Essentially, the FASB proposes purchase accounting with an impairment test for goodwill. In other words, goodwill is amortized "only if it was found to be impaired or was associated with assets to be sold or otherwise disposed of." Appendix A explains that the Board believes that this approach is more decision-useful than other methods. More telling is the comment that many constituents (read "top corporate managers who have gained access to important Washington people through large soft money contributions") approve the scheme.

For the sake of argument, I shall assume that it is indeed more useful than other techniques. But only for the sake of argument. I shall also concede, only for the purposes of this argument, that goodwill meets the definition of asset and that this approach meets the definition of expense. Concepts Statement No. 5, paragraphs 75-77, still requires the item to be reliable. "To be reliable, information about an item must be representationally faithful, verifiable, and neutral." I submit that the FASB’s proposal meets none of these criteria.

As measured by this exposure draft, goodwill is not representationally faithful, mostly because it doesn’t even have a referent. It’s hard to be representationally faithful when the surrogate exists for its own sake. True, financial economists talk about objective goodwill and subjective goodwill, such as in the classic The Theory and Measurement of Business Income by Edwards and Bell, but there is a huge disconnect between accounting and economic goodwill. Economists and accountants may employ the same word "goodwill" but they are referring to quite different phenomena. When economists apply the word, they mean either market value or value in use. When accountants utilize the term, they have in mind whatever plugs the journal entry for the business combination.

Goodwill after the date of acquisition also fails the test of representational faithfulness. To measure goodwill you will obtain the market value of the firm ("reporting unit") and subtract out the sum of the market values of the individual net assets of the organization (paragraphs 19-24 and 85-96). What is absurd about this algorithm is that it assumes that the residual amounts relate solely to the goodwill purchased as a result of a previous business combination. This computed goodwill, however, can easily be tainted with "internal goodwill" generated from other activities, events, resources, and social arrangements. The calculated number may thus have little to do with the purchased goodwill.

The proposed impairment test is not verifiable, for managers will have the opportunity to deny impairment in many situations. Even when the economy turns sour so that managers look very foolish claiming no impairment, the executives will simply find those appraisers and consultants who will provide the desired numbers. Auditors will have no benchmark by which to evaluate the treatments, so they will sign off on almost anything the managers want. When anything goes, nothing is verifiable.

The impairment test likewise is not neutral. Instead, it empowers managers to lie through their teeth. Unfortunately, there’s nothing neutral about data manipulation.

If the exposure draft is accepted as is in the final draft, I predict that most firms will show no goodwill amortization except during big baths and restructurings. I further predict an increase in shareholder lawsuits and SEC investigations because of management abuse. FASB is opening a can of worms with this accounting, and lawyers will stay busy for years to come.

Finally, consistency demands reopening a lot of other accounting techniques. I suggest that we capitalize R&D and use an impairment test. I propose that we drop depreciation and depletion and limit ourselves to impairment tests on plant and equipment and natural resources. Let’s also capitalize all leases and amortize these values only with an impairment test. Oh, I forgot -- the constituents might get unhappy with that suggestion.

More by J. Edward Ketz