As I mentioned during a recent session of office hours that I hosted, I believe there are a number of persistent myths in accounting that seem to get passed on from generation to generation.  I promised to share them on the blog, so here they are. 

These are essentially a collection of statements that I have encountered over the years, which I believe to be either incorrect, incomplete, or misleading.  These statements have come not only from students, but also from textbooks, colleagues, research papers, and the financial press (after all, students generally get their information from somewhere).  Along with each statement is a short synopsis of my thoughts on the issue.

Disclaimer: While I am the incoming Research Fellow at the FASB (beginning July 1, 2009), these our my own views and in no way reflect the views of the Board or staff.

1)    “There is an inherent trade off between relevance and reliability.”
I certainly agree that delaying recognition of information about a company’s economic resources, obligations, or performance until the measurement error surrounding those items is reduced (e.g., through verifiability) clearly reduces the timeliness of information (a problem for relevance).  But, at the same time, by failing to capture information about the underlying economic situation of the company, delaying recognition also results in financial statements that are less representationally faithful, which is a problem for reliability.

In other words, to be relevant, information should help users to assess the timing, amounts, and uncertainty of a future cash flows; to the extent that information is not representationally faithful, it will also lack predictive and feedback value.  Consequently, relevance and reliability should also (and perhaps primarily) be viewed as complements.

2)    “Allowing multiple accounting treatments harms comparability.”
While this is often true, it need not be so.  Financial statements are comparable to the extent that they allow users to identify real economic similarities and differences between firms and over time.  In the same way that changing the form, but not the substance, of a transaction should not allow an entity to receive a different accounting treatment from its peers, uniformly applying similar accounting procedures to substantially different business practices can obscure the true differences between companies.

3)    “Historical cost continues to be the primary basis for measuring most assets and liabilities in the US.”
No asset or liability is accounted for strictly according to historical cost under US GAAP.  Inventories are subject to lower of cost or market measurements.  Land is subject to impairment.  Even cash is translated into dollars according to current exchange rates.  Thus, while a unit of US currency can truly be said to be measured at historical cost, I’m not sure that much else can.

The notion that we have statements that reflect (mostly) historical costs is dangerous because it sweeps under the carpet the large amount of estimation and adjustment that permeates financial statements prepared in accordance with US GAAP.  In doing so, it also creates a false dichotomy in debates about whether US GAAP should abandon historical cost and shift to more fair value measurements.

Lastly, let me also point out that historical cost of an asset or liability is often the best estimate of its fair value at the time of initial acquisition and at the time of eventual derecognition (if the asset or liability is ultimately sold in a transaction).  As such, it is largely in the interim periods that these two measurement attributes can diverge from one another.  But it is important to keep in mind that this divergence is not driven solely by changes in fair value while historical cost stays fixed and constant and objective – rather, US GAAP almost always requires subsequent adjustments (systematic and/or periodic) when implementing historical cost as a measurement principle.

4)    “Historical cost is more reliable than fair value.”
One part of the problem here is that the statement masks people’s true feelings on the matter.  In my mind, what people often mean by this statement is that, deep down, they are more sure about something in the past (what was paid) than they are about something in the future (what would be received for transferring the asset or liability).  The problem with that comparison is that, as I just pointed out, few items are truly carried at historical cost.  Once a cost gets adjusted for reporting in a subsequent period, it is the reliability of the adjusted number that matters.

The other part of the problem with this statement is that it is vague.  Reliability has three components, representational faithfulness, verifiability, and neutrality.  What some argue is that adjusted historical costs, while not necessarily more representationally faithful or neutral than fair value, are in many cases more verifiable than fair value.  If this is what people mean, then this is what they should say.  Otherwise, confusion and misunderstanding are likely to prevent meaningful discussion from taking place.

5)    “Historical cost has a big advantage over fair value measurements:  it is verifiable.”
Verifiability is defined as “the ability through consensus among measurers to ensure that information represents what it purports to represent or that the chosen method of measurement has been used without error or bias” (SFAC 2).  There are two aspects to this definition.  The latter part relates to the ability to replicate a measurement.  To the extent that a financial statement preparer has documented the process used to generate a fair value measurement (e.g., observing a price in a specific market at a particular point in time, or applying a specific model with a specific set of inputs to generate an estimate of fair value), these measurements can be verified much in the same way that an allocation process can be verified.  In both cases, an auditor can reduce measurer error by verifying that the measurement technique has been correctly applied.

The other part of the definition relates to our ability to ensure that the measurement technique results in a measurement that is representationally faithful.  Verifiability of the measurement technique cannot reduce this type of measurement error.  Because of this, both historical cost, through outdated information or erroneous assumptions, and fair value, through ill-suited measurement models or invalid inputs, can fail to faithfully represent the underlying economic phenomena in the financial statements.

Thus, whether this statement is a myth or not depends on what one means when they say it.  However, I believe it is generally true that fair values are much more verifiable than is often assumed (they too can be verified to reduce measurer error) and that historical costs are often much less verifiable than is often implied (like fair values, they are not immune to measurement error or bias).

6)    “Fair value measurements are more subject to manipulation than are historical costs.”
While there are seeds of truth in this statement, it creates the impression that “mark-to-myth” fair values measurements are constrained only by the limits of management creativity, rather than being disciplined by rules such as SFAS 157.  According to SFAS 157, fair value measurements should be consistently applied in the same manner.

Specifically, the pronouncement states the following:
“Valuation techniques used to measure fair value shall be consistently applied.  However, a change in a valuation technique or its application (for example, a change in its weighting when multiple valuation techniques are used) is appropriate if the change results in a measurement that is equally or more representative of fair value in the circumstances. That might be the case if, for example, new markets develop, new information becomes available, information previously used is no longer available, or valuation techniques improve. Revisions resulting from a change in the valuation technique or its application shall be accounted for as a change in accounting estimate.”

Thus, like amortization schemes, fair value measurements can be altered in any given period, but doing so requires compliance with SFAS 154, which includes making the case that the new measurement technique results in more useful information.

7)    “Fair values are irrelevant when a company does not plan to dispose of the asset or liability in the marketplace.”
This only true if the assets or liability is certain to produce or consume the underlying cash flows at the times and in the amounts that were expected at the time of the initial transaction.  If there is uncertainty, then providing information about how the market currently values those assets or liabilities can provide information on, not only the amounts that are now likely to be received or paid in expectation, but can also provide timely information on the underlying risk of those cash flows, which is a crucial component of not only making investment and credit decisions, but also for evaluating whether the returns that management is generating are sufficient for the risk being taking on.

8)    “Fair values are exit prices and so measure an asset’s liquidation value.”
This is statement is wrong conceptually if it is implying liquidation in a fire sale of individual assets.  As defined in SFAS 157, fair value is an estimate of the price that would be obtained in an orderly transaction. Moreover, fair value measurements are to assume the highest and best use of an asset or asset group, which might be based on an in-exchange valuation premise, but may also be based on in-use valuation premise (whichever is believed to provide the highest and best use to market participants).

What the above statement highlights is the danger in blindly relying on exchange prices to estimate fair value.  An in-exchange premise may or may not be appropriate and, if markets are thin, or if market conditions are otherwise suspect, then the prices observed in those markets will not necessarily be the best estimate of an asset or liability’s fair value.

In response to confusion on this matter, the FASB has recently issued guidance on how to estimate fair value when markets are thin (FSP FAS 157-3) and how to determine when markets are inactive (FSP FAS 157-4).