Reaching a Fair Settlement on Fair Value

In the year I’ve been writing for Compliance Week, one subject I haven’t broached is fair-value accounting. I’m not alone; the Securities and Exchange Commission’s Advisory Committee to Improve Financial Reporting has decided not to touch it either, while offering recommendations on all manner of other nettlesome problems. Nonetheless, the debate continues over fair value versus historical cost accounting. Since I’ve seemingly heard every argument on all sides of this debate, it’s probably time for me to speak up, too.

In early drafts of this column, I tried to summarize the positions for and against fair value that have been espoused by various voices in the financial reporting community. This proved an exercise in futility, because the strong views that have been expressed on both sides don’t lend themselves to quick summarization. Instead, I will discuss my views and make some suggestions about how to move forward on the measurement debate in an orderly fashion—rather than in the haphazard fashion that has typically characterized progress on measurement guidance prior to the arrival of Financial Accounting Standard No. 157, Fair Value Measurement.

The Case for Change

Accounting rulemakers have a hard time selecting measurement attributes. The Financial Accounting Standards Board and the International Accounting Standards Board have typically chosen measurement attributes on a standard-by-standard basis, trying to maximize relevance and reliability for the particular asset or liability in question. Of course, relevance and reliability (or faithful representation) are the primary qualitative characteristics of accounting information, as described in FASB’s conceptual framework. Financial information is relevant if it provides information that is useful for decision making. Information is reliable to the extent that it succeeds in portraying whatever it is supposed to portray. So any particular measurement attribute can exhibit varying levels of relevance and reliability when applied to different assets and liabilities.

The accounting literature employs a wide variety of measurement characteristics, from those based solely on historical cost to those based entirely on future value, and everything in between. As a result, it can be difficult to understand exactly what the balance sheet figures are intended to represent. Although discussion usually uses the terms “historical cost” and “fair value” as if they are the only potential attributes, numerous cost-based and value-based measurement attributes exist.

The trend in recent years clearly is toward more value-based measurements, but each move in that direction is met with resistance and raises new challenges for accountants, auditors, and regulators. The discussion is characterized by strong view and, at times, distrust. Many preparers believe standard setters have a secret goal of requiring the reporting of all assets and liabilities at fair value; some investors and others believe preparers want accounting standards that will facilitate manageable results, rather than those that clearly communicate the company’s financial position.

Despite the often-rancorous debate, most everybody agrees that the current situation is untenable. Today’s accounting model can result in accounting fluctuations that don’t reflect economics, includes complex rules to determine which items are measured in which way, allows multiple measurement attributes for the same asset or liability, and provides many opportunities for preparers to make mistakes. All of this might be worth it if financial statement users were getting the information they wanted, but I don’t believe that’s the case either. Change, of course, is difficult to manage and requires a time commitment from everybody affected to learn, understand, apply, and interpret the new guidance. But the status quo in this area isn’t good for anybody.

Current Value for All Financial Instruments

One reason fair-value accounting has become more popular is that fair value is always a relevant piece of information, while cost often is not. The current value of an asset or liability invariably is useful, but the cost of some items, such as derivatives, becomes essentially meaningless as time passes. Hence a measurement attribute based on current value is the only kind of attribute that could logically be applied to all assets and liabilities. Some investor groups, including the CFA Institute, believe that we should move to a full fair-value balance sheet, entirely ending the problems and confusion caused by the mixed-attribute model.

I’m not ready to go that far, I do believe that financial instruments should be reported at current value. Yes, some people believe current values of assets that management doesn’t plan to sell are not relevant, and that those assets are best reflected at accreted cost; this is the thinking that led to the Held-to-Maturity category under FAS 115, Accounting for Investments in Debt and Equity Securities. But a financial instrument is valuable to the holder only because it provides the right to cash, either through a current sale in the market or the payments required by the terms of the instrument. That’s why I believe that an estimate of the current value of the future cash flows associated with the financial instrument is the only meaningful way to report financial instruments.

Granted, the current value of some financial instruments can be difficult to determine. But if cost-based measures aren’t relevant, it doesn’t matter that they are more reliable than value-based measures. Accurately reporting a piece of information that is not useful doesn’t improve financial reporting. What should happen instead is an effort to improve how we determine the current value of complex financial instruments, so that we can constantly improve the reliability of those measures.

Reporting financial instruments at current value would also eliminate the need for fair-value hedge accounting, a source of complexity and many restatements. And it would eliminate situations where the same asset can be reported using multiple different measures based on choice and intent, which confuses financial statement users. Finally, it would eliminate the incentives for certain accounting-motivated transaction structures, as these structures often seek to take advantage of the ability to change measurement attributes with little real economic change.

Cost-Based Measures for Most Other Assets and Liabilities

Once we get into assets that can be consumed, rather just exchanged or settled for cash, I’m not sure current value is the best measurement attribute. For example, machinery used to create products and provide services has value not only because it can be sold in the market, but also because it can be used to produce products or provide services that then are sold.

