The Decision Usefulness of Fair Value Accounting – A Theoretical Perspective

ABSTRACT

Regulators such as the SEC and standard setting bodies such as the FASB and the IASB argue the case for the conceptual desirability of fair value measurement, notably on the relevance dimension. Recent standards on financial instruments and certain non-financial items adopt the new measurement paradigm. This paper takes issue with the notion of decision usefulness of a fair-value-based reporting system from a theoretical perspective. Emphasis is put on the evaluation of the theoretical soundness of the arguments put forward by regulators and standard setting bodies. The analysis is conducted as economic (a priori) analysis. Two approaches to decision usefulness are adopted, the measurement or valuation perspective and the information perspective. Findings indicate that the decision relevance of fair value measurement can be justified from both perspectives, yet the conceptual case is not strong. The information aggregation notion that underlies standard setters' endorsement of fair value measurement turns out to be theoretically restricted in its validity and applicability. Also, comparative analysis of fair value accounting vs. historical cost accounting yields mixed results. One immediate implication of the research – a condition for the further implementation of fair value accounting – is the need to clarify standard setters' notion of accounting income, its presumed contribution to decision relevance and its disaggregation.

Acknowledgements

I am indebted to Christoph Kuhner for his critical and insightful comments at various stages of this project. Also, the helpful suggestions of two anonymous reviewers and the journal's previous editor, Kari Lukka are gratefully acknowledged. Finally, participation in the European Accounting Association's 2002 doctoral colloquium provided useful comments at a very early stage of this project, and lent valuable stimulus to my research.

Notes

¹See, for example, IAS 39, para. 9; IAS 41, para. 8; IFRS 3, Appendix A; IFRS 4, Appendix A. The broad correspondence of the two standard setters' concepts is demonstrated by the IASB's reaffirmation to adopt SFAS 157 as a basis for the project on Fair Value Measurements.

²In December 2006, the IASB issued a Discussion Paper which states broad agreement with SFAS 157. Thus, final convergence concerning the definition and estimation of fair value is on the horizon. However, since some IAS/IFRS standards appear to employ an entry value notion of fair value, the IASB considers scope limitations for an IFRS on fair value measurements, or, alternatively, introduction of an entry price measurement attribute for those standards. See IASB (2006, paras. 12–17). For an analysis of the current IAS/IFRS fair value guidance, see Cairns (2006, pp. 7–10).

³See the definition in SFAS 157, para. 24. Under IFRS, a uniform definition of an active market is applied on the standards level; see IAS 36, para. 6; IAS 38, para. 8; IAS 41, para. 8.

⁴ Thomas S. Kuhn in his ‘Structure of Scientific Revolutions’ extensively discusses this term in the context of scientific methodology and develops his influential theory of the process and drivers of paradigm shifts. While Kuhn's ideas have been subject to contentious debate, with himself backtracking on the definition, the general idea of a paradigm as a set of shared beliefs appears useful for tentatively conceptualizing the theoretical basis of fair value measurement.

⁵The implementation of the decision usefulness objective drew from the Trueblood Report. However, the foundations had been developed much earlier both in the academic and the standard setting sphere. Significant contributions involve the AAA's ASOBAT and APB Statement No. 4. For an historical overview see Hendriksen and van Breda (1991, pp. 92–115, 126–131).

⁶See also SFAS 115, para. 40; SFAS 107, para. 39; SFAS 133, para. 220; and SFAS 157, para. C32. For similar conjectures in IASB pronouncements, see, for example, IAS 36, para. BCZ11; IAS 40, para. 40, B36; IASC (1997, ch. 5, paras. 2.6–2.12); and, more recently and broadly in the context of initial recognition, IASB (2005, paras. 99, 101–104).

⁷See SFAC No. 7, especially the section on ‘Present Value and Fair Value’ (paras. 25–38), and the observations on the evolution and implementation of the fair value paradigm in the following Section 2.3.

⁸For an even earlier exposition of a similar approach, see Schmalenbach (1919) on the concept of a ‘dynamic balance sheet’.

⁹In an influential paper, Sprouse (1978) termed those balance sheet positions that are mere by-products of the matching process and do not conform to notions of assets or liabilities ‘what-you-may-call-its’ (p. 69). He observes: ‘Under the revenue/expense view, (…) what constitutes “proper matching” and “nondistortion” is very much in the eyes of the beholder (…) “Matching costs and revenues” is too often an attractive but empty slogan rather than a meaningful concept that one can look to for guidance.’ See also Gellein (1992, p. 198) and Schuetze (2001, pp. 9–11).

¹⁰The perception thus was that under a revenue–expense approach, the balance sheet merely served as ‘mausoleum for the unwanted costs that the double-entry system throws up as regrettable by-products’ (Baxter, 1977, p. x).

