic, primary and fully diluted EPS.

In the mean time, accountants and auditors are saddled with complicated rules of questionable usefulness. Survey results published in the February 1979, CPA Journal by Boyer and Gibson suggest that significant numbers of academics and practitioners were unable to correctly answer questions about the rules at that time. We believe that until and unless the weight of the evidence demonstrates the usefulness of complicated schemes, simpler rules should govern EPS reporting.

Simplifying EPS Reporting

The questionable aspects of EPS reporting previously highlighted result largely from the incorporation of untested or unproven assumptions into accounting standards. Simplifying EPS calculations and eliminating those assumptions should partially alleviate the standards overload problem without sacrificing useful information. We propose four specific amendments to achieve that objective.

Replace Primary EPS with Basic EPS.

The importance of anticipating equity dilution when making investment decisions is self-evident. We believe that the merits of presenting historical data, such as basic EPS, are equally apparent. The remaining question is why discard primary EPS rather than make it part of a three- level disclosure?

Research findings have repeatedly demonstrated that common stock equivalency testing, the cornerstone of primary EPS computations, does not produce interpretable results. This result appears to hold for a variety of testing schemes. At least one empirical comparison of basic and primary EPS supports the assertion that primary EPS is merely a garbled transformation of basic EPS. From a preparer's perspective, primary EPS is a burdensome exercise in applying arbitrary rules. From a user's perspective, primary EPS is a noisy statistic, based on questionable assumptions. We recommend that its presentation be discontinued.

Eliminate the Treasury Stock Method from Fully Diluted EPS. Two difficulties with the treasury stock method have already been discussed. First, the assumption that option and warrant proceeds are used to acquire treasury shares is unverified. Expanding plant facilities, building up inventories, and funding pension obligations are all equally plausible uses for cash inflows. Second, the treasury stock method dampens the dilative effects of assumed option and warrant exercise. Thus, fully diluted EPS does not present the maximum dilution scenario.

Discarding the treasury stock method addresses both of these concerns. Allowing the full dilative impact of options and warrants to enter into diluted EPS calculations is equivalent to assuming that option and warrant proceeds are held as a cash balance. This treatment makes fully diluted EPS a disclosure of the maximum dilution possible at financial statement date. Also, speculation about alternative uses for option and warrant proceeds is eliminated from the financial reporting process.

Discontinue the Three Percent Test for Dual Presentation. The apparent purpose of the three percent materiality test is to minimize the EPS reporting burden when potential dilution is trivial. An obvious problem with all quantitative materiality standards is establishing the level at which disclosure ceases to affect decisions. Researchers have not established a universal materiality standard, nor are they likely to. Similarly, there is no evidence that three percent dilution is the point at which financial statement readers become concerned with the effects of convertible securities.

The test is, thus, unsupported. We recommend that companies with simple capital structures present basic EPS, and that those with complex capital structures present basic and fully diluted EPS. Financial statement readers can then make their own judgments about materiality.

Present Information About Individual Dilative Securities. Presenting detailed information about the dilution potential of individual securities completes the EPS disclosure picture. Access to such information permits users to make their own assumptions about conversion timing and likelihood. The full range of EPS figures, from no assumed dilution, to maximum assumed dilution, is thus available to those who wish to perform a few simple calculations.

We believe that financial reporting and financial analysis are separate roles. Eliminating questionable analytical techniques from EPS calculations would allow accountants to concentrate on providing historical information, the activity in which they enjoy a comparative advantage. Financial statement users would benefit from having access to clearly defined, interpretable statistics, such as basic EPS, and raw data needed to perform additional analyses. Comparing the proposed disclosure format to current practice demonstrates that only minor adjustments would be necessary.

Implementing the Proposed Changes

While the resulting disclosure format is only one of many possibilities, it illustrates a shift from selective, hypothetical disclosure to comprehensive, historical disclosure. We also believe that such a format would satisfy the SEC's 10-K reporting requirements for EPS. As previously stated, much of the information needed to prepare the EPS disclosures proposed in this article is already available in the 10- K reports of public companies. SEC regulations require a detailed explanation of EPS computations unless those computations are obvious in the registration statement or financial report. The SEC regulation does not provide specific guidance for the preparation of this schedule, but the schedule presented in Exhibit 1, which is excerpted from the 1988 financial statements and Form 10-K of the Bethlehem Steel Corporation, is representative of a typical disclosure.

