Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

Over the past two decades, through the impetus of Anglo-Saxon standard setters, the fulcrum of the traditional accounting model, namely the accounting standards under which a company’s equity and income are measured and reported, have led to much debate. This wide-reaching movement, originally initiated in the United States by the Financial Accounting Standards Board (FASB), and spread at an international level by the International Accounting Standards Board (IASB), aims to replace historical cost—the basis for the measurement of a company’s income statement and valuation of its assets and liabilities under most continental European accounting standards—with the notion of fair value. Fair value is currently defined by the IASBFootnote 1 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.” The concept of fair value is broader and can be applied more generally than that of market valueFootnote 2: Indeed, where there is no quoted market price available on an active market, the valuation is determined by the exchange value agreed by two independent parties, by the market price of an item with similar characteristics or by the net present value estimate of future cash flows. Fair value can potentially be used for a large number of non-financial assets and liabilities (e.g. goodwill impairment, post-retirement scheme valuation) and can therefore serve as the basis for a new corporate accounting model which aims to provide a more accurate reflection of the uncertainties affecting future cash flow estimates and investment opportunities in financial reports.

This article discusses the usefulness of financial information based on the measurement of a company’s wealth and net income under the fair value approach. In this respect, the key question—is fair value relevant?—can be analysed as follows: (1) Do fair value-based “accounting numbers” help better estimate the value of a company and the intrinsic risk relating to its activity? (2) How informative are they for financial statements’ users? (3) How useful is fair value information for decision-making?

In order to address these concerns, we shall first examine the basis and boundaries of the traditional accounting historical cost model. We will then analyse the key determinants of the emergence of the fair value approach, with particular regard to the role attributed to financial statements and the needs of their users. Lastly, we will present a summary of the empirical studies carried out to assess the usefulness of fair value accounting information and will draw out future research avenues.

1 Traditional Accounting Model: Basis, Boundaries and Alternatives

1.1 Corporate Accounting (Re)presentation: A Value-Oriented Framework

The accounting (re)presentation of a company is a contingent structure based on generally accepted principles in a given historical and economic context, which together form a model. The legitimacy of this model is based on its capacity to apprehend, evaluate, summarise and monitor over time information relating to transactions that have an impact on the company’s wealth and performance. Ideally, this model should enable the definition of a system of measurement for company net income and shareholders’ equity that is socially acceptable and also meets the requirements of the wide panel of users of financial statements (employees, managers, suppliers, customers, fund providers, tax authorities etc.).

The standard corporate accounting model is based on an underlying concept of valueFootnote 3 which, as in economics, can refer to cost, exchange value or value-in-use. The approach selected will determine the structure of the accounting presentation and the criteria used to measure a company’s performance—net income—and wealth—net assets.

As shown by Richard (2001), “while valuation at cost price seemed to take precedence in the earlier European regulations, from 1673 to around 1800, realizable value was recommended by jurists and became widely used in accounting regulations in Germany and France during the nineteenth century and even beyond.” The current debate surrounding fair value therefore echoes the debate that took place in France and Germany up until the Second World War, between advocates of dynamic accounting (referring to cost) and supporters of static accounting (referring to realizable value in the context of going concern). However, in the United States, banks started to drop references to market price (in favour of historical cost) in 1938, following the Great Depression (Swenson and Buttross 1993). After the Second World War, the historical cost model gained precedence both in Europe and the United States.

1.2 The Mainstream Accounting Model: Value as a Reference to Cost

Today’s mainstream accounting model for preparing statutory and consolidated accounts is based primarily on the concept of cost value (i.e. previously accumulated costs), associated with the realisation principle and, depending on national practices, with the prudence principle.Footnote 4 Variants of this model differ in the extent to which they recognise the principle of placing (economic) reality over (legal) appearance.

