Fair Value and the financial crisis ( Accounting )

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009) pages 189 – 213 © CAAA / ACPC doi:10.1506/ap.8.3.1

Fair Value Accounting and the Financial Crisis: Messenger or Contributor?*

MICHEL L. MAGNAN, Concordia University

AbstrACt This commentary discusses how fair value accounting (FVA) affects the nature of financial reporting, especially for financial institutions that were deeply affected by the 2007 – 9 finan- cial crisis. Toward that end, I address four questions. First, I review FVA’s role in financial reporting, emphasizing its development over time. While the commentary’s focus is on the interface between financial instruments and FVA, its reach extends well beyond financial instruments. Thereafter, I discuss the measurement and valuation challenges that arise from the use of FVA in financial reporting. Then, I analyze the evidence, analytical and empirical, on the role that FVA may have played in the financial crisis of 2007 – 9. Since, to some extent, the crisis is still unfolding, there is limited yet very insightful empirical evidence on this issue. The evidence does suggest that FVA, in combination with its use by regulators, may have severely undermined the financial condition of some institutions. The effect was am- plified for institutions holding assets in markets that saw their liquidity dry up during the crisis. In other words, FVA may have amplified the crisis. Finally, I discuss some implica- tions that we can draw from the crisis about the merits and risks underlying FVA. For instance, I conclude that, in a search for relevance, the use of FVA in financial reporting may accelerate its disconnection from a firm’s business reality.

Keywords Conservatism; Fair value accounting; Financial crisis; Liquidity; Relevance; Reliability

LA CoMptAbiLité à LA juste VALeur : MessAgère ou protAgoniste de LA Crise FinAnCière ?

résuMé L’auteur s’intéresse à l’incidence de la comptabilité à la juste valeur sur la nature de l’information financière, en particulier pour les institutions financières qui ont été durement touchées par la crise financière durant les années 2007 à 2009. Il se penche plus particulièrement sur quatre questions. Il examine d’abord le rôle de la comptabilité à la juste valeur dans l’information financière, en retraçant son évolution au fil du temps. Bien que ses observations portent principalement sur l’interface entre les instruments financiers et la comptabilité à la juste valeur, elles débordent largement ce sujet. L’auteur traite ensuite des difficultés que soulèvent la mesure et l’évaluation que nécessite le recours à la

* Financial support from the Center for the Interdisciplinary Research and Analysis on Organizations (CIRANO) and from the Lawrence Bloomberg Chair in Accountancy is acknowledged. The author benefited from discussions on this topic with Ross Watts, Trevor Harris, Dan Thornton, Antonio Parbonetti, and Andrea Menini. Moreover, useful comments and questions were received from participants at workshops organized by CIRANO, the Montreal CFA Association, and University of Padova.

190 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

comptabilité à la juste valeur en information financière, après quoi il étudie les données, analytiques et empiriques, sur le rôle que peut avoir joué la comptabilité à la juste valeur dans la crise financière des années 2007 à 2009. Puisque, dans une certaine mesure, cette crise se poursuit, les données empiriques relatives à cette question sont encore limitées, bien que très éclairantes. Les faits suggèrent que la comptabilité à la juste valeur, en conjugaison avec son utilisation par les autorités de réglementation, peut avoir sérieusement contribué à la détérioration de la situation financière de certaines institutions. L’effet a été plus marqué encore pour les institutions détenant des actifs sur des marchés qui ont vu se tarir leurs liquidités au cours de la crise. En d’autres termes, la comptabilité à la juste valeur pourrait avoir amplifié la crise. Enfin, l’auteur traite de certaines conclusions pouvant être tirées de la crise au sujet des avantages et des risques sous-jacents à la comptabilité à la juste valeur. Il en déduit, par exemple, que dans une recherche de pertinence, l’utilisation de la comptabilité à la juste valeur en information financière peut accélérer la rupture de cette information avec la réalité économique de l’entreprise.

Mots clés : comptabilité à la juste valeur, crise financière, fiabilité, liquidité, pertinence, prudence

Despite its almost universal adoption by accounting standard-setters, the merits of fair value accounting (FVA) continue to generate intense, passionate debates among academics, busi- nesspeople, regulators, and investors. A surprising element underlying these debates is the apparent irreconcilable positions adopted by some participants in favor of or against FVA. The recent financial crisis has exacerbated the debate further. For instance, at the request of the U.S. Congress, the Securities and Exchange Commission (SEC) conducted an inquiry into fair value accounting’s contribution to the fall 2008 financial crisis (SEC, 2008). In its report, the SEC endorses the use of fair value accounting and concludes that mark-to-market accounting did not cause or contribute to the crisis. However, the SEC recommends that the Financial Accounting Standards Board (FASB) provide more implementation guidance to financial statements’ auditors and preparers. In addition, earlier in 2008, key standard-setters, such as Canada’s Accounting Standards Board, the FASB, and the International Accounting Standards Board (IASB), introduced temporary provisions waiving some aspects of fair value accounting for financial institutions. In the spring of 2009, the FASB and other standard- setters issued additional guidance regarding how to account for securities in illiquid, distressed, or disrupted markets. Still, despite these changes, several issues surrounding the measurement and recognition of fair values in financial statements remain unsettled. The recent financial crisis actually provides a unique context to gain insights into the implemen- tation of FVA and to assess the broader consequences deriving from its adoption.

The purpose of the paper is fourfold. First, I briefly describe the reach and foundations of FVA in financial reporting, underlining that it goes well beyond financial instruments. However, I focus on its application to financial instruments and financial institutions. My review encompasses the conceptual and empirical underpinnings of FVA. Second, I analyze the measurement and valuation challenges that arise from using FVA in financial reporting. Third, I review the evidence, analytical and empirical, regarding the role that FVA may have played in the financial crisis of 2008 – 9. Because, to some extent, the crisis is still ongoing, empirical evidence on this issue is limited. Fourth, I raise other issues that standard-setters,

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 191

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

regulators, and investors should consider when using or promoting the use of FVA in financial reporting. For example, I conclude that, in a search for relevance, financial reporting may become disconnected from business reality. I discuss FVA’s role as either a messenger or contributor to the crisis.

An analysis of the theoretical and empirical foundations of FVA suggests four observa- tions about its potential role in the crisis. First, despite its unifying label, FVA actually encompasses different measurement modes. The validity of these alternatives is conditional on market efficiency. Second, demonstrations that FVA-based figures are value-relevant do not necessarily imply that they are optimal from a societal perspective. As such, FVA ac- celerates a standard-setting trend away from a broad societal viewpoint toward an exclusive focus on shareholders. Third, the implementation of FVA challenges many traditional ac- counting concepts such as verifiability, reliability, and conservatism. However, the rationale for these concepts (i.e., to restrain managerial opportunism and facilitate contracting) is still with us. Finally, the widespread adoption of FVA potentially increases volatility in reported earnings and accentuates the disconnect between financial statements and the real economy.

Overall, despite its conceptual appeal, I conclude that the implementation of FVA is fraught with many difficulties that severely undermine its potential advantages. Moreover, although its role in the financial crisis may have been exaggerated, there is emerging evi- dence that the use of mark-to-market may have contributed to a failure contagion effect among financial institutions. Hence, accounting may have been more than just a messenger. Paradoxically, there are also reasons to believe that the use of fair value accounting (mark- to-model) may have allowed some financial institutions to delay the recognition of underlying problems as the crisis was building up. In addition, the move toward fair value accounting has the potential to further detach financial reporting from the true and fair view of reflecting underlying business reality, making financial markets and financial reporting self-reflecting mirrors. Finally, the use of fair value accounting by bank regulators raises several questions about the mapping between the accounting standard-setting process (which focuses exclu- sively on investors’ needs) and regulators’ objectives (which include the stability of the banking system and the well-being of society as a whole).

FAir VALue ACCounting: An oVerView

Historical background

Fair value accounting is neither a novel concept nor a new practice. In the late 19th century and early 20th century, it was common for firms to value their capital assets using appraised values (i.e., estimates of the net realizable values that the assets would bring in the market). Many early economists also believed that the exit value, or the amount a firm would realize by selling an asset in the market, was the only appropriate basis to construct financial state- ments (see, among others, Diewert, 2005). However, by the 1930s, abusive valuation practices by some managers led to the enactment of more formal accounting standards by the accounting profession. As a result, historical cost emerged as the dominant practice for reporting most assets and liabilities. Nevertheless, fair value accounting remained an attrac- tive concept to many accounting theorists who formalized its appeal and form over the

192 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

years. For instance, Staubus (1961), Chambers (1964), and Sterling (1970) argue for the use of exit values in financial reporting (i.e., the net realizable value for the asset).