In this way, these kinds of assets can generate value by selling them directly, or by consuming them while creating other items of value. When value can be obtained from an asset through its use, allocating the cost of that asset to the products or services created is logical. The goal isn’t to report the value of the asset at any point in time, but simply to allocate the cost of the asset to the periods and activities in which the asset it used. Likewise, it also seems appropriate to report liabilities that will be settled through the provision of goods or services at the amount the company received to take on the obligation. As such, I do think that for non-financial items, cost-based measures are often relevant.

A financial instrument is

valuable to the holder only

because it provides the right

to cash ... I believe that an

estimate of the current value

of the future cash flows

associated with the financial

instrument is the only

meaningful way to report

financial instruments.

In addition, reliability considerations are different for financial and non-financial items. Historical cost information is generally reliable for most items, as the amount paid for an asset or received for taking on a liability is usually evident and verifiable. In contrast, the reliability of value-based measures can vary significantly. Yes, current values of assets that are traded in active markets generally can be portrayed accurately by reference to the prices in those markets. But current values of assets that are unique or rarely exchanged must generally be determined based on projections or models and can be somewhat hypothetical in nature. Still, it is far more common for financial instruments to be traded in active markets than it is for non-financial items. So the reliability of value-based measures of non-financial items is, on average, lower than the reliability of such measures of financial instruments.

Given the combination of the greater relevance of cost-based measures of non-financial items and the decreased reliability of value-based measures of such items, the argument to report these items at a cost-based measure is much stronger. Furthermore, the effort needed to determine the current value of these non-financial assets is going to be a lot more than the effort needed for financial assets.

Surveys of financial statement users about the usefulness of current value measures have not been very illuminating to me, suggesting some confusion over the issue. Without user requests for change on this front, I’m inclined to keep things more or less as they are. As such, I believe that standard setters should generally retain cost-based measures for non-financial assets.

Of course, cost-based measures are generally accompanied by an impairment test, meant to ensure that an asset is not reported at a value that is too high—although the definition of “too high” varies. These tests seem most sensible when they are based on whether costs of an asset can be recovered or whether costs to fulfill a liability will be higher than the amount received for taking on the obligation. Why? Because these kinds of analyses are consistent with the logic that makes cost-based measures relevant: the item is being consumed (or provided) over time, so the cost and benefits should be allocated over the related activities.

Volatility Should Be Identified

So far, this column has considered the effects of measurement decisions only on the balance sheet. These decisions also affect income. Concern about income is what has caused most of the shouting that happens whenever standard setters suggest more remeasurement. Some of the concern arises over recording changes in the value of some items without recording offsetting changes in the value of other items. This has led to hedge accounting and to the fair-value option, which bring their own set of problems.

Other concerns regarding volatility stem from a desire to align external reporting with internal performance measures. The argument generally goes that investors should not hold management responsible for unrealized changes in asset values because management is not held responsible internally for such items. Others more directly express the desire to have more “manageable” results of operations to report than would otherwise be possible if one had to report significant changes in value. All of these concerns are exacerbated when there is a significant amount of uncertainty surrounding the measure of current value, because the reported results may be less precise.

Reporting all financial instruments at fair value, as I have suggested today, would alleviate many of the situations where “false volatility” is reported. Other situations where this occurs should be examined and addressed if possible. But it seems clear to me that reporting volatility in current values that actually exist should be encouraged rather than avoided. If senior management or the board of directors wants to evaluate its personnel using measures that ignore volatility, they are free to do so—but that doesn’t mean investors should be denied the information about that volatility.

That being said, unrealized changes in value are seen by some as a different kind of income or expense than income that arises from transactions. While I don’t know if this view is always logical, I see no harm in reporting some or all unrealized changes separately from realized ones. Companies should be allowed, perhaps even encouraged, to disaggregate some or all unrealized changes in values from other components of income. So long as the classification or disaggregation methodology is explained, I believe investors can determine what to do with the information.

What About FAS 157?

With the implementation of FAS 157, virtually all current value-based measurements in financial statements prepared under U.S. Generally Accepted Accounting Principles are based on this one particular definition of fair value. Many aspects of FAS 157 have proven either difficult or controversial, or both, in its first period of use. For example, FAS 157 states that fair values shall always be based on exit prices rather than entry prices, and shall be based on assumed (or hypothetical) marketplace information rather than entity-specific information. Both of these provisions have generated multiple implementation issues.

Throughout this column, I have used the terms “value” or “current value” rather than “fair value” because I don’t think it is appropriate to limit the thinking on measurement to whether or not FAS 157’s version of “fair value” is correct or not. There are plenty of other current value-based measures that could be used, and time might show that these measures are better than fair value for certain items. The financial reporting community needs to be open to that possibility. The choice doesn’t have to be narrowed to historical cost and FAS 157’s view of fair value unless we arbitrarily limit it that way.

As for the question of whether FAS 157 is the right definition of fair value, that will need to wait for a future column, along with the question of whether FAS 157 is somehow partially to blame for today’s problems in the financial markets.