Generally speaking, this model functions like an asymmetrical filter favouring the recognition of potential losses and carrying over gains from the effective completion of a transaction (Basu 1997). It is based on a prudent and non-volatile measurement of net income and equity: since the 1990s, this view has been the object of much criticism, particularly with regard to the actual relevance of the accounting information reported.Footnote 5 Conversely, its proponentsFootnote 6 attribute a number of qualities to this method, which they feel justify its continuing use and prevalence. For instance, Ijiri (1971, 1975) supports that historical cost is more objective and reliable than alternative methods which are supposedly more relevant, and it is therefore a more effective means of resolving conflicts of interest. In fact, except in periods of inflation, it is a particularly robust model that is highly appreciated by the non-banking business world.Footnote 7 The FASB (1984) and the initial International Accounting Standards Committee (IASC)Footnote 8 (1989) conceptual frameworks present it as the most frequently used valuation basis.

1.3 Historical Cost Accounting: What Information Do “Accounting Numbers” Provide?

The first empirical studies to assess the usefulness of accounting data for the decision-making process were carried out by Ball and Brown (1968) and Beaver (1968), then taken up by various other researchers in the 1970s and 1980s. This work drew on efficient markets theory and event study methodology, with the aim of highlighting the market’s reaction to published accounting information (annual or interim financial statements) in the form of abnormal returns. All of these empirical studies found that the only information content taken into account by the markets is net income. Paradoxically, this work highlighted the extent to which markets anticipate the informational content of accounting data well in advance of its publication in financial statements. The results of these studies thus undermined the assumed usefulness of “accounting numbers” and their relevance in decision-making, leading researchers to suggest they have more than a mere contractual utility (Watts and Zimmerman 1986), and to give precedence to the disciplinary role of accounting, i.e. the accountability.

1.4 Alternative Valuation Methods: A Lot of Proposals But No Obvious Practical Benefits

Researchers put forward a number of alternative valuation conventions in order to restore some relevance and utility to the traditional accounting model chiefly based on the historical cost approach. These accounting models can first be classified on the basis of two types of independent criteria (Boussard 1997), the first relating to the choice of method for the valuation of assets (historical cost versus concept of “value”), the second relating to the unit of monetary measurement (nominal value versus purchasing power). Where there is no inflation, this typology can be further refined by combining the “valuation method” criterion with respect (or relaxation) of the realisation principle, thus highlighting the essential dichotomy in the way results are established (transaction result versus holding result).

The principal models put forward in accounting literature can be separated into three categories, according to the notion of value used as a reference:

  • The “current entry values” model, based on the entry value which refers either to the acquisition price or the replacement cost. This model was advocated by Edwards and Bell (1961) in relation to the maintenance of physical capital.

  • The “current exit values” model, based on the exit value, which is the price at which an asset can be sold or liquidated. This was put forward by Chambers (1966) and by Sterling (1970).

  • The “value-in-use” model, based on the added-value of the asset to the company, and measured by the present value of future cash flows estimate.Footnote 9

In 1984, when it adopted its current conceptual framework, the FASB recognized the following as a basis for valuationFootnote 10: (1) historical cost, (2) current value or replacement value, (3) market value and net market value (excluding any case of forced liquidation) and (4) present value estimates of future cash flows. Method (2) complies with the transaction principle, while models (3) and (4) do not apply this principle strictly.

2 The Emergence of the Fair Value Concept: The Issues at Stake and Leading Factors

Representing a genuine change in direction, fair value emerged as a result of four combined factors to become a central accounting convention under US and international accounting standards.

The first leading factor in this change was the attitude adopted in the new Anglo-Saxon conceptual frameworks—that of the FASB (1984) and then the IASC (1989)—with regard to issues such as the purpose of accounting (assistance with decision-making versus rendering of accounts and controls), the various meanings assigned to the notion of “users” of financial statements (the investor in the generic sense versus many types of users) and, implicitly, certain qualities expected from financial and accounting informationFootnote 11 (relevance versus accuracy). The objectives attributed to financial statements were clearly biased towards meeting the needs of users—mainly creditors and investors—in terms of forecast data, placing particular emphasis on the usefulness of accounting information for external parties in their economic decision-making. Many researchers support the idea that board members with a background in financial services have been taking over the Boards of the FASB and the IASB and are thus able to make the accounting standards purport that fair value is the correct measurement basis to run a company as it is in the financial industry (Allen and Ramanna 2013).