Although FVA was kept as a default option in accounting for some assets (i.e., lower of cost and market), it effectively reentered U.S. firms’ financial statements only in 1993 through SFAS No. 115 (FASB). Going beyond prior fair value disclosure requirements, SFAS No. 115 mandates that some securities be accounted for at their fair value, thus directly affecting a firm’s balance sheet and income statement. The main rationale underlying SFAS No. 115 is the reduction of gains trading by financial institutions’ managers (i.e., the ability to choose how and when unrealized securities portfolio gains are recognized into the income statement). Other accounting pronouncements followed, reinforcing fair value accounting’s reach within financial statements, culminating with the enactment of SFAS No. 157 (FASB 2006a). SFAS No. 157 formally defines fair value and frames its measurement and disclosure. The evolution of standards in Canada closely parallels that in the United States with the enactment of the Canadian Institute of Chartered Accountants’ Handbook section 3860 and, more recently, section 3855.

Measurement

Fair value is broadly defined as “the price that would be received to sell an asset or paid to transfer a liability in an orderly transaction between market participants at the measurement date” (SFAS No. 157, para. 5) (FASB, 2006a). The parties to the transaction are assumed to be willing and knowledgeable. Under FVA, assets and liabilities are categorized according to the level of judgment (subjectivity) associated with the inputs to measure their fair value, with three levels being considered. At level 1, financial instruments are measured and re- ported on a firm’s balance sheet and income statement at their market value, which typically reflects the quoted prices for identical assets or liabilities in active markets. It is assumed that the quoted price for an identical asset or liability in an active market provides the most reliable basis for fair value measurement because it is directly observable to market partici- pants and represents the best proxy for the price that would be received assuming a decision to sell the asset. However, in some cases, an asset will not trade or will not have an active market. For these particular cases, if valuation inputs are observable, either directly or indi- rectly, but do not qualify as level 1 inputs, the level 2 fair value assessment of a financial instrument will reflect (1) quoted prices for similar financial instruments in active markets, (2) quoted prices for identical or similar financial instruments in markets that are not active, (3) inputs other than quoted prices that are observable (e.g., yield curve), or (4) correlated prices. For example, a bond issued by Exxon may not trade but, if there is an active market for a Royal Dutch Shell bond with similar terms and maturity, then the market price of the Royal Dutch Shell bond will serve as a level 2 input to value the Exxon bond. Finally, certain financial instruments that, for example, are customized or have no market, will be valued by a reporting entity on the basis of assumptions that presumably reflect market partici- pants’ views and assessments (private placement investments, unique derivative products, etc.). Such valuation is deemed to be derived from level 3 inputs and is commonly referred as “mark-to-model” because it is often the outcome of a mathematical modeling exercise with various assumptions about economic, market, or firm-specific conditions (FASB, 2006a, b): the objective in level 3 is to infer what the price of the asset would be, if the

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 193

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

market existed. In all cases, any unrealized gain (or loss) on financial instruments held by an institution translates into an increase (decrease) in its stockholders’ equity and, conse- quently, an improvement (deterioration) in its capitalization ratios.

Applications

The initial implementation of FVA focused on financial instruments such as debt and equity securities (SFAS No. 115) as well as derivatives (SFAS No. 133) (FASB, 2008b). For the purpose of SFAS No. 115, the appropriate measurement method depends upon management’s intention, which leads to the identification of three types of securities.1 However, while ac- counting treatments between the three types of securities differ, fair value accounting directly or indirectly underlies the measurement of all three. First, debt securities that an enterprise has the positive intent and ability to hold to maturity are classified as held-to- maturity securities and reported at amortized cost. However, if these securities suffer a decline in fair value below the amortized cost basis that is other than temporary, their ac- counting cost must be written down to that value and the write-down must be included in net earnings as a realized loss. Once written down, held-to-maturity securities cannot be written up. Second, debt and equity securities that are bought and held principally for the purpose of selling in the near term are classified as trading securities and reported at fair value, with unrealized gains and losses included in income. Finally, debt and equity securi- ties not classified as either held-to-maturity securities or trading securities are classified as available-for-sale securities and reported at fair value, with unrealized gains and losses ex- cluded from earnings and reported in a separate component of shareholders’ equity (other comprehensive income).

However, it is important to mention that while FVA for financial instruments has at- tracted considerable attention, with specific standards being devoted to its application, it now reaches most aspects of financial reporting. For instance, the goodwill impairment test (FASB 2001; CICA, 2009: section 3062) essentially relies on a fair value assessment of business units. The new approach for consolidation following an acquisition also relies on FVA because it assigns to the noncontrolling interests’ liability a value from a transaction in which the noncontrolling interests were not involved (i.e., the purchase price paid for the controlling interest is assumed to provide an indication of the value of noncontrolling in- terests) (e.g., IASB, 2008; FASB, 2008a; CICA, 2009: section 1602). A similar argument can be made regarding the proposed standard on lease accounting, which would capitalize all lease payments to arrive at a fair value of discounted future cash flows (joint discussion paper, IASB / FASB, 2009). Moreover, because not all changes in fair market value flow di- rectly into net income, most notably because of timing differences (e.g., hedge accounting) or prescribed accounting treatment (e.g., unrealized gains or losses on available-for-sale securities), other comprehensive income has emerged as a way to temporarily soak up such transitory fluctuations until the full transaction cycle is completed.

1. For specialized firms (pension funds, investment banks, etc.), fair market value may apply to all assets and/or liabilities. Moreover, although SFAS No. 115 does not apply to unsecuritized loans, it does apply after their securitization.

194 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

theoretical and empirical Foundations underlying FVA

Conceptual Perspective

The premise for the current application of FVA can be traced back to accounting-based aca- demic research. From a conceptual perspective, Milburn (2008) analyzes how relevant market efficiency is in defining fair value for financial reporting purposes. In a comprehen- sive commentary, and after reviewing prior empirical evidence, he concludes that active, well-regulated, capital markets typically exhibit a reasonable level of efficiency. In his view, a presumption of reasonable market efficiency underlies IASB and FASB positions for fair value relevance in financial reporting. However, the relation between market efficiency and fair value breaks down when considering assets measured using level 3 inputs, which are not based on real or observable market prices. For these types of assets, the use of FVA is less conceptually grounded. Hence, contrary to its stated purpose, SFAS No. 157 does not provide an integrated perspective on the measurement of fair value accounting because it encompasses different measurement bases. Milburn (2008) concludes that more work is needed to assess how to use FVA when conditions of market efficiency are unlikely to be present: such work in finding adequate measurement substitutes should be guided conceptu- ally by market efficiency considerations. In that regard, through an analytical model, Plantin, Sapra, and Shin (2008) show that fair value accounting is most likely the best estimator of underlying values when the assets being held have short-term maturities, are traded in liquid markets that are well normalized, and represent intermediate or junior claims on an entity’s underlying cash flows.

Empirical Perspective

From an empirical perspective, there is consistent evidence, accumulated over the past 20 years, that a firm’s stock price is more closely associated with the market value of its un- derlying financial or real assets than with their historical cost (i.e., their purchase price plus related expenses) (e.g., Barth, Beaver, and Landsman, 2001; Landsman, 2006). The superior relevance of market-derived values is even more obvious in the case of financial derivatives where historical cost is often close to zero, but market value can fluctuate widely (e.g., Venkatachalam, 1996; Ahmed, Kilic, and Lobo, 2006). In other words, fair values, or marked-to-market values, have been found to be more relevant indicators of firm value than traditional historical cost-based figures.2 An early study on the relevance and implications

2. Although studies take many different forms, the most widely used approach closely resembles the following (simplified version of a regression):

Priceit = β0 + β1Assetsit + β2Liabilitiesit + β3Unrealized Gain (Loss)it ,

where i represents a specific firm, and t, a given year-end. Variables are measured in dollars or dollars per share, or are standardized by proxies for firm size. Price equals a firm’s stock market price while both assets and liabilities are as on the balance sheet (consistent with generally accepted accounting principles [GAAP]). Unrealized gains (losses) reflect the difference between an asset market value (according to FVA) and its book value (according to GAAP). FVA-measured information is deemed to be relevant for investors if results from the regression model show that β3 is positive and statistically significant.

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 195

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

from FVA is performed by Bernard, Merton, and Palepu (1995). Prior to the adoption of international financial reporting standards (IFRS) or the advent of Basle II banking regula- tions, Denmark’s accounting standard-setting and banking regulatory authorities relied on mark-to-market valuation for the assets of their commercial banks. Bernard et al. find that Danish banks’ book values, which reflect mark-to-market valuations, provide more reliable information to investors than historical cost-based figures that were then provided by U.S. banks. Moreover, they do not find evidence that Danish bank executives manipulate mark- to-market numbers to circumvent regulatory capital ratios. However, they also point out that that the Danish and U.S. capital markets exhibit significant differences and, thus, their findings may not hold in a U.S. setting.

