U.S. Securities & Exchange Commission
SEC Seal
Home | Previous Page
U.S. Securities and Exchange Commission

Speech by SEC Staff:
Remarks Before the 2006 AICPA National Conference on Current SEC and PCAOB Developments


Cathy J. Cole

Associate Chief Accountant, Office of the Chief Accountant
U.S. Securities and Exchange Commission

Washington, D.C.
December 11, 2006

I plan to discuss several issues related to the two-class method of computing earnings per share. However, before I get started, I ought to explain why I decided to talk about this topic. There seems to be some concern among my colleagues that a general audience, even of accountants, might not share my enthusiasm for this topic. With that in mind, I want to assure you that the issue of computing and presenting EPS using the two-class method affects more companies than you might think. And, in my opinion, the two-class method is an under-discussed topic in accounting. So, I want to do my part to make up for the discussion shortfall.

Now, on this topic I have good news and bad news… since the staff did agree with a registrant in one case but did not agree with a registrant in another case. Well, I will give you the bad news first. Then you will have the good news to look forward to.

The first situation relates to diluted earnings per share (EPS) when one class of common is convertible into another class of common. FASB Statement 128 requires the use of the two-class method of computing EPS in certain circumstances, namely, when there is more than one class of common stock and the classes have different dividend rates or when there are other securities that have a right to participate in dividends with common stock.

You may recall, the two-class method has been around for years and was carried forward from APB Opinion 15 into Statement 128. As with APB 15, Statement 128 has only a modest amount of computational guidance on the mechanics of the two-class method. The general guidance is first to reduce income by dividends declared to arrive at undistributed earnings. Then undistributed earnings are allocated to common stock and participating securities. Finally, the total of distributed and undistributed earnings allocated to each class of common is divided by outstanding shares to arrive at the EPS number for each class. While that sounds simple enough, there is little guidance in the statement itself on how to go about allocating undistributed earnings between common stock and participating securities.

Now, that's where EITF Issue 03-6 comes in. The Emerging Issues Task Force provided guidance on the allocation of undistributed earnings in EITF Issue 03-6. This consensus clarifies that allocations to participating securities under the two-class method should be based on the contractual participation rights to share in current earnings under the assumption that all of the earnings for the period are distributed. It also indicates that the allocation should not be based on arbitrary assumptions, but instead upon objectively determinable, nondiscretionary participation rights.

With this background in mind, several two-class method issues were recently brought to our attention by the staff in the Division of Corporation Finance. While there may be a variety of reasons for companies having more than one class of common stock, it is often the case that one class may be entitled to higher dividends than the other while having lower voting rights. It is also frequently the case that one class of common stock is convertible into another.

As an example, a company may have two classes of common stock, say Class A and Class B. Class A may be entitled to dividends at a rate that is 1.2 times that of Class B, but Class B may have twice the voting rights of Class A. Class B may also be convertible at any time into Class A on a 1 to 1 basis.

Statement 128 states that participating securities convertible into a class of common stock should be included in EPS on an if-converted basis. EITF Issue 03-6 clarifies that for participating securities, the "if-converted" guidance should be applied only for diluted EPS. Neither Statement 128 nor EITF 03-6 directly address how to handle a class of common stock that is convertible into another class of common stock. Further, it is not clear whether the guidance in EITF Issue 03-6 on participating securities is applicable to multiple classes of common stock, as opposed to securities that participate with common stock, but that are not themselves common stock.

So, where is this all leading? Well, the staff believes that for diluted EPS, a company with two classes of common stock must actually present both a basic and diluted earnings per share number for each class of common stock regardless of conversion rights. One might ask, "If for diluted EPS, I have assumed the conversion of one class into the other, how can I then present two diluted earnings per share numbers, one for each class? Doesn't assuming the conversion of one class into the other essentially eliminate the second class for diluted EPS purposes?" Well, for the class being converted into (Class A in the example), we believe diluted EPS should be computed using the if-converted method for the convertible class (Class B), if doing so would be dilutive. Diluted EPS for the convertible class (Class B in the example) should be computed using the two-class method. Diluted EPS must be presented for both classes.