The second key determinant is the growing use of increasingly complex financial instruments and the high level of market volatility. The widespread use of these instruments has increased levels of risk and aptly highlights the issue of the relevance of accounting information and notably the recognition of financial instruments in financial statements. Derivative products demonstrate this problem particularly well as they initially mobilise modest amounts of capital but generate a significant risk at a later stage. The financial institution bankruptcies observed in the United States at the end of the 1980s underlined the boundaries of the standard accounting model: the historical cost model was unable to provide a timely indicator to the users of financial information of the financial health of banks using derivatives (Barth et al. 1995). For some, the need to prevent systemic crises in the financial sector and to increase the relevance of financial information were behind the emergence of a range of proposals for a fair value method (Cornett et al. 1996).

The third determining factor is the will, of the Securities and Exchange Commission (SEC) in particular, to reduce the discretionary power of directors over earnings management (Levitt 1998; Lewis 2012). The historical cost model gave directors some considerable leeway both to make provisions, and hence integrate uncertainty, and to establish ad hoc results. As provisions are based on a subjective assessment of risk, cost forecasts and non-definitive amortisation of assets, they can be used as a political accounting tool. Conversely, directors can take an opportunistic approach to the concept of transaction results and dispose of assets that conceal unrealised capital gains (while deferring the disposal of assets carrying potential capital losses) in order to generate income, delay the appearance of losses or smooth out results. Consequently, the use of fair value is presented, notably by the US market regulator, as a solution for a more reliable valuation of assets and shareholders’ equity (Beatty et al. 1996).

The last leading factor is the introduction of finance theory into accounting research since the early 1980s. According to some authors (e.g. Ramanna 2013; Holthausen and Watts 2001), this may have caused to widespread the study of the association of accounting data with market data and gave intellectual rationales to the proponents of fair value.

As a result of the dysfunctions caused by the irregular use of the historical cost model—which has not fulfilled its role as a safeguard—and in order to improve the relevance of information disclosed on certain financial instruments, the standards authorities implemented a programme to break away from the historical cost model—most often used according to the company’s intended purpose for holding the assets—and to promote the notion of fair value. The FASB, supported by the SEC and part of the academic worldFootnote 12 and, on an international level, by the IASC, was decisively involved in this process. However, as noted by Garmilis (2001), “the adoption of standards based on fair value has always caused some conflict, between regulators and standards authorities on the one hand, and companies on the other”. In fact, since information requirements have always been badly defined and are often assumed, the promotion of the fair value model is motivated more by the large number of proposals put forward than by any explicit request on the part of users.Footnote 13

3 Fair Value Accounting: Evolution or Revolution?

Having briefly looked at the stages in the accounting standardisation process that led to the emergence of a fair value model, we now discuss the qualities and disadvantages attributed to this measurement model by academics and professionals.

3.1 The Introduction of Fair Value into Accounting Standards

Although the concept of fair value first came into use in the 1950s, it had a limited scope of application and very specific accepted meanings.Footnote 14 Actually, the term has only been used in its current form since the 1990s. Over this period, fair value began to be used for the disclosure (information on financial instruments in the notes to financial statements), and later for the accounting recognition—i.e. the booking of unrealised profits and losses under earnings—of certain financial instruments and derivatives. It is now becoming more widespread and is being used in a variety of fields, not only as a principle for initial valuation, but also for monitoring the value of assets and liabilities.

Developments in accounting standards in the United States illustrate the increasing popularity of the notion of ‘fair value’,Footnote 15 driven by both the SEC and the FASB, and linked to the growing use of financial instruments. In 1991, SFAS standard 107 (1991)Footnote 16 referred to fair value as the valuation basis for disclosure. SFAS 119Footnote 17 (1994) similarly referred to fair value as the basis for the disclosure of information on derivatives. With SFAS 115,Footnote 18 which institutionalised accounting on the basis of the management perspective in 1993, the notion of fair value became the reference method for reporting certain securities (investment and transaction securities) in the balance sheet and earnings were measured according to the variation in the fair value of these securities. Since then, this valuation method, which takes account of unrealised gains and losses, has underpinned the FASB’s strategy to eliminate or limit the effects of discretionary choices in accounting based on intent. Adopted in 1998, despite considerable opposition from banks, SFAS 133 (1998)Footnote 19 made fair value accounting of derivative financial instruments—assets and liabilities—in the balance sheet mandatory,Footnote 20 both on their initial entry into the accounts and subsequently. SFAS Standards 141 and 142 (2001) extend beyond the scope of financial instruments, stipulating that, in business combinations, all identifiable assets and liabilities should be reported at fair value, and these valuations monitored over time. Similarly, the replacement of the amortisation of certain intangible assets by impairment tests is also based on fair value monitoring.