On the basis of these and other empirical findings, many accounting professors have actively lobbied standard-setters such as the Financial Accounting Standards Board to (1) introduce FVA into financial statements, initially through footnote disclosure, (2) gradu- ally reduce the relative scope of historical cost-derived assets and liabilities in financial reporting, and (3) modify the conceptual framework underlying standard-setting to state more clearly that the primary goal of financial reporting is to provide information that is relevant to investors (presumably, stock market investors) and that, as such, FVA should be emphasized over historical cost.3 Academic research’s influence over the standard-setting process has been greatly enhanced by the involvement of many leading accounting profes- sors promoting FVA into the decision-making process of standard-setters or regulators such as the FASB or the SEC.4 There is currently a joint project between FASB and the IASB to adopt a unified conceptual framework for accounting standard setting. The draft framework, which should be adopted within the next year, clearly states that the main purpose of finan- cial reporting is to provide information that is relevant for investors, with emphasis on market values and cash flow forecasts as the most critical drivers underlying financial reporting.

Overall, from the above review, we can draw some conclusions. First, FVA has deep roots in accounting thinking and practice. Second, there is widespread empirical evidence suggesting that FVA provides relevant and useful information to investors as they attempt to value firms. Third, standard-setters worldwide have a clear and defined agenda with re- spect to FVA, which is likely to result in financial statements further reflecting FVA-derived information. However, the move toward FVA constitutes a major shift in the basic tenets of financial reporting and raises many implementation issues because it changes how manag- ers and other stakeholders view the firm. Such changes are likely to affect the decisions and actions of managers and stakeholders. We discuss these challenges in the next section.

3. For instance, as early as 2000, the Financial Accounting Standards Committee of the American Accounting Association stated that “The Committee generally supports the FASB position that financial instruments be reported in the financial statements at fair value” (Wahlen, Boatsman, Herz, Jonas, Palepu, Ryan, Schipper, Schrand, and Skinner, 2000).

4. For example, a leading proponent of FVA, Katherine Schipper, was a member of the FASB between 2001 and 2006. One of the researchers who pioneered empirical work on FVA, Mary Barth, is currently a member of the International Accounting Standards Board and was previously involved with the American Institute of Certified Public Accountants and the FASB.

196 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

MeAsureMent And VALuAtion CHALLenges surrounding FVA iMpLeMentAtion

undermining the Critical Foundations of Financial reporting

Despite its many tangible or perceived benefits to investors, the adoption and use of FVA undermine several critical foundations of extant financial reporting.

Verifiability and Reliability

Watts (2003: 219) argues that accounting standard-setters should focus on accountants’ core competence—that is, “providing verifiable conservative information that market partici- pants can use both as inputs in their own valuation and as calibration for their own and others’ unverifiable information”. In his view, FVA falls short in this regard because it is often not verifiable or not conservative. He argues that FVA, especially for nonfinancial as- sets, invites excessive managerial discretion. In further work, Watts and Ramanna (2007) argue and find that, for firms subject to SFAS No. 142, the magnitude of goodwill impair- ments decreases as the proportion of unverifiable fair values increases.

Watts’s argument can also be extended to financial instruments that are valued accord- ing to mark-to-model, because these models essentially rely on projections that are unverifiable. Benston (2008) asserts that the definition of fair value contained in SFAS No. 157 (para. 5) raises many implementation problems, especially if fair value is not based on actual market prices. He also considers, consistent with Watts’s argument, that FVA, espe- cially level 2 and level 3 values, facilitates manipulation by managers. Some of the firms involved in the scandals that arose early in the 21st century were using variants of FVA (e.g., see Flesher and Flesher, 1986; Cudahy and Henderson, 2005). The Enron case illustrates the potential negative consequences from using either mark-to-market or mark-to-model accounting, with management strategically selecting bid or ask prices to value its energy contracts and using mark-to-model values for many real asset transactions. Enron was a key market-maker, sometimes the only market-maker, in some markets, thus facilitating mana- gerial discretion (Weil, 2001).

Some recent studies also show that FVA can undermine the reliability of financial state- ments. For instance, focusing on accounting for stock options, Aboody, Barth, and Kasznik (2004) find that managers select valuation model parameters to strategically manage esti- mates of disclosed employee stock option fair values. Their finding raises the broader question of whether managers will behave similarly when selecting model parameters for fair value estimates of other financial instruments. Such a situation provides a striking con- trast to historical cost, for which it is possible to verify exactly what an asset’s purchase price is, as well as related acquisition costs, using reliable sources. Furthermore, in the case of financial instruments that are not traded on an organized market, their valuation for fi- nancial reporting purposes relies on numerous assertions by management, assumptions about the appropriate benchmarks or markets, or the reasonableness of a valuation model inputs, all of which are of varying reliability and verifiability.

Benston’s (2008) views are generally consistent with those of Aboody et al. (2004) and Watts (2003). Focusing on SFAS Nos. 157 and 159, he argues that both standards provide a

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 197

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

potential boon to manipulative managers and extend fair value to all aspects of financial re- porting since the Board “also believes that, with the passage of time, historical prices become irrelevant in assessing an entity’s current financial position” (excerpt from SFAS No. 159, para. A3).

Conservatism

Within a conservatism perspective, financial statements anticipate bad news (i.e., before a transaction is actually concluded; hence, an asset is written down if it is deemed that it has suffered a permanent impairment or if expected economic conditions suggest that the firm will not be able to recover its value). Moreover, such a write-down is permanent (i.e., the asset will not be reevaluated upward in the future even if economic conditions change in the meantime). Still within a conservatism perspective, financial statements will only reflect good news if there is an arm’s-length transaction that validates the value of such news: the impact of any appreciation in the value of an asset or of the signing of a new contract will be reflected in a firm’s financial statements only if the asset is actually sold or converted into revenue from operations. In contrast, within a FVA perspective, both realized and un- realized losses and gains are recognized on financial statements. Moreover, assets that have been marked down can be reevaluated upward. IFRS extend this practice into a new realm as they allow fair value write-ups for operating assets.

Watts (2003) observes that FVA induces severe cost asymmetry whereby it is much costlier to make an excessive payout (based on earnings that contain unrealized FVA esti- mates) that cannot be recovered than to retain earnings and other resources one period too long. So markdowns should require less verification than markups. Seemingly unbiased FVA is open to an excessive degree of moral hazard when it serves as a benchmark for divi- dend payouts and regulatory ratios like bank capital. In contrast, Barth (2007: 12) argues that, “Although opponents of more comprehensive use of fair value have some legitimate concerns, standard setters are unaware of a plausible alternative.” The debate is on two levels. On one hand, Barth is talking about the measurement of a final output. On the other hand, Watts refers to the validity of the various measurement inputs under moral hazard, the out- put being of some importance but mostly in terms of providing financial statement users and other stakeholders with the ability to adapt, modify, or “test-drive” the resulting output. Beyond fair values, measurement assumptions and hypotheses are probably more critical because they allow users to reconstruct the reality according to their own priors.5

5. There is an alternative position that fair value accounting may be too conservative under some conditions, contradicting the going-concern assumption. For instance, Benston (2008) argues that fair value accounting-based information may be useful to some creditors and shareholders of firms facing liquidation, but not to investors in other firms with no going-concern uncertainty. Because the fair value of an asset is defined as “the price that would be received to sell an asset … at the measurement date” (SFAS No. 157, para. 5), it implicitly assumes that, at the end of each reporting period, an entity sells its assets at market or model-estimated prices at that same time. A liquidation balance sheet is not prepared very differently.

198 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

revisiting Accounting’s position within society

Financial Reporting as a Social Good

The implementation of FVA explicitly confirms the primacy of financial markets and inves- tors in the determination of accounting standards. The broader social issues and implications arising from accounting standards for stakeholders beyond investors are assumed away. The potential danger of relying on capital-market-based findings to directly prescribe account- ing standards was highlighted more than 30 years ago by Gonedes and Dopuch (1974). Following a first wave of capital-market-based studies that mapped their findings directly into standard-setting issues, Gonedes and Dopuch explain that observing an empirical rela- tion between accounting amounts and equity prices or returns does not provide sufficient evidence about the desirability or effects of a particular standard, even if markets are infor- mationally efficient. Their conclusion rests on the fact that accounting standards are essentially a public good. Therefore, standard-setters’ mandate and responsibility are to de- velop standards after making the appropriate social welfare trade-offs, which involve more parties than just investors, such as creditors, governments, regulators, and employees. Hence, deciding about a particular accounting standard requires that social preferences be specified. From a different angle, Holthausen and Watts (2001) argue that the value-relevance literature has little to say about standard-setting issues. In their view, without an underlying theory that explains, predicts, and links accounting, standard setting, and valuation, value- relevance studies simply report statistical associations.

