Is Diluted EPS Becoming More Art Than Fact?

By A. Bruce Caster, Raymond J. Elson, and Leonard G. Weld

E-mail Story Print Story

EPS and Management

Because of current tax law, reported EPS can have a significant impact, directly or indirectly, on the wealth of many top corporate executives. IRC section 162(m) limits tax-deductible compensation for covered employees (including, specifically, chief executive officers) to $1 million, unless that compensation is tied to performance goals. Thus, a large share of the total compensation to highly paid executives is based on achieving performance goals, and many of the commonly used performance goals are related to EPS.

A survey of 177 publicly traded U.S. corporations (K. Murphy, “Executive Compensation,” Handbook of Labor Economics, Vol. 3, O. Ashenfelter and D. Card, eds., North Holland, 1999) found that approximately one-third of the respondents tie executive compensation directly to EPS or EPS growth. Other common performance-based plans, particularly those involving stock options, tie executive compensation to changes in share price.

If a performance-based executive compensation plan is tied directly to EPS, accounting rules that change EPS, even by just a few cents, will directly affect the wealth of a company’s executives. Because reported EPS also affects share price, accounting rules that affect EPS will also have at least an indirect effect on executives whose compensation plans are tied to stock price.

The result is tremendous pressure on top executives to ensure that the company regularly meets its EPS targets, which could manifest itself in pressure to “manage” the components involved in the EPS computation. The challenge is to raise capital while controlling the number of shares of common stock issued. Sometimes that pressure can also be felt by a company’s board of directors. In April 2006, the Coca-Cola Company announced a share-based compensation plan for its board of directors under which the directors will receive share grants equal to a flat fee of $175,000. The grants will be payable in cash in three years provided the company increases earnings per share 8% each year. If the company fails to meet its EPS targets, the directors will go without pay [M. Credeur, “Coca-Cola Ties Director Pay to Company’s Performance (Update3),” Bloomberg.com, April 5, 2006].

There is a long history of allegations that companies have used devices like aggressive revenue-recognition policies, discretionary accruals, and adopting or changing accounting principles in order to be able to report the desired amount of earnings each period (L. McGrath and L. Weld, “Case Histories of Fraud and Abusive Earnings Management,” Ohio CPA Journal, April–June 2002). Recent academic research has even provided evidence that some companies may manipulate earnings in such a way that the EPS calculation will be rounded up to the next penny (S. Das and H. Zhang, “Rounding-up Reported EPS, Behavioral Thresholds, and Earnings Management,” Journal of Accounting & Economics, April 2003).

Similarly, companies can engage in a wide variety of activities that alter EPS by changing the number of common shares outstanding during the period. For example, share repurchases are a legitimate method of reducing shares outstanding, and such a decision will increase EPS. Many companies have active share-repurchase programs:

• General Mills reported an increase in after-tax income from $183 million to $252 million, and EPS jumped from $0.45 to $0.64. Average diluted shares outstanding dropped 3% because of an active repurchase program, and this accounted for $0.02 of the $0.64 EPS.
• Clear Channel Communications reported net income and EPS of $253.8 million and $0.41, respectively, in the second quarter of 2004. In the second quarter of 2005, net income fell to $220.7 million, but EPS only dropped to $0.40. An active share-repurchase program accounted for a $0.04 EPS increase; without share repurchases, 2005 EPS would have been only $0.36.
• AutoZone reported that net income increased 3.1% in its third quarter of 2005, but EPS benefited from share repurchases and increased 10.5%, from $1.68 to $1.86. Since 1998, AutoZone has repurchased 85.8 million shares, at a cost of approximately $4 billion.