You may ask, "What is the point of this second computation?" Why can't the company just present diluted EPS for Class A and be done with it? Well, there are several reasons for the presentation of diluted EPS for Class B. The first reason is that while presenting diluted EPS for Class A assuming conversion of Class B may be more dilutive for Class A, the assumed conversion might in fact be antidilutive for Class B. In our simple example, if Class A is entitled to $1.20 of dividends for every $1.00 of dividends for Class B, then if Class B is allowed to share equally in earnings with Class A, then Class B would be entitled to a higher share of earnings on a diluted basis assuming conversion, than it would be entitled to absent the conversion. Further, Class B shareholders might be interested in knowing what diluted EPS would be for Class B shares, if the shares were not converted into Class A. This might be especially true in circumstances where potential common shares would also affect the computation.

The second reason for presenting EPS for Class B is that presenting diluted EPS only for Class A is inconsistent with the guidance of Statement 128, paragraph 61(d), which that says that basic and diluted EPS data should be presented for each class of common stock. Thus, if there are two classes of common stock that means two sets of EPS data. I just think of it as double the numbers…double the fun. Obviously though, the company should also provide the appropriate footnote disclosures explaining its computation of basic and diluted EPS.

Now on to another two-class issue. The facts in the next case were as follows. There were two classes of common stock, let's call these A and B. Class A was entitled to one vote per share; Class B was entitled to ten votes per share. The company's articles of incorporation stated that dividends could be paid on Class A without an equal or any dividend being paid on Class B. However, dividends could not be paid on Class B without an equal amount being paid on Class A. Further, Class B was convertible into Class A on a 1-to-1 basis at any time. And, Class B common stock controlled approximately 75% of the shareholder votes of the company. Classes A and B shared equally in the net assets of the company on liquidation. The company had historically paid dividends equally to both classes. In applying the two-class method, the company had allocated the undistributed earnings equally on a per share basis to both classes.

In this case, there was no contractually predetermined ratio of dividends between the classes of common stock. EITF 03-6, if applied literally in this situation, would have resulted in no allocation of undistributed earnings to the Class B common stock. Since the articles of incorporation provided that all the earnings could be entirely distributed to Class A without any distribution to Class B, no allocation would be made to Class B. Issue 03-6 states that "…if an entity could avoid distribution of earnings to a participating security, even if all of the earnings for the year were distributed, then no allocation of that period's earnings to the participating security would be made."

However, company management believed it was appropriate to allocate undistributed earnings between the two classes on a 1-to-1 basis. Management believed that the allocation of undistributed earnings must be done considering all of the rights and privileges of the different shareholder groups. They thought that it was unreasonable to think that the Class B common stockholders that control the Board would systematically allow the Board to pay dividends on Class A shares without paying them on Class B as well. Also, since the Class B was convertible at any time into Class A, the Class B stockholders could convert and capture any differential dividend, after it was declared. But, although Class B controlled the Board, the Board could not pay dividends on Class B without paying an equal amount on Class A.

After additional consideration and consultation, we came to believe that the particular facts in this case justified a different answer from a strict application of EITF Issue 03-6. The SEC staff believes that the EITF consensus was generally written in the context of classes of participating securities that do not control the company and that applying the EITF guidance by analogy in this case did not produce the most transparent reporting. Thus, the staff did not object to the company's judgment on how to apply the two-class method in this case. We did suggest that the company provide disclosures of the factors it considered in determining that its methodology was appropriate.

The "moral" of this tale of two-classes is that when applying the two-class-method to several classes of common stock, one ought to consider all of the rights and privileges of the classes in determining the allocation of undistributed earnings to the individual classes of common stock. And, for good measure you may want to ask the staff, about the issue as well.

Now before wrapping things up, I did want to point out that further guidance on two-class EPS is in sight. The FASB staff is currently drafting a FASB Staff Position to illustrate the computation of diluted EPS using the two-class method. Thank-you for your attention today. If you have any questions please send them up.


Modified: 12/11/2006