Resident fellow Alex J. Pollock submits a letter of comment to the Securities and Exchange Commission and explains why the conceptual deficiencies of the fair value accounting theory make it a poor choice for most companies as an accounting standard.

Thank you for the opportunity to comment on the SEC's consideration of "fair value" accounting standards. I respectfully submit the following comments and recommendation.

Summary

In my opinion, the conceptual deficiencies of the fair value accounting theory make it a poor choice for most companies as an accounting standard.

Discussion

Fair value accounting is an appropriate accounting standard for securities brokers. But should everyone else in the world be accounted for as if they were securities brokers? This is highly dubious.

The fair value theory inappropriately treats operating businesses, whose business is the generation of net cash flows over time, as if they were securities brokers, whose business is daily buying and selling of securities. These two business models suggest the "Banking View" and the "Securities View," respectively. The Banking View is set out in detail by the recent white paper of the International Banking Federation, "Accounting for Financial Instruments Conceptual Paper," which defines a clear, practical alternative to fair value accounting.

Accounting theories are like those of politics and philosophy: they are debated over years without clear demonstration of which is correct--they are neither like logical proofs nor scientific results. This makes them subject to developing ardent adherents and opponents, and to varying fashions over time. For example, the SEC absolutely demanded historical cost accounting for decades (unless you were in the securities business), but now the fashionable theory is fair value.

Proponents of the fair value theory usually claim that they are only insisting on "the facts" of market prices. Of course they admit that in many cases there is no active market or no market at all, thus no real price. They also admit that markets can be panicked and reach fire sale prices that will be judged by later observers as irrational--or in the opposite direction, become irrationally euphoric. For example, everybody knows that in the Great Housing Bubble, now collapsing, the market was simply wrong. But fair value proponents focus on "the fact" of the market price or estimated price right now. A typical expression of this view: "Let's not hide from the truth."

But what is accounting truth? It never is and never can be simply the facts. It is the facts treated according to some theory to calculate what the theory defines as its results, for example, the defined concepts of "profit" and "capital." There is no such thing as accounting truth in any simple way--there are only facts, and estimates, projections and guesses, turned into accounting results as defined by some theory. So one can easily grant that today's panicked bid is a fact, but not agree with how this fact translates into financial statements.

As is often pointed out, the fair value theory has particularly perverse results when applied in the midst of a market panic, when markets are neither liquid, active or orderly. What is the meaning of a "market price" when there is no market? Of course you can make estimates by projecting cash flows and applying a discount rate. But which discount rate? The fair value theory forces the huge uncertainty premium or panic discount of distressed markets into the profit and capital calculations of entities whose contracted-for cash flows may all be realized.

Moreover, fair value accounting becomes dependent on the prices of derivatives when the cash market is not trading. What is the meaning of a derivative price when there is no market price for the underlying cash asset? Use of derivatives in this situation exposes the market to manipulation, when speculators may drive down the price of the derivative to force reported losses by those holding the assets.

Fair value proponents like to argue that "historical cost is irrelevant." But for fixed income assets, this is not the question. Fixed income assets have a principal to be repaid and a stream of interest payments. If all principal and interest is going to be paid as agreed, what is the right accounting representation?

Consider a five-year bond financed with a five-year certificate of deposit to yield a spread of 2% and a net cash flow of 2% every year. Should the accounting reflect that steady cash flow, as long as the credit is good? "Of course," says the Banking View. "Of course not," say the fair value theorists, "we have to move the reported profits and capital around every quarter to imbed the different discount rates the market, or the lack of a market, is implying.

As is well known, fair value accounting means that deterioration in the quality of a company's own debt, which lowers its value in the market, is reported to the public as an increase in the company's profit and capital. Correspondingly, improvement in the company's credit quality, resulting in making its debt more valuable, creates an accounting loss and a reduction in capital. One article recently called this "a mere accounting illusion." It is worse than that: it is an absurdity which follows necessarily from the fair value theory--a reductio ad absurdum.

From a public policy point of view, fair value accounting is distinctly pro-cyclical, building market excesses of both optimism and pessimism into reported profits and capital. In the midst of severe bust, where we are now, this reinforces the downward cycle of panic-falling prices-losses-illiquidity-credit contraction-more panic-further falling prices-greater reported losses-no active markets. Fair value accounting adds momentum to a destructive downside overshoot. A recent FASB slide presented at a conference concludes, "Fair value reflects losses that have been incurred, it does not cause losses." In a downward cycle, this appears untrue.

Recommendation

I believe many accounting problems arise from the belief that there can be only one set of financial statements which must be "right" or "true." A better conclusion is that accounting truth cannot be completely captured by any single official approach. The various theories of what an asset is, what profit is, what capital is, give rise to different financial statements. The best solution would be to publish statements from different perspectives, using different theories applied to the same facts, estimates, guesses and transactions. For example, one could be current GAAP, one fair value, one the tax books, one the regulatory books, one what management thinks is the real story. Like the others, fair value would be one perspective, but not the one-and-only perspective. With this approach, we could all know more, because we are more likely to approach accounting truth through multiple perspectives.

Alex J. Pollock is a resident fellow at AEI.