Despite investors’ expectations that share repurchases will always reduce total shares outstanding, it appears that not all companies’ share-repurchase programs are designed for that purpose. For example, Intel spent $42 billion between 1990 and 2004 repurchasing 2.2 billion shares, but the number of Intel shares outstanding in 2004 (split-adjusted) was about the same as it was in 1990. Over the three years ending June 30, 2005, Microsoft repurchased 674 million shares, but it also reported issuing 666 million new shares over that same period. In the end, Microsoft spent $6.6 billion (after tax) to reduce its outstanding shares by 8 million, or 0.08%. BusinessWeek (D. Henry, “The Dirty Little Secret About Buybacks,” January 23, 2006) cited a study of the companies in the S&P 500 that showed they collectively had spent $197 billion in 2004 on share buybacks, yet the total number of shares outstanding in those companies increased 1.8% for the year. Only one-third of the companies reported actual reductions of at least 1% in the number of shares outstanding.

The Wall Street Journal (M. Maremont and S. Ng, “Stock Buyback Now May Spur a Big Bill Later,” January 31, 2006) reported that some companies are starting to accomplish their buybacks through “accelerated share repurchase” (ASR) programs. At least 30 major corporations, including DuPont, Northrop Grumman, Sara Lee, and Duke Energy, have engaged in ASRs in the last year.

In an ASR, a company buys back a large block of shares by purchasing them from an investment bank, which, in turn, usually borrows the shares from its institutional customers. Once the ASR has been effected, the company can announce that it has immediately repurchased a significant number of shares. And it can report the immediate increase in EPS—and can expect to receive the immediate boost in share price—that normally accompanies such a buyback.

Unfortunately, the company must later complete the transaction by purchasing enough shares in the market to repay the loans made by the investment bank’s institutional customers, and it must do so at the higher share price created by the buyback announcement. One thing that ASRs clearly demonstrate is that the ability to announce an immediate increase in EPS may be sufficiently important to managers that they are willing to engage in complex, risky transactions in order to accomplish that goal.

Other Financial Instruments and Diluted EPS

Share repurchases, by whatever means they are accomplished, represent one strategy for increasing EPS by minimizing the number of common shares outstanding. Another strategy is simply to avoid issuing additional common shares.

Companies have found ways to issue financial instruments that provide the investor with most of the advantages of common stock, without actually issuing common stock. Some long-standing examples are convertible securities (convertible debt and convertible preferred stock) and stock options. Because these instruments do not represent additional common shares, they do not automatically impact EPS. But because they can readily be turned into additional common shares in the future, financial statement users must be able to judge their potential impact on future EPS.

The response of the accounting profession to these sorts of securities has been to provide a second EPS computation, diluted EPS. Basic EPS is a historical report, based on events that really did occur in the reporting period. Diluted EPS is a pro forma presentation, showing the potential effects of events that did not occur but that easily could occur in future periods.

The unfortunate limitation of diluted EPS is that it is based on a fixed set of assumptions regarding the specific categories of financial instruments that must be included in its computation. That limitation has provided intriguing opportunities for companies wishing to design new financial instruments that have both desirable economic characteristics and desirable effects on the computation of diluted EPS. Among the instruments that might be categorized in this way are contingent convertible debt and freestanding financial instruments.

Contingent Convertible Debt

The distinguishing feature of convertible debt is that debt holders can exchange this type of security for some other type of security, generally a fixed number of shares of common stock. Contingent convertible (CoCo) debt contains a conversion contingency that must be met before the debt holder is permitted to convert the debt to common stock. The most common conversion contingency is a price premium, though other sorts of contingencies are possible. A price-premium contingency requires that the market price of the common stock exceed the conversion price by some predetermined amount—often 10% to 20%—before the debt becomes convertible.

By requiring that an additional condition be fulfilled before the debt becomes convertible, the contingency delays the dilution of voting control and the decline in market value that typically result from conversion. Additionally, the contingency increases the probability that the corporation will be able to call the debt and never have to issue new shares, thereby avoiding these two problems.