Increased Volatility in Earnings

Most prior research shows that the adoption of FVA translates into more volatile financial results (earnings).6 One reason underlying this finding is that not all assets and liabilities are marked to market (Ryan, 2008). For instance, assuming a rise in market interest rates, FVA may translate into somewhat artificial volatility if one side of the balance sheet is val- ued at mark-to-market (e.g., investments that would presumably go down in value), while the other side relies on another method (e.g., actuarial reserves for an insurance company that are estimated using specified long-term rates or deposits for a bank). However, if both sides are valued at market, the decrease in the value of investments will be offset by a cor- responding decrease in the value of liabilities being discounted at higher rates, as long as the durations of assets and liabilities are matched.

In that sense, FVA may induce pro-cyclicality in reported earnings because it adds un- realized gains to earnings in an up market, thus making the subsequent drop in reported earnings in the down period even more dramatic (Laux and Leuz, 2009). One example illus- trates the potential impact of FVA on the volatility of reported earnings.

Crédit Suisse: On February 12, 2008, Crédit Suisse, one of the world’s leading commercial banks, reports record income from continuous operations of 8.5 billion Swiss francs. Seven days later, on February 19, 2008, Crédit Suisse announced that

6. For example, Barth, Landsman, and Wahlen (1995) show that fair value-based measures of net income are more volatile than historical cost-based measures.

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 199

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

some additional control processes have led to the repricing of certain asset-backed positions in its structured credit trading business, with the current total fair value reduction of these positions being reduced by an estimated $U.S. 2.85 billion. A month later, on March 20, 2008, Crédit Suisse reported that its 2007 operating in- come had been revised downward by 1.18 billion Swiss francs (789 million Swiss francs after tax), reducing initially reported 2007 earnings by around 10 percent.

The Crédit Suisse case illustrates the difficulty of pinning down the fair value of many assets when the underlying valuation methodology is complex and subject to shifting hy- potheses and assumptions about the future. Crédit Suisse’s experience also shows that reported results for a given period may be subject to a wide margin of error or discretion or even restated.

The volatility potentially induced by FVA implies that the reported values of financial institutions’ assets can fluctuate more than business fundamentals would suggest: a financial institution’s value goes beyond the separate market values of its different assets and liabili- ties and incorporates reputation, client networks, fee-producing services, etc. If there are downward market trends, which translate into a severe shrinking of their capital ratios, fi- nancial institutions may be forced to deleverage and sell additional assets at distressed prices, thus feeding a downward spiral. Fundamental values may not be going down as fast or as deeply. In that scenario, the issue is not necessarily the accounting itself but how fi- nancial regulators use such accounting information. First, firm value is quite different from the sum value of its assets and liabilities. Second, FVA-based financial reporting is only the messenger that a firm’s solvency is undermined by its financial strategies or lending prac- tices. It is up to regulators to figure out how to use such information.7

Control

By emphasizing market- or model-based measurement, using FVA also affects the relative role of accountants in the preparation of financial statements. Although historical cost- based financial statements are squarely under the control of accountants, FVA-derived assets and liabilities often require the expertise of other professionals such as actuaries, valuation experts, or financial engineers, with accountants being more likely to play a sec- ondary role (e.g., verifying underlying assumptions or testing hypotheses). This raises questions as to the education and training that accountants now need in this new financial reporting reality. Moreover, the advent of FVA requires accountants to have a better under- standing of markets: how they work and when they do not work. Allen and Carletti (2008b) provide some useful insights in that regard. They argue that markets work when market prices reflect future payoffs. However, in times of financial crisis (such as the Russian crisis of 1998, which saw the demise of Long Term Capital Management), prices diverge from the asset’s underlying value and instead reflect the amount of cash or liquidity that is avail- able to buyers who are active in the market. Allen and Carletti argue that such conditions

7. For instance, Irene Wiecek argues that “the credit crisis is not the fault of accounting. It is the fault of overly lenient lending practices” (Jeffrey, 2008). Furthermore, she adds that there was a lack of oversight and regulation in the area.

200 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

may lead to contagion and force banks into insolvency despite the fact that they would be able to fully pay up their obligations if they were able to operate until the assets mature. Under such a context, the question arises as to what accountants should be measuring and reporting, especially if the assumption of going concern is maintained.

Summary

Several conceptual and practical arguments have been raised against the use of FVA in fi- nancial reporting. There is also some evidence that it may facilitate managerial discretion. However, one of the most serious and intense attacks against using FVA in financial report- ing has occurred during the last few months as the world’s financial system teetered on the brink of an abyss. The next section reviews the arguments suggesting that FVA may have contributed to the financial crisis and analyzes the emerging academic evidence in this regard.

FVA And tHe FinAnCiAL Crisis

the debate surrounding FVA’s role in the Financial Crisis

Viewpoints from the Financial and Regulatory Communities

The 2008 financial crisis was characterized by extreme volatility in financial markets as well as by the failure or quasi failure of many financial institutions such as Bear Stearns, Lehman Brothers, AIG, Merrill Lynch, Citicorp, Royal Bank of Scotland, and Dexia. Fair value de- tractors, among them David Dodge, the former governor of the Bank of Canada, argue vehemently that FVA has accelerated and amplified the current financial crisis (e.g., Mc- Farland and Partridge, 2008). Their argument can be summarized as follows. Starting in 2007, the drop in the price of many types of financial instruments led financial institutions to mark down the asset values reported on their balance sheets, thus weakening their capi- talization ratios. To improve their financial profile and to enhance their safety zone with respect to regulatory capital requirements, these institutions started to sell securities or close down positions on some financial instruments in markets that were increasingly illiquid as a result of the emergence of a liquidity crisis. These sales magnified the downdraft in quoted prices, thus bringing additional devaluations. Along these lines, William Isaac, former chair of the U.S. Federal Deposit Insurance Corporation, argues that “mark-to-market accounting has been extremely and needlessly destructive of bank capital in the past year and is a major cause of the current credit crisis and economic downturn” (Jeffrey, 2008: 27).

However, FVA can count on broad support from the accounting profession, standard- setters, and regulators. For instance, in a recent speech, Nick Le Pan, Canada’s former superintendent of financial institutions, argued that FVA is only a messenger and should not be criticized for merely reflecting the poor underlying economic outlook (McFarland and Partridge, 2008). Barbara Roper, from the Consumer Federation of America, argues that sound accounting principles, such as FVA, led to the exposure of underlying problem assets. In her view, FVA provides more accurate, timely, and comparable information to investors than any other accounting alternative. The CFA Institute, representing the financial analysts’ community, has also vigorously supported FVA, with open letters to the SEC and various position papers (CFA Institute, 2008).

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 201

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

Viewpoints from the Academic Community

Two comprehensive commentaries provide a nuanced interpretation of the role of FVA in the 2008 financial crisis and are consistent in their conclusions with the findings reported by the SEC in December 2008. Ryan (2008) concludes that fair value accounting did not contribute to the crisis because it provided investors with needed accurate and complete in- formation concerning subprime positions. However, he concedes that there is a need for additional guidance as to what is an orderly transaction, because the use of inputs for fair value accounting is driven by conditions at the measurement date.

Laux and Leuz (2009) conclude that FVA probably did not contribute to the crisis to such a large extent but was also not merely a messenger. In their opinion, the FVA debate is a new version of the relevance versus reliability argument. Moreover, there are legitimate concerns about marking asset values to market prices in times of financial crises because they are tied to various contracts and regulations (e.g., capital ratio regulations for banks). However, the standards do allow for some flexibility and deviation from market prices under these conditions. They consider that the implementation of FVA may lead to some disrup- tions from the standards’ initial intent. For instance, SEC enforcement actions can imply that auditors and firms abide by certain rules or practices, not using the judgment initially envisioned in the standards. In contrast, too much managerial discretion in the determina- tion of fair values may translate into manipulations. The challenge for standard-setters is to find an appropriate balance between the potential contagion effects from using fair value accounting and investors’ needs to receive timely information about impairments. Overall, FVA may be pro-cyclical, accentuating the magnitude of earnings in booms and of losses in busts. Hence, the use of fair value accounting information needs to be adapted, but the underlying accounting standards do not need to change. They also observe that historical cost accounting is unlikely to solve problems of market inefficiency and that its use for loans contributed to the securitization rush that led to the crisis.

Fair Value Accounting and the Financial Crisis: the evidence

Analytical Insights

Two analytical papers model how reliance on fair value accounting affects managerial deci- sion making and financial markets. Their conclusions raise some concerns over the use of FVA in financial institutions, especially by regulators, and provide useful insights into the 2008 crisis.

Allen and Carletti (2008a) analyze the impact of FVA through a two-industry model (insurance and banking). They put forward two contrasting views. On the one hand, FVA provides the relevant and true value of underlying assets. On the other hand, FVA induces excessive and artificial volatility that does not reflect underlying value. They conclude that using market prices to value assets may not be beneficial when financial markets are illiquid. More specifically, during a liquidity crisis, prices do not reflect the appropriately discounted value of future expected cash flows but are instead determined by the cash available in the market at that moment. However, through FVA, these market conditions affect bank asset values as well as portfolio contract choices and induce default contagion across financial institutions, especially for those with long-term assets (versus a situation where historical

202 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

cost is used). Hence, according to Allen and Carletti, market prices under conditions of market illiquidity should be ignored because they understate fundamental values and distort portfolio and contract choices.

