A few weeks ago FASB Board member Leslie Seidman joined us for office hours.  We covered a lot of ground, and Ray has already written on some of her ideas for research.  I thought I would summarize her rather cautionary views on fair value accounting for financial instruments.  Playing to the academic audience, Ms. Seidman presented a three-dimensional matrix that captures

  1. the marketability of the instrument,
  2. the variability of its terms, and
  3. the activity with which the firm is actually trading them.

In one corner of this cube, you have highly marketable securities that whose terms (and hence value) are constantly fluctuating, and that are being actively traded by the firm.  Examples might include purchased debt securities or derivatives.  Ready marketability makes it easy to value the securities, because market prices will exist for similar or identical assets, be recent, and be relatively untainted by illiquidity; the fact that terms change regularly (e.g., every day that passes alters the option value of many derivatives, and may alter an issuers creditworthiness) means that regular remeasurement conveys relevant information; and the fact that the firm actually trades the assets implies that they can realize the returns captured by fair value remeasurement.  Thus, high scores on all three dimensions indicate an instrument that Leslie would be pretty open to reporting under fair value standards.

At the other extreme, a firm’s private annuity guarantee provides a good example.  The private nature of the contract makes it unmarketable, the fact that it is an annuity means that value is determined only by the discount rate (assuming high creditworthiness of the guarantor), and the guarantor is probably unable to transfer the liability to another party, so even if there were a change in value, the firm could not directly realize it through trade.  (In an email exchange as I was writing this post, Leslie indicated “I feel less comfortable applying a fair value measurement to [private and untransferable instruments], because there is no ‘exit market.’  That does not mean that this type of arrangement would be carried at cost; rather they might be measured at a current amount that does not purport to represent a hypothetical sale to a market participant.”)

I have a few observations on this model.

  • The model does seem pretty consistent with a conceptual framework.  Marketability captures reliability and verifiability, while variability in terms and trading activity seem to capture the purported relevance of unrealized gains and losses (are there any, and if so, are they realizable). So I think it is a framework that is workable for those who want to persuade standard setters that fair value is not always the solution–because persuasion requires discussing the issue using vocabulary and arguments that standard setters use.
  • The model is essentially silent on the macroeconomic effects of fair value accounting.  Not a day goes by that I don’t see an argument that fair value accounting should be eliminated because it is pro-cyclical and can amplify bubbles or crashes or both.  I wonder if Leslie would defend the omission by saying that such questions are not the appropriate purview of the FASB (and are not part of the conceptual framework).  Alternatively, maybe she would add a 4th dimension that would capture this type of economic impact.
  • The model is not symmetric, in that a purchaser of debt may measure an asset at fair value, while the issuer does not.  This asymmetry arises because the contract terms are asymmetric–in most cases, anyone can trade an IOU but the person who issues it.  Would this allow for some form of accounting arbitrage?

Now, a couple of research implications.

  • The matrix does suggest a fairly natural research program–step 1 being to assess whether the three dimensions capture variations in the value-relevance of fair values for financial instruments.  Is it possible to identify two types of instruments that are similar on two dimensions, but vary only in a third?
  • I think many academics would disagree about the importance of being able to realize unrealized gains through trade.  For example, Barth and Hodder have an article indicating that earnings numbers are pretty useful when they reflect gains on debt remeasurement arising from a decline in a firms own credit quality.