International accounting standards also evolved during this period. IAS 32 (1995), relating to the disclosure of information on financial instruments, adopted fair value as a valuation basis. However, it was only after 1998, subsequent to a complete review of its standards, that the IASC introduced the notion of fair value into all of its standards, including those relating to business combinations, post-employment retirement scheme, intangible assets, securities portfolios, fixed asset revaluations, etc. Also in 1998, the IASC adopted IAS 39 which, like SFAS 133, generalized the use of fair value for the reporting of numerous financial instruments. Highly criticized by credit establishments, the original version of this accounting standard raised a number of problems with regard to its implementation (enhanced volatility of firm performance indicators, pro-cyclical accounting treatment, difficulties encountered during the 2008 financial crisis to identify reliable market data…) although the IASC through this standard aimed to improve the monitoring of risk exposure—through a periodical revision of market values or discounted present values—and to monitor the creation of value more effectively, regardless of the intentions of the parties. However, once again, strong opposition from the professional world during the 2008 subprime crisis deeply undermined this project oriented towards a more ‘full fair value’ approach.

3.2 Qualities Attributed to the Fair Value Model

A number of benefits have been cited to justify the use of fair value in the accounting for financial instruments.Footnote 21

The fair value method is based on the discounting of future financial flows and thus provides information which integrates market trends (Allen and Carletti 2008b). It is therefore perfectly in line with current methods used by investors to compute their cash flow estimates.

Furthermore, the fair value method provides better comparability of financial statements, by giving equivalent values for a financial instrument, regardless of the date on which it was acquired, thus eliminating opportunities for “cherry picking” arising from the improper application of the realisation principle. It also ensures that the measurement of performance is exhaustive: by integrating not only transaction gains and losses but also holding gains and losses, the fair value model provides an entirely faithful representation of the strategy adopted for financial instruments—disposal versus holding. Moreover, it guarantees that the value accounted for is exhaustive, particularly for derivatives where the initial cost is zero.

Ensuring that this approach is consistent with the approach used for operational risk management (interest rate, currency or price) would facilitate the reconciliation of accounting income with economic income. Moreover, the use of fair value would ensure that the information produced is neutral, as it is based on data which is exogenous (market values or, if no active market is involved, model values based on external parameters) and easily accessible (market values).

3.3 Criticisms of the Fair Value Model

Several criticismsFootnote 22 have been made against the use of fair value.

Many of these arguments refer to the increased volatility of fair value accounting measuresFootnote 23 and the consequences thereof. However, in doing so, they in fact raise the more fundamental issue of the actual function of accounting models and the relevance of filtering or, on the contrary, better reflecting the actual volatility of economic activity. Conversely, other criticisms underline the unjustified increase in the volatility of earnings and shareholders’ equity as a result of an implicit abandonment of the going concern principle.

The most frequently cited criticism concerns assets not traded on an efficient market,Footnote 24 which are consequently valued on the basis of internal models. Detractors stress the lack of objectivity and neutrality in these valuations, and the loss of reliability and comparability due to the use of internal models (Allen and Carletti 2008a).

Other critics argue that the use of the fair value accounting model implies taking a short-term approach to the management of a company.

Lastly, certain detractors of this model stress the prohibitive cost of obtaining information, given the limited usefulness of fair value data for users.

4 Fair Value and Information Usefulness: Empirical Evidence

Despite the increasing popularity of this accounting model, the use of fair value as a general valuation principle poses a number of practical problems and raises major issues. Although the benefits attributed to fair value methods are generally the result of deductive reasoning, assumptions and even affirmations, the criticism often stems from fears rather than actual inadequacies that have been pinpointed in an empirical manner.