Recent academic research (C. Marquardt and C. Wiedman, “Earnings Management Through Transaction Structuring: Contingent Convertible Debt and Diluted Earnings per Share,” Journal of Accounting Research, May 2005) suggests that voluntary conversions of CoCos are exceedingly rare. Most companies that issue CoCos call them before they become convertible, and conversions generally occur only when the CoCo contains a provision that allows conversion when the securities are called, even if the conversion contingency has not yet been satisfied.

One further advantage of CoCo debt is that it may receive favorable treatment in the computation of diluted EPS. Ordinary convertible debt is immediately incorporated into the diluted EPS computation. Until recently, however, CoCo debt was considered to be the equivalent of contingently issuable shares, and it was not included in diluted EPS until the contingency (the price premium) was satisfied.

That changed in November 2004, when EITF 04-8, “The Effect of Contingently Convertible Debt on Diluted Earnings per Share,” declared that CoCos whose conversion triggers are based solely on market price should be treated exactly the same way as any other convertible debt in computing EPS; that is, they should immediately be included in the computation of diluted EPS. Because EITF 04-8 only applies to CoCos with a market-price contingency, CoCos whose conversion triggers are not based on market price are still excluded from diluted EPS. The result is that companies can still use this low-yield convertible debt while avoiding the earnings dilution that results from issuing ordinary convertible securities.

It also appears that the favorable treatment of CoCos in the computation of EPS is a major consideration in financing decisions. The above study by Marquardt and Wiedman reported evidence that companies whose “CEOs’ compensation is more dependent on earnings results are significantly more likely to issue CoCos.”

Unresolved Questions Regarding Financial Reporting of CoCos
The question of non–market-based contingencies is only one of several issues related to CoCos that were not resolved by EITF 04-8. Other questions include how to account for CoCos that become convertible, but only temporarily, once the necessary conditions for conversion have been satisfied. For example, some existing CoCos contain provisions that may cause the debt to be convertible—

• on the next day only, based on the price the previous day (and not on days after that);
• only during specified quarters of the year; or
• some combination of the above.

Both SFAS 128 and EITF 04-8 primarily focus on instruments that become convertible, and then continue to be convertible, once the conversion contingency is satisfied. There is currently no authoritative guidance regarding how “temporarily convertible” instruments such as the ones described above should impact diluted EPS.

Freestanding Financial Instruments

A freestanding financial instrument is a financial instrument that is entered into separately and apart from any of the entity’s other financial instruments or equity transactions, or an instrument that is entered into in conjunction with some other transaction and is legally detachable and separately exercisable. SFAS 150, Accounting for Certain Financial Instruments with Characteristics of Both Liabilities and Equity, describes three main classes of freestanding financial instruments:

• Mandatorily redeemable instruments. These instruments must be redeemed by transferring assets on a fixed or determinable date, or upon an event certain to occur, and redemption is outside the control of the issuer and holder. Examples of mandatorily redeemable instruments include stock to be redeemed upon the death of the holder, trust-preferred securities that are redeemed upon maturity of debentures or other loans, and conditional redeemable stock that must be redeemed upon the occurrence of an event (e.g., change of control).
• Instruments with repurchase obligations. These instruments embody an obligation to repurchase an issuer’s equity shares, and the obligation may require settlement by transfer of assets. Examples of instruments with repurchase obligations include written put options and forward purchase contracts that require physical or net-cash settlement.
• Instruments with obligations to issue a variable number of shares. These instruments embody an obligation that the issuer settle by issuing a variable number of its equity shares if the monetary value of the obligation is based solely or predominately on 1) a fixed monetary amount known at inception, 2) variations in something other than the fair value of the issuer’s equity shares, or 3) variations inversely related to changes in the fair value of the issuer’s equity shares. Examples of instruments with obligations to issue a variable number of shares include: 1) debt settled with a variable number of the issuer’s equity shares; 2) instruments indexed to the S&P 500 and settled with a variable number of the issuer’s equity shares; 3) written put options that can be settled by one party delivering stock equal to the current fair value of the counterparty’s gain; and 4) forward purchase contracts that are net-share settled.

