Why fair value accounting can't work.

Like it or not, we're inching toward a fair value accounting model. While the FASB and some academics have been beating the drums for fair value or current value accounting, financial executives and managers have been equally articulate on the other side. Their arguments seem to boil down to three points:

* "If it ain't busted, don't fix it." What we have today (historical cost) works, and people understand it. In short, leave well enough alone.

* If unrealized changes in values have to flow through the income statement each period, they'll distort the results of operations.

* Determining fair value is far from exact. It involves costs to develop the information. It requires imprecise estimates and is susceptible to manipulation.

Yet one characteristic of this controversy stands out. Never in my 30 years in the appraisal business has a client said, "I'm curious; I'd just like to know what my assets are worth." Always, in every assignment, there's been an actual or potential transaction that required information regarding FMV.

Unless a transaction is contemplated, or has been consummated, information on values is of no practical use. People do not pay, and should not pay, good money for useless information.

How Much Do You Want for That?

Fair value is never going to work in accounting for a good reason. The fact is, at any time, there's never a single value that's fair. Take an asset familiar to everyone, a private residence. What's the fair value, say, for a personal financial statement?

* Is the fair value the cost of reproduction, the amount you'd want to collect from your insurance carrier if the house were destroyed by fire?

* Is the fair value the amount you'd receive, after commissions and fix-up costs, if you were to sell the house on the open market?

* Is the fair value the amount you want the local property tax asses r to place on it for purposes of collecting ad valorem taxes?

* Is the fair value what a virtually identical house next door sold for last year - when interest rates were higher?

It's clear that these four values would be different - in some cases, markedly so. Each purpose for which value information is needed calls for a different answer. And utilizing the incorrect answer may lead to erroneous decisions or actions. For example, insuring the house at the amount you wish it were assessed at might cause you to under-insure it. Or, putting the assessed value on your financial statement might understate your net worth, perhaps unfairly reducing your borrowing power.

One might argue that this example isn't typical of real world business conditions, but here's a true case study.

Company A bought Company B and had to allocate (APB 16) the purchase price over the assets acquired. The transaction was highly leveraged and the lenders were concerned about the value of the assets. If the company didn't pay its debt, the banks' liens would have to satisfy the amount owed.

Two different valuation reports were prepared simultaneously, each with a carefully described narrative section. One report gave the liquidation value of the assets in case the banks had to sell them to a third party. Of course, these values were relatively low. The other report was for financial reporting, where the premise of value was continued use for the purpose for which the assets had originally been acquired - an amount that was much higher.

Not surprisingly, the two amounts were substantially different. The bank didn't want the allocation values, and the SEC wouldn't have accepted the liquidation values. If, in addition, the appraiser had been asked to value the assets for insurable value and for personal property tax assessment, he would have had to give the client two additional reports, a total of four values for the same assets at the same time.

While appraisers are often accused of giving clients whatever answer they want, the truth is that we always determine first the purpose for which the value information is going to be used. Then, and only then, can we begin our research and analysis, before reaching our conclusions of value and writing the report. "Change the purpose for which the report is required and you'll change the value" isn't the message accounting theoreticians or regulators want to hear.

What's Fair?

Those who believe fair value would be preferable to today's historical cost model almost always assume there's but one value, and that shareholders and creditors (the recipients of accounting information) would be better off with it.

But what is that value? How should it be defined? If it's a non-operating asset, perhaps marketable securities or excess land, the premise of value should be what it could be sold for in a reasonable period.

Those are the easy assets theoreticians and regulators like to sort out.

In dealing with operating assets, the nuts and bolts of a manufacturing business, what is fair value? What do the proponents of fair value accounting want?

* Is it the cost to reproduce new the exact assets in place - with or without accumulated depreciation from all causes?

* Is it the cost to replace the assets in place with the latest new technology, often at lower cost?

* Is it the cost to acquire the same or similar assets from a used-equipment dealer?

* Is it the price that would be received from a dealer if the assets were sold individually?

* Is it the price that would be realized at an industrial auction?