Plantin et al. (2008) start with the premise that most banks’ assets, especially loans, are traded over-the-counter (OTC). In their view, OTC markets do not have the same depth as normalized regulated markets. Relying on an analytical model, they conclude that mark-to- market accounting injects excessive volatility into transaction prices (i.e., accounting standards are an endogenous source of volatility). In other words, prices drive accounting measurement but measurement has an effect on pricing. More specifically, mark-to-market accounting injects artificial risk into the market that degrades the information value of prices, thus inducing suboptimal real decisions such as asset allocations or investment horizons. In their view, the value distortion is greatest when assets are (1) long-lived, (2) illiquid, and/or (3) senior, leading to a shortening of managers’ time horizons. More specifically, the use of mark-to-market leads to a shortening of decision horizons among firms, mostly because of agency problems. Thus, firms become more sensitive to short-term price movements, which are magnified for long-lived assets with their cash flows extending far into the future. In addition, illiquidity implies that the price of an asset responds to the decisions of other fi- nancial institutions, thereby creating an endogenous risk. In this regard, their argument differs from that of Allen and Carletti (2008a), who posit that illiquidity leads to asset prices being a function of liquidity available in the market. Finally, senior securities will be most affected because there is typically not much upside potential with respect to their value. In these cases, for reporting purposes, historical cost accounting may actually dominate fair value accounting because it induces less distortion into asset values. However, historical cost accounting is typically inefficient because it ignores price signals.

The findings from Allen and Carletti (2008a) and Plantin et al. (2008) suggest that the reliance on FVA for long-lived senior assets induces suboptimal decisions by managers when markets are illiquid. Moreover, in illiquid markets, price information is itself affected by the use of FVA and may become an unreliable value measure. Finally, under conditions of illiquidity, the use of FVA facilitates the propagation of distorted values into financial institutions’ balance sheets, potentially inducing failure contagion. Both studies suggest that the use of historical cost accounting alleviates some of these problems and may be a better measurement tool under some conditions.

Empirical Evidence

Early empirical evidence in the aftermath of the crisis supports some of the analytical find- ings presented above. Moreover, they suggest that FVA may have played more than just a messenger role and justify some of the concerns raised about the implementation of FVA and its lack of internal consistency regarding measurement.

Kolev (2008) examines the value-relevance of the three levels of assets reported under fair value accounting by U.S. commercial banks in the first two quarters of 2008. His sample comprises 177 banks in the first quarter and 172 banks in the second quarter. His basic valu- ation regression is the following:

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 203

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

Share priceit = β0 + β1Net Equity Book Valueit + β2Level 1 Net Assetsit + β3 Level 2 Net Assetsit + β4 Level 3 Net Assetsit + Other control variables,

where i indicates a given firm; t indicates a given quarter; and fair value is defined in terms of net assets (i.e., assets less liabilities).

Kolev (2008) finds that, in the first quarter, investors perceive reported values under all three levels to be value-relevant, with minor differences between β2 (0.811), β3 (0.772), and β4 (0.709) (see Kolev’s table 4). In other words, one dollar of level 1 net assets is valued at $0.81 by stock market investors. However, the valuation coefficients decline between the first quarter and the second quarter, reflecting greater uncertainty about underlying asset values: all levels of FVA net assets now sell at significantly less than their book value. Moreover, there is now a valuation gap between the valuation of level 1 net assets (β2 = 0.604), level 2 net assets (β3 = 0.582), and level 3 net assets (β4 = 0.419), the latter being valued at a significant discount by investors. The extent of the valuation gap is reduced if the expertise of audit committee members is enhanced.

Goh, Ng, and Yong (2009) conduct a similar investigation for the first three quarters of 2008. After removing observations without stock market data or with missing financial data, their sample comprises 516 banks. Their basic valuation regression is similar to Kolev’s but includes a more comprehensive set of control variables.

They find that, in the first quarter, investors perceive reported values under all three levels to be value-relevant, with minor differences between β2 (0.743), β3 (0.650), and β4 (0.583) (see Goh et al., 2009: table 8). In other words, one dollar of level 1 net assets is val- ued at $0.743 by stock market investors, a significant discount to book value. However, by the second quarter, there is a significant shift in the valuation of FVA assets. On one hand, the valuation of level 1 net assets now matches their book value (β2 = 1.086), a significant increase from the first quarter. On the other hand, investors significantly downgrade the valuation of level 3 net assets, which are now valued at a deep discount to their book value (β4 = 0.296). The valuation of level 2 assets remains more or less the same between the two quarters. The situation does not change much in the third quarter, with level 1 net assets still being valued at close to book value (β2 = 1.153), while level 2 net assets (β3 = 0.512) and level 3 net assets (β4 = 0.254) remain valued at a discount. A higher capital ratio and a higher-quality auditor do allow a bank to enhance the value-relevance of its level 2 and level 3 assets close to their book values (i.e., coefficients close to 1.00).

Song, Thomas, and Li (2009) look at the valuation relevance of both assets and liabili- ties reported by a sample of 431 banks during the first three quarters of 2008. In contrast to Kolev (2008) and Goh et al. (2009), they do not report distinct valuation figures by quarter and do not combine fair value assets and liabilities together for regression purposes. Other- wise, their basic valuation regression resembles the following:

Share priceit = β0 + β1Net Equity Book Valueit + β2Level 1 Assetsit + β3Level 2 Assetsit + β4Level 3 Assetsit + β5Level 1 Liabilitiesit + β6Level 2 Liabilitiesit + β7Level 3 Liabilitiesit + Other control variables,

204 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

where i indicates a given firm; t indicates a given quarter; and fair value is defined in terms of net assets (i.e., assets less liabilities).

Song et al. find that, on average, level 1 (β2 = 0.968) and level 2 assets (β3 = 0.972 ) are valued by investors at close to their reported book (i.e., fair market) values (see Song et al.: table 3). While value-relevant, level 3 assets are priced at a significant discount (β4 = 0.683). Moreover, both level 1 (β5 = −0.818) and 2 (β6 = −1.006) liabilities are also valued at close to their book (i.e., market) values. However, level 3 liabilities are priced at double their reported fair market values (β7 = −2.185), thus implying that investors severely doubt the reliability of FVA-based liabilities. Song et al. also observe that strong (weak) govern- ance enhances (reduces) the value relevance of fair value measures reported by banks, especially for level 2 and level 3 assets and liabilities. Governance encompasses board in- dependence, audit committee activity, ownership, internal controls, and auditor size.

The key conclusion to draw from these findings is that, in the eyes of investors, level 3 asset (liability) fair values are manipulated upward (downward). Moreover, under crisis conditions (i.e., second or third quarter of 2008), it appears that investors revise downward the valuation of level 2 and, especially, level 3 assets because there is more uncertainty and potentially more managerial discretion underlying their measurement. However, the evolu- tion in the valuation of level 1 assets is more difficult to pin down, with Kolev (2008) reporting a fall in their valuation and Goh et al. (2009) reporting an upward reevaluation. Potential explanations for this discrepancy include different samples (Goh et al.’s being more comprehensive) as well as different empirical models. Further work is needed to rec- oncile both sets of findings.

Khan (2009) looks into how using fair value accounting may translate into failure con- tagion within the banking system. His sample period encompasses the 1988 – 2007 period and includes all U.S. bank holding companies with available stock market data (for up to 98,162 bank-month observations). His variable of interest is banks achieving extremely poor stock market returns in a particular month (bottom decile of the sample). He observes that the banking system’s proportion of assets recorded at fair market value reached around 70 percent in the summer of 2007. He reports three key findings. First, he shows that the extent of a bank’s use of fair value accounting increases its propensity to achieve poor stock market performance (i.e., to exhibit failure contagion attributes). Second, he finds that the less liquid the market, the greater the failure contagion effect. Third, the illiquidity condi- tion and the extent of a bank’s reliance on FVA reinforce each other in exacerbating the failure contagion effect. However, well-capitalized banks were less likely to be affected by the contagion effect than banks with weak capital ratios. His findings are consistent with the analytical predictions derived by Allen and Carletti (2008a) and Plantin et al. (2008) in showing how the use of FVA in contexts of illiquid markets such as 2008 facilitates the propagation of distorted values into financial institutions’ balance sheets, thereby inducing failure contagion.