SFAS 150 requires that the above freestanding financial instruments be classified as liabilities or assets under some circumstances, even though they have characteristics of equity. These instruments are initially measured at fair value; however, freestanding “physically settled” forward contracts that obligate the issuer to purchase a fixed number of its own shares for cash are initially measured at the present value of the amount to be paid at the settlement date. All instruments (with a few exceptions) are subsequently measured at fair value, with changes in the liability flowing through the income statement.

Common shares that are subject to physically settled forward purchase contracts or are mandatorily redeemable are excluded from both basic and diluted EPS calculations. At the same time, amounts (including contractual dividends and participation rights in undistributed earnings) attributed to shares that are to be redeemed or repurchased and not recognized as interest costs should be deducted from income available to common shareholders. Among the reasons cited by FASB for this conclusion was that the accounting for physically settled forward contracts reduces equity even though the shares are still outstanding. Therefore, the shares are effectively accounted for as if retired and should not be treated as outstanding in EPS calculations.

EITF 04-8 specifically addressed the effect of contingent convertible instruments on diluted EPS. The EITF is considering whether to address the effect of non–market-based contingencies for issued instruments and market-based contingencies for freestanding instruments on the calculation of EPS.

Observations and Recommendations

When one steps back from the gritty details of various categories of securities that may or may not be included in the computation of diluted EPS, one starts to notice a pattern. Creative minds develop new and innovative securities that are designed to raise capital while minimally affecting diluted EPS—at least under the accounting rules existing at the time. This leads FASB to reexamine the computation of diluted EPS, to make sure these new securities are being handled appropriately. When those securities no longer enjoy favorable treatment in the computation of diluted EPS, they are replaced by other, even more creative types. Such has been the story of CoCo debt securities, and it will undoubtedly be the story of other securities yet to be invented.

What is missing is guidance from FASB regarding how to continue improving the reporting of diluted EPS as new securities are developed over time. Also missing is a definition for diluted EPS.

That is not to say that no definition currently exists. The current definition of diluted EPS describes a long, complex computational process, and it defines diluted EPS as being the result that one obtains at the end of the computation. Using the jargon of current accounting debates, this definition is completely rule-bound and essentially lacks any conceptual basis. It also seems oddly out of step with the other definitions that FASB has provided for the accounting profession.

Statement of Financial Accounting Concepts (SFAC) 6 takes great pains to define the elements of financial reporting in terms of real circumstances that exist in the business world. That statement asserts that the elements reported in the financial statements are supposed to reflect those external realities, and it carefully describes which external reality each element is supposed to reflect. For example, assets are “probable future economic benefits obtained or controlled by the enterprise,” and liabilities are “present obligations of the enterprise to transfer assets or perform services to another entity in the future.”

This sort of definition can be used to judge the quality of a proposed accounting standard. One can ask, “If we adopt this standard, will the items reported in the financial statements do a better job of reflecting the realities that they are supposed to represent?” And readers of financial statements understand what they are supposed to be seeing in those statements.

The current definition of diluted EPS does none of that. It does not identify any particular external reality; it only describes the process by which one performs a computation. As such, it provides no basis for judging whether the current method of determining diluted EPS is appropriate, and provides no guidance on how to incorporate new financial instruments into the reporting of diluted EPS over time. It simply tells an accountant what to do, without saying anything about what it is trying to accomplish.

The authors recommend that FASB provide a conceptual, or “principles-based,” definition of diluted EPS. Such a definition could form a basis for significant improvement, both in the process for determining diluted EPS and also in users’ understanding of its purpose. Furthermore, the authors fear that if the profession continues to operate under the current definition, diluted EPS may eventually come to be viewed as simply an arbitrary result of computational art, not an attempt to report a useful fact.