* Is it the price the company would receive if the entire business unit were sold, perhaps including intangibles?

* Is it the original cost, adjusted by some price index?

Simply asking these questions highlights the difficulty of picking one approach as the answer in all cases.

The first method is usually used in an asset allocation in a business combination. The second was at one time mandated by the SEC. In valuing a going concern, we sometimes use the third approach. Banks and finance companies want the fourth and fifth. Management wants to know what the business would sell for in total, as well as if the assets were disposed of individually. Finally, the FASB required the indexation approach in SFAS 33, fortunately now withdrawn.

When appraisers are accused of developing unreliable information, or two different appraisers differ in a tax case or lawsuit, it's often asserted that developing values is uncertain. That's not true. The principles of valuation are well known, and professional appraisers follow them.

Almost invariably, when there's a question about the reliability of values, the answer lies in the premise of value. If one appraiser values on the basis of continued use, and a second assumes liquidation, of course the answers will differ. As long as the appraisers agree on the basic assumptions, two independent appraisers should come within 10 percent of each other. Frankly, two separate accountants preparing a set of financial statements for a company probably wouldn't come as close, in terms of net income.

Who's the Fairest of Them All?

The FASB and many accounting professors appear to want fair value information either in the face of the financial statements or in the footnotes thereto. Financial executives, to the contrary, seem to such proponents to be dragging their feet when it comes to fair value reporting. Is this because financial managers explicitly or implicitly know there's no such thing as the value of anything?

Those of us in the appraisal business are often accused of promoting fair value accounting. After all, it's asserted, "Think how great it would be for your business." But no one knows better than we that rulemakers cannot promulgate an accounting standard that would develop the right answer at the right time.

As appraisers, we can develop any type of value information desired, based on any definition of value our clients want (or are required to provide). What we cannot do is create value information based on any single premise of value, and be sure it won't be more misleading than useful.

Different circumstances demand different value information. Who's going to write a one-size-fits-all rule? If critics complain that today's historical cost-based financials don't provide adequate information to users of financial statements, what will those same critics say if they're given reproduction cost information when they want liquidation values, or vice versa?

It would be nice to let financial managers decide what information, whatever premise of value, they should use in developing fair value. But then there'd most likely be a total lack of comparability for critics to complain about. Specifying just one definition of value in advance would be equally useless, as both the SEC and FASB found out in their competing approaches to inflation accounting.

There appears to be only one reasonable approach, an approach that should both satisfy the needs of users and recognize the abilities of preparers to develop the data. That would be to prepare and disseminate certain value disclosures as supplemental information, perhaps in the MD&A so it needn't be audited.

Management would have to determine if any material changes in assets were contemplated or reasonably likely to occur. These would include potential disposition of assets or asset categories, as well as replacement of existing facilities. SFAS 121 for impairment calls for an analysis of possible impairment, whether or not a writedown is taken. In my proposal, management would estimate whether assets are liable to be converted into cash, i.e., sold or exchanged. If so, they would disclose the likely outcome in terms of future cash flows. Similarly, if major cash expenditures are required to replace existing facilities, whether fully depreciated or not, this, too, would be disclosed.

What's key is that fair value information would be neither developed nor disclosed for assets that will continue to be used for the purpose for which they were acquired. Assume there are going to be no cash consequences other than production and sale of goods and services in the normal course of business. Then developing and disclosing the theoretical value of assets (whatever the premise of value might be) that will be neither bought nor sold would be no more than an exercise in futility, so why do it?

To summarize, then, fair value information could and probably should be developed and disclosed when, and only when, future cash flows are likely. When cash flows relating to assets aren't likely, let's stick with the tried and true accounting model that's served us so well for almost 100 years.

Alfred M. King is chairman of Valuation Research Corp., based in Princeton, N.J. You can reach him at (609) 4520900. King also is a member of the Brookings Institution working group on intangible assets. He has worked with the APB and FASB as a member of both FEI and IMA technical committees. As an accountant, corporate controller and CFO, he's heard the arguments on both sides.