Exhibit 2 illustrates the proposed changes using the same data. Aside from format, the exhibits differ in two principal ways. First, the modified disclosure does not present primary EPS. In the case of Bethlehem Steel, stock options are the only common stock equivalents. Their impact on primary EPS is very small, in part due to the mitigating effect of the treasury stock method. The basic EPS figure in the modified disclosure is, thus, only a penny higher than primary EPS reported in the 10-K. In most of the reports examined by the authors, this difference was very small.

The second difference results from the elimination of the treasury stock method from the modified disclosures. Bethlehem's 1988 10-K reports the net dilative effect of options on primary and fully diluted EPS (141,000 and 174,000 shares, respectively). These assumed issuances are net of assumed treasury stock acquisitions at the average price for primary EPS and at the higher of ending or average price for fully diluted EPS. The modified disclosure does not present primary EPS, so the effect of the options on this calculation are irrelevant. Fully diluted EPS in the modified disclosure reflects the full dilative effect of the options. In this case, options on 1,833,000 shares were outstanding and exercisable at the end of 1988. The change lowers fully diluted EPS from $4.91 to $4.81.

Exhibit 2 illustrates only one of many possible EPS reporting formats, but it does incorporate the four modifications proposed in this article. First, primary EPS, a statistic of questionable value, is eliminated. Second, the full dilution potential of options and warrants is disclosed. Third, although it does not apply in this case, the three percent dilution test is eliminated. Fourth, information is provided to facilitate the calculation of diluted EPS under a variety of assumptions about conversion timing and likelihood. The proposed disclosure also appears to satisfy the SEC requirements that the EPS computation be evident.

End Twenty Years of Complexity

The principal authoritative guidance for EPS reporting was issued over twenty years ago. The complicated rules set forth in APB 15 have been criticized because they are costly to implement and because they result in statistics that are difficult to interpret. The authors believe that amending current standards would benefit both preparers and users of financial statements. Simpler rules for EPS reporting would partially alleviate the standards overload problem faced by preparers and auditors. Comprehensive, historical disclosures would enable financial statement readers to obtain a clearer picture of the potential for earnings dilution through conversions into common stock.

Of course, responsibility for reconsidering EPS reporting rests with the FASB. Several factors must be considered before an issue is formally addressed by the FASB. These include: 1) the pervasiveness of the problem, 2) the likelihood of developing alternatives to improve financial reporting, 3) the technical feasibility of alternative solutions, and 4) the practical problems associated with achieving general acceptance for a solution. EPS reporting appears to be a good candidate in terms of all four factors. Whether the issues raised in this article will ultimately be addressed by the FASB depends largely on the interest shown in them by individuals and groups such as the readership of The CPA Journal We hope that further discussion among preparers and users will encourage the FASB to reconsider EPS reporting and require a more comprehensive historical disclosure of per share amounts.

(1)Seeforexample:

Arnold,DonaldF.,andThomasE.Humann,

"EarningsPerShare:AnEmpiricalTestofthe

MarketParityandtheinvestmentValueMeth

-ods,"TheAccountingReview,January,1973,pp.

23-33.

Frank,Werner.,andjerryJ.Weygandt,"A

PredictionModelforConvertibleDebentures,

JournalofAccountingResearchSpring,1971,

pp.116-26.

Fulmer,Jr.,John,andJamesE.Moon,Tests

forCommonStockEquivalency,"Journalof

Accounting,Auditing,andFinance,Fall,1984,

pp.5-14.

Givoly,Dan,andDanPalmon,"Classification

ofConvertibleDebtandCommonStockEquivalents:

SomeEmpiricalEvidenceontheEffects

ofAPBOpinion15,"JournalofAccounting

Research,Autumn,1981,pp.530-543.

Hofstedt,ThomasP,,andRichardR.West,

"TheAPB,Yieldindices,andPredictiveAbility,"

TheAccountingReview,April,1971,pp.329-337.