It is still too early to reach conclusions concerning FVA’s role in the current financial crisis: not all data are available, additional analyses must be completed, and all its conse- quences cannot be observed. However, relying on prior research findings and on available data, it is possible to draw some inferences about the contribution of FVA to the financial

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 205

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

crisis. Overall, findings reported by Kolev (2008), Goh et al. (2009), and Song, Thomas, and Yi (2009) suggest that (1) FVA-derived information is value-relevant, consistent with extensive prior research; (2) investors perceive level 3 FVA asset and liability values to be measured less reliably and, accordingly, price them at substantially less than 100 percent; (3) strong governance and capital ratios enhance confidence in reported figures; and (4) by the end of summer 2008, investors were significantly discounting level 3 input values, con- sistent with a perception that management was exerting significant discretion in its measurement. The implicit prices assigned by investors to level 1 assets rose over the year, suggesting that investors perceived market prices to be depressed. To some extent, Khan (2009) complements these findings by showing that the pre-crisis period, the use of FVA, and situations of market illiquidity did induce a failure contagion effect.

FAir VALue ACCounting, tHe MArKets, And inVestors: ConsiderAtions For tHe Crisis And beyond

FVA and option Compensation

Many financial institutions involved in the current crisis made extensive use of stock options or other equity-based awards to reward their senior executives, allowing unrealized gains on assets to be converted into cash. For example, the compensation packages of Richard Fuld (Lehman’s CEO), Stanley O’Neal (Merrill Lynch’s CEO), and Charles Prince (Citi- corp’s CEO) were laden with equity-based instruments. The increased volatility brought forward by FVA is conducive to the use of equity-based compensation, especially stock options, because their value is then enhanced. According to the Black-Scholes model, vola- tility is positively associated with option values. Prior research suggests that there is a strong association between performance volatility and the use of stock options (e.g., Magnan, 2006). Through FVA, the outcomes from aggressive risk taking in investment and financing strategies will directly flow into reported earnings, thus further leveraging the potential gains to be derived from stock options and other incentives, and making these compensation strategies even more attractive to management.8

does FVA reflect underlying business performance or Allow Financial institutions to delay the day of reckoning?

Some of the fiercest critics of FVA argue that, far from enhancing transparency and relevant financial reporting, it actually provides corporate managements with ways to avoid the day of recognition and to delay asset impairments. For instance, Watts (2003) argues that the elimination of conservatism brought by FVA leads to the capitalization of unverifiable fu- ture cash flows on the balance sheet.9 Such unverifiability and managerial opportunities to make strategic valuation choices introduce significant noise into the financial reporting

8. It is relevant to note that, prior to the change in Canada’s Bank Act in 1992, Canadian banks were forbidden to offer stock options to their executives.

9. There is empirical evidence that effective auditing of FVA-derived numbers requires very specialized valuation knowledge that may be difficult for auditors to gain and maintain (Martin, Rich, and Wilks, 2006).

206 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

process that may be costly to investors. Moreover, by moving firms away from transaction- based accounting, FVA is contradicting SEC efforts to tighten revenue measurement and recognition standards to ensure that only completed sales transactions are reported in earn- ings. For instance, in 1999, the Securities and Exchange Commission released Staff Accounting Bulletin No. 101 (SAB No. 101) (SEC, 1999), which prescribes specific criteria to indicate when a transaction has been concluded, thus considerably reducing managerial discretion in the recognition of revenues. In contrast, fair value accounting does not rely on the conclusion of a transaction to estimate the value of a financial asset or contract. Experi- ence shows that, until the advent of SAB No. 101, several firms had applied aggressive revenue recognition criteria that dramatically boosted reported earnings and growth rates. Earnings restatements following the enactment of SAB No. 101 were often sizable and led to significant stock price falls, even if reported cash flows were not affected. In other words, conservative accounting provides information that is useful beyond the estimated cash flows from a particular contract and protects investors and creditors from managerial opportun- ism. However, from a strict FVA perspective, were the accounts receivable reported by these firms unwarranted or unjustified because they reflected future payoffs? The economic asset certainly existed: the issue was to what extent it had been earned.

The case of Lehman Brothers is consistent with Watts’s argument. As of November 30, 2007, 75.1 percent of fair value assets were measured according to level 2 or level 3 inputs. In other words, the majority of assets supposedly valued at fair value were not valued on the basis of directly observable quoted prices. By May 31, 2008, that proportion had in- creased to 81.7 percent of assets measured at fair value, suggesting that barely 18 percent of assets supposedly valued according to FVA were marked to market. Further empirical work as well as the liquidation of Lehman Brothers will eventually provide additional evi- dence regarding the extent to which its assets may have been overstated or purposely shifted into level 2 or 3 to hide developing losses and give management more discretion. At the very least, it suggests that FVA reporting may work well for investors when assets trade in deep and efficient markets but may become less transparent when market conditions become more difficult or less liquid. It is telling that Lehman Brothers was an early adopter of both SFAS No. 157 (fair value measurements) and SFAS No. 159 (FASB, 2006b), deciding to implement their provisions in the first quarter of its 2007 fiscal year.

The Lehman case, as well as many others, raises the issue of FVA’s applicability as it is being extended from instruments traded in liquid and organized markets to credit-type instruments that are often securitized and that are not transparent about their underlying as- sets. The valuation of these credit-type instruments is made difficult by the lack of direct information. Moreover, the market for these instruments is not as deep and liquid than as that for traditional instruments such as bonds, equities, or foreign currencies. Apparently, markets were not efficient in assessing the value of these structured investment vehicles or securitized pools of assets, and investors may have relied too much on the judgment of par- ties such as credit-rating agencies that had partial information and were facing some potential conflicts of interests because they charged fees to render opinions on specific securities.

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 207

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

Accounting and the Market: Mirrors Facing each other

The mapping of market values on corporate balance sheets mandated by accounting standard- setters contrasts with the trend by many analysts and sophisticated investors to use financial statement data to gauge whether a firm’s stock market value has moved away from its fun- damental or intrinsic value. In that regard, Lee, Myers, and Swaminathan (1999) show that value and price are co-integrated but can diverge from each other when information acquisi- tion is costly (i.e., prices can veer away from fundamentals). Moreover, they also show that value-relevance has two dimensions. When price is a noisy measure of value, value-relevance may be assessed according to the ability of an information signal to contribute to return prediction (instead of current value), which suggests that accounting standard-setters’ map- ping between value-relevance and FVA may be defined too narrowly and that both dimensions could potentially be considered.

These divergent trends raise a fundamental question as to the grounding of financial statements. Macintosh, Shearer, Thornton, and Welker (2000) argue that the market uses accounting earnings, along with other information, to value firms’ stock and other securi- ties. However, the prices of many of these securities underlie derivatives’ prices, which then find their way into financial statements through FVA, thus completing a circular sequence! As Macintosh et al. (2000: 36) say: “Companies’ earnings determine security prices, which determine derivative prices, which determine companies’ earnings. In short, neither the ac- counting sign nor the financial market sign appear to be grounded in any external reality. Instead, each model appeals to the other model for the only ‘reality check’ available.”

According to Macintosh et al. (2000), financial reporting becomes a self-referential simulacrum, potentially detached from reality, or a hyper-reality. Lehman Brothers’ experi- ence in the months leading to the crisis provides a recent illustration of that trend. During its last complete fiscal year before its bankruptcy (2007), Lehman Brothers reported earn- ings of around $4.3 billion. Investors presumably relied on Lehman’s reported earnings to assess its prospects and value its shares. The chain of decisions exactly matches the above quotation from Macintosh et al. A significant proportion of Lehman Brothers’ assets was actually shares and share-based derivatives (more than a third of its assets measured accord- ing to FVA). Shares traded on a stock market are all affected to a varying degree by the same secular trends and fluctuations (e.g., systematic risk). Hence, one can argue that Lehman Brothers earnings and its stock price were mutual reflections of one another, possibly detached from underlying real operations. Such a conclusion can probably be extended to many financial institutions deeply involved in the current crisis or engulfed by it.

Lehman Brothers’ equity-based compensation provides another illustration of the self- referential sequence that FVA introduces into financial reporting and stock market prices. In 2007, Lehman granted close to 39 million deferred share units to its executives and em- ployees. On the basis of the firm’s quoted stock price on the dates at which these grants were made, the overall value of the grant was around $2.7 billion. Since 2006, SFAS No. 123 (FASB, 1995) has mandated the measurement and recognition of equity-based compensation at fair value, using an amortization method for grants that have a long-term vesting period. Thus, in 2007, Lehman Brothers reported an expense for equity-based compensation of $1.8 billion, close to 25 percent of earnings before income taxes and equity-based compensation

208 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

expenses. Hence, the amount reflected as an expense by Lehman on its financial statements reflects the current quoted price of its stock at grant date. Moreover, as its stock price fluctu- ates, Lehman’s reported expense will fluctuate as well since the amortization of the deferred equity compensation grants is reevaluated according to current market prices (in contrast to stock options, which are valued only at the grant date). Therefore, a significant portion of Lehman’s reported earnings in the period leading to the fall 2008 crisis were actually driven by Lehman’s own stock price. One can further assume, if financial statements had been available, that compensation expense would have shown a significant drop in the last quarter of 2008 as Lehman’s stock price went down.