¹¹See Robinson (1991, p. 110). For the respective definitions, see SFAC No. 6, paras. 25–43, 70, 78–89; IASB Framework, paras. 53–64, 70 (Nobes, 2001, p. 9). The validity and necessity of the asset-liability approach as ‘conceptual anchor’ has only recently been reaffirmed by the SEC (SEC, 2003, p. 10).

¹²Former FASB member Arthur Wyatt refers to it as ‘possibly the most significant initiative in accounting principles development in over 50 years’ (Wyatt, 1991, p. 80), a notion emphasized by the testimony of SEC General Counsel James Doty to the US Senate, who made it clear that ‘the time has run out on “once-upon-a-time-accounting”‘.

¹³A similar reasoning can be found in the IASB's 2005 Discussion Paper on Measurement Bases, which emphasizes that the relevance of fair value as a measurement attribute for initial recognition stems from its ‘market value properties’, which are assumed to hold in principle for any fair value, that is, prevail irrespective of its estimation. See IASB (2005, paras. 111, 227–231).

¹⁴For instance, market valuation of long-lived non-financial assets was well established in the USA prior to the Great Crash and the ensuing security market regulation in the 1930s (Walker, 1992, pp. 4–8). Although the SEC later prohibited such practice, the issue of market value accounting for certain items was discussed regularly in the USA as well as in other countries (Wyatt, 1991; Christie, 1992, p. 97).

¹⁵For an overview see Nelson (1971). The debate can be traced back as far as to Canning (1929) or Simon (1899). Later, it focused on the scope of current value measurement and the selection of the appropriate measurement basis. Given the diversity of this literature, only some of its more famous contributors are mentioned: for example, Chambers (1965) as an advocate of exit value, Revsine (1970) and Sterling (1981) as advocates of entry value, Edwards and Bell (1964) and their concept of business profit.

¹⁶While Beattie (2002, p. 109), Demski et al. (2002, p. 163) and Scott (2003, p. 174) perceive the move towards fair value measurement as a renaissance of the measurement perspective, Beaver (1998, p. 4), for instance, frames this development in an information perspective context.

¹⁷A thorough textbook description of the information economics approach to financial reporting is given by Christensen and Demski (2003).

¹⁸Thus, we do not subscribe to the implications of the impossibility result (Demski, 1973): the denial of the usefulness of applying conceptual criteria on the grounds of the specifity of individual decision contexts employs a Paretian notion of economic efficiency which is not adequate for economic analysis. Rather, economic analysis requires an assessment of the welfare implications of different alternatives which can only be properly evaluated on the basis of Kaldor–Hicks efficiency. While Pareto efficiency prescribes that no user will suffer a reduction in welfare from a new accounting regulation and at least one user will experience an increase in personal welfare, Kaldor–Hicks efficiency is less restrictive in that it requires an increase in aggregate welfare, that is, it implies that those users better off can compensate those users that suffer decreases in welfare. See Posner (1998, p. 16); in the context of normative evaluation of accounting alternatives Beaver and Demski (1974, p. 174), Cushing (1977, p. 311) and Ohlson (1987, pp. 2, 6).

¹⁹It is not the aim of this paper to give a thorough discussion on the merits of normative vs. positive research approaches. Based on the arguments given, it is simply assumed that there is a case for a priori reasoning, especially with respect to standard setting questions. For a more general discussion, see Cushing (1977), Christenson (1983), and Watts and Zimmerman (1990). For an assessment of potential standard setting implications of empirical research, especially value-relevance research, see Holthausen and Watts (2001).

²⁰Curiously, the active market criterion outlined in IFRS standards as a means for discriminating sufficiently informative market prices (level one estimates, see note 2) exclusively refers to the time dimension of liquidity, that is, the speed with which a transaction partner can be found, rather than the price dimension as theoretically valid indicator of information quality, which can be measured, for example, on the basis of bid–ask spreads.

²¹The latter also applies to a traditional mixed model, where assets carried at historical cost are mingled with assets valued at a lower current value. However, as pointed out earlier, a stand-alone analysis of the informational properties of historical cost-based measures and thus, balance sheet captions, appears misguided since, unlike fair values, there is no explicit claim to approximating economic value and aggregating cash flow information.

²²The (re-)emergence of the EBO model can therefore be seen as a ‘renaissance’ of the measurement perspective (Beattie, 2002, p. 109).

²³For the distinction between economic income in a narrow sense and this ‘economic profit’ see Christensen and Demski (2003, pp. 40, 50).

²⁴Although, as pointed out, the frameworks elaborate no concept of informative earnings, several paragraphs hint at the notion of persistent earnings. See, for example, SFAC 1, para. 44; SFAC 5, para. 31; IASB Framework, paras. 28, 72–80.

²⁵These results are turned upside down for residual income. Fair value residual income is constant in time and reflects the competitive advantage, that is, the fraction of sales that is earned on top of market expectations. Transaction-based residual income, on the other hand, increases in time and thus appears less indicative of the competitive advantage.