The introduction of valuation in financial reporting raises the question of the legitimacy of accounting methods and highlights the need for a theoretical analysis framework. The latter generally refers to the efficiency of capital markets as a working theory.Footnote 25 The most widely used methodology involves examining the effect of an accounting choice—such as fair value—on the market value of a sample of companies, where the price is perceived as an aggregate measure of future cash flow estimates. The theory of informational usefulness associated with an accounting choice is valid if it is possible to establish a significant link between the accounting choice and the change in the price or Market to Book Ratio.

Because a certain amount of time is required before the application of standards can be assessed, the available empirical studies mainly concern either the banking sector and, more often, the effects of the introduction of SFAS 107 or 115 or the goodwill impairment value-relevance over the post-SFAS 142 regime. However, it is already possible to draw certain conclusions in relation to our objectives from this work.

With regard to the information content of “accounting numbers” for the market, empirical studies are not generally able to show that the fair value model is significantly better than the historical cost model.Footnote 26 However, some of the work carried out appears to establish a link between fair value and the market price of a company. For example, in discussing the problem of valuing a securities portfolios (Eccher et al. 1996), establish that there is a strong correlation between the fair value of securities and the market value of a company. However, this result cannot be applied to all balance sheet items, such that the fair value of financial instruments, taken on its own, can only account for a small part of the variation in Market to Book Ratio.

With regard to the effect on the market of adopting different accounting regulations (Cornett et al. 1996), show that the events—that is, the different phases in the standardisation process—relating to the introduction of fair value have a negative impact on the value of banks.

Moreover, concerning the problem of volatility, if the earnings of banks are significantly more volatile using fair value than on the basis of historical cost, this amplification does not appear to have any significant effect on stock market returns (Barth et al. 1995). In fact, behind this finding lies a fundamental question concerning the expected properties of the accounting model. Should it be founded on measurements that filter risk—that is, constructed to reduce systemic entropy—or, on the contrary, should it be as neutral as possible in order to pass on information on risks to the users of financial statements?

Lastly, with respect to the usefulness of fair value information according to the method by which it is integrated into financial reporting—that is, whether it is used in accounting recognition or disclosure—(Beatty et al. 1996) show, by studying the reaction of stock market prices to the adoption of SFAS standard 115, that the accounting of unrealised gains and losses is useful solely for regulatory bodies and not for other users. In parallel, researches analysing the impact of fair-value based goodwill impairment over the post-SFAS 142 implementation period find that the association between goodwill and future cash flows is stronger (Lee 2011; Li et al. 2011).

Although these empirical studies are still only partial, the results obtained so far have rekindled the debate regarding the usefulness of accounting information based on fair value. As regards its usefulness for decision-making, one interpretation favours the contractual role of “accounting numbers”. As Jeanjean (2001) shows, the introduction of the fair value model underlines the highly “disciplinary” role of this method. In fact, as the completion of transactions is no longer determinant in the calculation of income, it is possible to control the activities of managers more effectively and eliminate the opportunistic management of holding gains. Moreover, this model would simultaneously introduce an accounting framework that centred management decisions on the creation of value, and a method of financial reporting that complies with current standards for performance measurement based on shareholder value.

5 Conclusion

The problems surrounding the introduction of fair value into accounting models goes beyond the scope of financial instruments or even the banking sector. Indeed, through the standards relating to business combinations and the depreciation of intangible items, it potentially concerns all companies of a certain size. Moreover, this change in accounting concepts forms the very core of the currently discussed conceptual framework to be implemented by the IASB for the establishment of future IFRS standards. The emergence of this new accounting model will require not only an adjustment of financial disclosures and practices, but also a redefinition of the respective roles of the balance sheet/financial position and the income statement which is currently under debate at the IASB level.Footnote 27

In an effort to re-establish the relevance of accounting information, standard setters are identifying consistent links between “accounting numbers”, management indicators and company value. The debate which has arisen over the introduction of fair value has thus shed light on the question of the purpose of financial statements and the usefulness of accounting information. The advantage of this is that it has extended the scope of discussions beyond the purely technical consideration of whether historical cost is a better valuation method than fair value, to look at the effects on the allocation of resources and the underlying economic issues (For whom and for which decisions should this information be produced?).