Messenger or Contributor? A disCussion

The above discussion suggests that assigning a messenger role to accounting potentially downplays its importance and relevance to the current crisis because the message is not neutral but conditioned by accounting standards. However, two issues arise from the use of FVA-derived information in regulatory oversight. First, FVA information is highly volatile and sensitive to market prices, which can enhance or distort investors’ perception of value during market bubbles or crises. For example, under FVA, the wild fluctuations of the stock market during the last quarter of 2008 and the first quarter of 2009, with some daily closings showing gains or losses from the preceding day of 5 – 10 percent, imply similar fluctuations in any stock-market-based assets. Hence, a firm may be solvent one day (assuming a large stock market gain), insolvent the next two days (assuming large stock market losses), and solvent again on the fourth day. While informative, is FVA-based financial reporting useful to regulators in planning and timing their interventions? The answer is that FVA information alone is probably necessary but is not sufficient. Other performance and risk metrics are needed to identify the targets of regulatory actions. A similar argument can be used to justify that FVA information is not sufficient for long-term governance purposes because it is not stable enough and is difficult to verify. The reliance on FVA-based information may have two opposite implications regarding the length and severity of the current crisis. On one hand, the discretion underlying FVA figures may have allowed managers to delay the day of reckoning when underlying subprime assets started to unravel. Moreover, the sensitivity to market prices that it introduces into financial statements may have amplified the impression of financial performance and solid capitalization in the bubble period. On the other hand, once the values of underlying assets started crashing, FVA-induced balance sheet realign- ments and recapitalizations may have further magnified the crisis.

Second, some argue that FVA values are not the issue. Rather, the issue is the quality of the accompanying disclosure (Leone, 2008). For example, Susan Schmidt, a former governor of the Federal Reserve Board and bank chief financial officer (CFO), argues that the focus should be on disclosure so that everyone, regulators and investors alike, under- stand the drivers behind fair value estimates. FVA information can be deceptive: until close to the crisis, both Lehman Brothers and AIG appeared solvent and sufficiently capitalized, with significant portions of their balance sheet relying on FVA. However, both firms’ FVA point estimate values did not tell the extent of the downside risk they were facing if events did not evolve according to expectations, Lehman because of its exposure to the increasingly

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 209

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

illiquid collateralized debt obligations market and AIG because of its exposure to rapidly deteriorating credit-default swaps.

Looking at both firms’ financial statements, it would have been difficult to assess the potential magnitude of losses to be incurred because of these exposures. For example, in the management, discussion, and analysis (MD&A) section of its 2007 annual report, Lehman Brothers reports that average daily trading revenue volatility increased from $35 million in 2006 to $48 million in 2007 (these figures are inferred from Lehman’s value- at-risk (VAR) computations based on the previous rolling 250 days). Lehman also reports that its largest daily trading loss was $137 million in 2007 (versus $59 million in 2006). While both sets of figures suggest a more volatile environment in 2007 than in 2006, they pale beside Lehman’s total FVA asset exposure of $291 billion at the end of 2007 (FVA lia- bility exposure was $149 billion): the uncertainty underlying the measurement of FVA assets and liabilities is likely to be several times the VAR figures reported in the MD&A. Hence, we can postulate that FVA without adequate additional disclosure is neither fair nor a good reflection of value that is at risk.

ConCLusion

The purpose of the paper is to briefly present FVA, its origins, application, and implications for financial reporting as well as its potential role during the current financial crisis. Al- though no definite conclusion can be reached at this early stage, there is reason to believe that fair value accounting is more than just a messenger carrying bad news and, therefore, may have contributed to the acceleration of the crisis, especially in the financial sector. While the relevance of FVA for investors cannot be questioned, its other qualities (or weak- nesses) may have been overlooked by standard-setters and regulators.

FVA for financial instruments is part of a broader trend in accounting standard setting to move away from “accounting” toward estimating expected future cash flows and incor- porating into financial statements (i.e., “forecounting”) (Magnan and Cormier, 2005). The trend undermines decades, if not centuries, of accounting practices and concepts such as conservatism and verifiability, and requires a complete set of valuation skills and knowledge from accountants. The current crisis constitutes the first serious challenge to this trend, and to FVA in particular, and is likely to generate abundant empirical research over the next few years, which will allow us to better assess the pros and cons of fair value accounting.

However, if not FVA, what else? Standard-setters and accounting academics argue that there is no alternative measurement or reporting model (Barth, 2007). However, various fi- nancial and economic interests are at play, such as additional powers for standard-setters, additional business for providers of accounting and valuation services, and increased un- certainty about bonuses for managers and executives. Hence, viewpoints and arguments from interested parties must be reframed accordingly. The debate goes further than account- ing and financial reporting and deals with the essence of what accountants are expected to contribute to society and, implicitly, what competences and skills they must possess to de- liver in that regard. One may surmise that current accounting standards, such as those relating to fair value, overstretch accountants’ current capabilities and prior learning and obscure other informational needs by investors and other interested stakeholders.

210 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

The IASB is revisiting the issue once again and has recently issued exposure drafts on fair value measurement and financial instruments’ classification: the IASB (2009a) expo- sure draft provides additional guidance on the measurement of fair value while another IASB (2009b) exposure draft overhauls and replaces the current standard for financial in- struments. With respect to the classification of financial instruments, the IASB proposes two measurement categories: amortized cost and fair value (IASB, 2009b). The classification into amortized cost is criterion-based: (1) Does the financial instrument have basic loan features? (2) Is the instrument managed on a contractual yield basis? Both criteria imply contractual and determinable cash flows, which typically can be estimated and verified. Fi- nancial assets measured at amortized cost are subject to a single impairment model. All equity instruments are reported at fair value.

On the surface, the IASB’s approach appears consistent with theoretical and empirical findings that suggest we need to rethink our blanket assumption that FVA works well under any context or conditions. Nevertheless, the assignment of financial instruments into two distinct categories with different measurement bases may be a deceptively simple solution because financial innovations over the past 30 years have clearly told us that financial instru- ments can be designed to fit specific purposes, even pass a particular criterion-based test. In addition, equity instruments also have the right to share cash flows, albeit on a residual basis, and are certainly exposed to the same market turbulences as contractual-type instru- ments (e.g., illiquidity). Hence, it remains to be seen whether these measures will address the core issues raised by many commentators.

In a commentary in the Financial Post, Thornton (2009) argues that, as accountants, “We need to start with a clean sheet of paper in revising the accounting so that it clearly and credibly distinguishes lemons from cream puffs.” In reviewing the market conditions during the crisis, he notes that auditors ought to drill down into bank assets to assess their underly- ing value, more specifically by verifying expected future cash flow projections. If market prices diverge from underlying or intrinsic values for lack of liquidity reasons, then such an audit would validate management’s claim. Otherwise, the securities being held are probably “lemons”, similar to used cars sold by unscrupulous car salespeople, and should be marked to market in the absence of audited verification of level 3 inputs. Thornton’s perspective on fair value is a call for accountants to regain control over the financial reporting process in- stead of passively relaying market-induced values. His views are also consistent with Allen and Carletti’s (2008b) view that knowledge of underlying markets is critical for reporting purposes. Moreover, his suggestion is also consistent with empirical findings that the valu- ation of fair value assets is enhanced when higher-quality auditors or expert audit committees are involved (see Kolev, 2008; Goh et al., 2009; Song et al., 2009). The intervention of ex- ternal parties such as auditors to drill down into the parameters underlying fair value figures also breaks the endogeneity loop referred to by Macintosh et al. (2000). However, he offers a major challenge to our profession to retool itself and rethink how auditors work, especially when markets are experiencing liquidity or other crises. Thus, the market crisis has provided the accounting and auditing profession with a rare opportunity to take a leadership role in managing the transition of financial statements toward FVA. The future of the profession rests on its success in managing such a transition.

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 211

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

reFerenCes

Aboody, D., M. E. Barth, and R. Kasznik. 2004. Firms’ voluntary recognition of stock-based compensation expense. Journal of Accounting Research 42 (2): 123 – 50.

Ahmed, A., E. Kilic, and G. Lobo. 2006. Does recognition versus disclosure matter? Evidence from value-relevance of banks’ recognized and disclosed derivative financial instruments. The Accounting Review 81 (3): 567 – 88.

Allen, F., and E. Carletti. 2008a. Mark-to-market accounting and liquidity pricing. Journal of Accounting and Economics 45: 358 – 78.

Allen, F., and E. Carletti. 2008b. Should financial institutions mark-to-market? Financial Stability Review 12 (October): 1 – 6.

Barth, M. 2007. Standard-setting measurement issues and the relevance of research. Accounting and Business Research (Special Issue): 7 – 15.

Barth, M. E., W. H. Beaver, and W. R. Landsman. 2001. The relevance of the value relevance literature for accounting standard setting: Another view. Journal of Accounting and Economics 31 (1 – 3): 77 – 104.

Barth, M. E., W. R. Landsman, and J. M. Wahlen. 1995. Fair value accounting: Effects on banks’ earnings volatility, regulatory capital, and value of contractual cash flows. Journal of Banking and Finance 19 (3 – 4): 577 – 605.

Benston, G. J. 2008. The shortcomings of fair-value accounting described in SFAS 157. Journal of Accounting and Public Policy 27 (2): 101 – 14.

Bernard, V., R. Merton, and K. Palepu (1995). Mark-to-market accounting for banks and thrifts: Lessons from the Danish experience. Journal of Accounting Research 33 (1): 1 – 32.

Canadian Institute of Chartered Accountants (CICA). 2009. Handbook. Toronto: CICA. CFA Institute. 2008. Investors overwhelmingly agree: Fair value should not be suspended and the

bailout plan will help the U.S. economy. http://www.cfainstitute.org. September 26. Chambers, R. J. 1964. Measurement and objectivity in accounting. The Accounting Review 39 (2):

264 – 74. Cudahy, R. D., and W. D. Henderson. 2005. From Insull to Enron: Corporate (re)regulation after

the rise and fall of two energy icons. Energy Law Journal 26 (1): 35 – 110. Diewert, W. E. 2005. Issues in the measurement of capital services, depreciation, asset price

changes, and interest rates. In Measuring Capital in the New Economy, eds. C. Corrado, J. Haltiwanger, and D. Sichel, 479 – 542. Chicago: University of Chicago Press.

Financial Accounting Standards Board (FASB). 1993. Statement of Financial Accounting Standards No. 115: Accounting for certain investments in debt and equity securities. Norwalk, CT: FASB.

Financial Accounting Standards Board (FASB). 1995. Statement of Financial Accounting Standards No. 123: Accounting for stock-based compensation. Norwalk, CT: FASB.

Financial Accounting Standards Board (FASB). 2001. Statement of Financial Accounting Standards No. 142: Goodwill and other intangible assets. Norwalk, CT: FASB.

Financial Accounting Standards Board (FASB). 2006a. Statement of Financial Accounting Standards No. 157: Fair value measurements. Norwalk, CT: FASB.

Financial Accounting Standards Board (FASB). 2006b. Financial Accounting Standards No. 159: The fair value option for financial assets and financial liabilities. Norwalk, CT: FASB.

212 ACCouNtiNg PersPeCtiVes / PersPeCtiVes ComPtAbles

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

Financial Accounting Standards Board (FASB). 2008a. Conceptual framework for financial reporting: The objective of financial reporting and qualitative characteristics and constraints of decision-useful financial reporting information. Exposure draft. Norwalk, CT: FASB.

Financial Accounting Standards Board (FASB). 2008b. Statement of Financial Accounting Standards No. 133: Accounting for derivative instruments and hedging activities. Norwalk, CT: FASB.

Flesher, D. L., and T. Flesher. 1986. Ivar Kreuger’s contribution to U.S. financial reporting. The Accounting Review 61 (3): 421 – 34.

Goh, W. B., J. Ng, and K. O. Yong. 2009. Market pricing of banks’ fair value assets reported under SFAS 157 during the 2008 economic crisis. Working paper, Massachusetts Institute of Technology and Singapore Management University.

Gonedes, N., and N. Dopuch. 1974. Capital market equilibrium, information production, and selecting accounting techniques: Theoretical framework and review of empirical work. Journal of Accounting Research 12 (1): 48 – 129.

Holthausen, R. W., and R. L. Watts. 2001. The relevance of the value-relevance literature for financial accounting standard setting. Journal of Accounting and Economics 31 (1 – 3): 3 – 75.

International Accounting Standards Board (IASB). 2008. International Financial Reporting Standards No. 27: Consolidated and separate financial statements. London: IASB.

International Accounting Standards Board (IASB). 2009a. Fair value measurement guidance. Exposure draft. May. London: IASB.

International Accounting Standards Board (IASB). 2009b. Financial instruments: Classification and measurement. Exposure draft. July. London: IASB.

International Accounting Standards Board (IASB) / Financial Accounting Standards Board (FASB). 2009. Conceptual framework — Joint project. London and Norwalk, CT: IASB / FASB.

Jeffrey, G. 2008. Mark market debate down as a draw. The Bottom Line, December, 27. Khan, U. 2009. Does fair value accounting contribute to systemic risk in the banking industry?

Working paper, University of Washington. Kolev, K. 2008. Do investors perceive marking-to-model as marking-as-myth? Early evidence

from FAS 157 disclosure. Working paper, New York University. Landsman, W. R. 2006. Fair value accounting for financial instruments: Some implications for

bank regulation. Bank for International Settlements Paper. Basel, Switzerland. Laux, C., and C. Leuz. 2009. The crisis of fair value accounting: Making sense of the recent

debate. Accounting, Organizations and Society 34 (6 – 7): 826 – 31. Lee, M. C., J. Myers, and B. Swaminathan. 1999. What is the intrinsic value of the Dow? The

Journal of Finance 54 (5): 1693 – 741. Leone, M. 2008. Fair value: It’s the disclosure, stupid. CFO.com. November 20. http://www.cfo.com. Macintosh, N. B., T. Shearer, D. B. Thornton, and M. Welker. 2000. Accounting as simulacrum

and hyperreality: Perspectives on income and capital. Accounting, Organizations and Society 25 (1): 13-60.

Magnan, M. 2006. Les options sur actions : Création de richesse pour les actionnaires ou enrichissement des dirigeants au détriment des actionnaires ? Finance-contrôle-strategie 9 (3): 221 – 35.

Magnan, M., and D. Cormier. 2005. From accounting to “forecounting”. Accounting Perspectives 4 (2): 243 – 57.

FAir VAlue ACCouNtiNg ANd the FiNANCiAl Crisis: messeNger or CoNtributor? 213

AP Vol. 8 No. 3 — PC vol. 8, no 3 (2009)

Martin, R. D., J. S. Rich, and T. J. Wilks. 2006. Auditing fair value measurements: A synthesis of relevant research. Accounting Horizons 20 (3): 287 – 303.

McFarland, J., and J. Partridge. 2008. Mark-to-market accounting rules fuel debate. Globe and Mail. November 20, Report on Business.

Milburn, J. A. 2008. The relationship between fair value, market value, and efficient markets. Accounting Perspectives 7 (4): 293 – 316.

Plantin, G., H. Sapra, and H. S. Shin. 2008. Marking-to-market: Panacea or Pandora’s box? Journal of Accounting Research 46 (2): 435 – 46.

Ryan, S. G. 2008. Accounting in and for the subprime crisis. The Accounting Review 83 (6): 1605 – 38.

Securities and Exchange Commission (SEC). 1999. Staff Accounting Bulletin No. 101: Revenue recognition in financial statements. Washington, DC: SEC.

Securities and Exchange Commission (SEC). 2008. Report and recommendations pursuant to section 133 of the Emergency Economic Stabilization Act of 2008: Study on mark-to-market accounting. Office of the Chief Accountant, Division of Corporation Finance, Washington, DC. http://www.sec.gov/news/studies/2008/marktomarket123008.pdf.

Song, C. J., W. B. Thomas, and H. Yi. 2009. Value relevance of FAS 157 fair value hierarchy information and the impact of corporate governance mechanisms. Working paper, Virginia Tech and University of Oklahoma.

Staubus, G. J. 1961. A theory of accounting to investors. Berkeley: University of California Press. Sterling, R. R. 1970. The theory of the measurement of enterprise income. Lawrence: The

University Press of Kansas. Thornton, D. 2009. Mark-to-audit values. Financial Post (FP Comment), March 26. http://network.

nationalpost.com/np/blogs/fpcomment. Venkatachalam, M. 1996. Value-relevance of banks’ derivatives disclosures. Journal of Accounting

and Economics 22 (1 – 3): 327 – 55. Wahlen, J., J. R. Boatsman, R. H. Herz, G. J. Jonas, K. G. Palepu, S. G. Ryan, K. Schipper, C. M.

Schrand, and D.J. Skinner. 2000. Response to the FASB preliminary views: Reporting financial instruments and certain related assets and liabilities at fair value. Accounting Horizons 14 (4): 501 – 8.

Watts, R. 2003. Conservatism in accounting part I: Explanations and implications. Accounting Horizons 17 (3): 207 – 21.

Watts, R. and K. Ramanna. 2007. Evidence on the effects of unverifiable fair-value accounting. Working paper 08-014, Harvard Business School (http://ssrn.com/abstract 1012139).

Weil, R. 2001. After Enron, “mark to market” accounting gets scrutiny. Wall Street Journal December 4, C1.