How Decision useful can it be? The case of fair value
accounting
Bachelor Thesis Joe Petrus Kok 10208585 30-6-2014
BSc Accountancy & Control University of Amstedam Supervisor: Mr. C. Clune MSc. Final version
Abstract
This thesis provides an analysis of the literature regarding the decision usefulness of fair value accounting and asks the question to what degree fair value accounting (FVA) provides financial information that is decision useful. To answer this question I divide decision usefulness in six different elements based on the model that the FASB and IASB introduced in the statement of concepts No. 8 (SFAC No.8). The overall findings suggest that FVA produces decision useful financial information according to the definition of the IASB and FASB. The reason of this result is the emphasis of the IASB and FASB on relevant financial information instead of reliable financial information. I disagree with this finding because I think that the literature implies that FVA information can never be reliable, because of its subjective nature and emphasis on estimates. Therefore I introduce a revised model with the emphasis on reliability and come to the conclusion that FVA is not decision useful when the financial information must be reliable instead of relevant. This summarizes the answer on my question which is that FVA is decision useful to the degree that it produces financial information that is relevant. This thesis also provides some ideas about future research in the different elements.
Samenvatting
Deze scriptie geeft een overzicht van de literatuur met betrekking tot de ‘decision usefulness’ van fair value accounting (FVA). De scriptie probeert de volgende vraag te beantwoorden: in hoeverre produceert fair value accounting financiële informatie die decision useful is. Om een antwoord te geven op deze vraag is het begrip decision usefulness verdeeld in zes verschillende elementen die komen uit het model dat de FASB en IASB hebben geïntroduceerd in hun statement of concepts No. 8 (SFAC No.8). Dit onderzoek vindt dat FVA financiële informatie produceert die decision useful is wanneer de definitie van de IASB en de FASB wordt aangehouden. Het belangrijkste kenmerk van deze definitie is dat de IASB en de FASB relevante informatie belangrijker vinden dan betrouwbare informatie. Ik ben het oneens met deze bevindingen omdat ik geloof dat de literatuur impliceert dat FVA informatie nooit betrouwbaar kan zijn vanwege zijn subjectieve karakter en het belang van schattingen. Dit is de reden dat ik een model introduceer waarin de nadruk ligt op betrouwbaarheid. Hiermee is het antwoord op mijn vraag duidelijk. FVA produceert informatie die relevant is, maar niet geheel betrouwbaar. Daarnaast geef ik een aantal aanbevelingen voor vervolg onderzoek in gebieden waarin de relatie tussen FVA en een element nog niet geheel is onderzocht.
Table of Content
List of tables and illustrations ... 4
1. Introduction ... 5
2 Fair value accounting and decision usefulness ... 8
2.1 What is Fair value accounting? ... 8
2.2 Decision usefulness ... 9
2.2.1 Relevance ... 11
2.2.2 Faithful representation ... 11
2.2.3 Enhancing Qualitative Characteristics ... 11
3. Decision Usefulness of fair value accounting ... 12
3.1 Relevance ... 12
3.1.1 Predictive value ... 12
3.1.2 Materiality ... 14
3.2 Representation faithfulness ... 15
3.2.1 Free from error ... 15
3.2.2 Neutral ... 18
3.3 Enhancing Qualitative Characteristics ... 20
3.3.1 Comparability ... 21
3.3.2 Verifiability ... 22
4 Discussion ... 24
5 Conclusion ... 29
List of Tables and illustrations
Figure 2-1: The model of decision usefulness ... 10
Figure 4-1: The revised model of decision usefulness ... 27
Table 4-1: The relevance of FVA... 25
Table 4-2: The representation faithfulness of FVA ... 25
1. Introduction
In the long history of the profession of accounting there has always been discussion around the topic of what the numbers should represent. Should they represent the costs they were acquired for? Should they represent the value when they are in use or should they represent a value that they are worth: a fair value? The International Accounting Standards Board (IASB) and the Financial Accounting Standard (FASB) state that the numbers should represent a fair value. They see fair value as the best alternative, since fair value accounting (FVA) leads to better decisions by investors, provides more relevant information and gives a better view of a firms current position (Penman, 2007). The decision by the IASB and FASB to make FVA the new way of accounting has started a debate regarding this claim. Can FVA be the new way of accounting? Is historical cost accounting better? Are there no alternatives (Barth & Landsman, 1995)? When the financial crisis emerged in 2007 FVA was blamed as one of the causes of this crisis (Véron, 2008). Even though people believe that this was mere scapegoating and FVA had no role at all (Véron 2008). The reasons for blaming FVA were not ungrounded. FVA is known for producing unreliable numbers and giving firms a higher systematic risk (Penman, 2007).
In the years of developing a FVA framework the standard setters have found a definition. Because of the drift for globalisation the IASB and FASB created their definition of fair value in the Statement of Financial Accounting Standards (FAS) No.157 and IFRS No. 13 (FASB, 2006; IASB, 2011). This definition has been criticised for its trade-off between relevance and reliability and its introduction of the fair value hierarchy that gives a lot of space for managerial estimates and increasing bias of fair value numbers (Benston, 2008).
To test the worth of FVA researchers mostly use decision usefulness or value relevance. In a combined conceptual statement the IASB and FASB introduced decision usefulness as the most important objective of accounting (FASB, 2010). Decision usefulness means that financial information has the power to make a difference in the decisions made by stakeholders. (FASB, 2010). The introduction of decision usefulness was in the early years of the FASB and came to its first definition in Concept Statement No. 1 (SFAC No. 1), but came to its most recent and internationally accepted definition in Concept Statement No. 8 (SFAC No.8) (FASB, 2010). This conceptual statement introduces a model for decision usefulness that is different from the one introduced in Concept Statement No. 2 (SFAC No. 2). The focus is more on relevance instead of reliability, which connects the most recent fair value definition with decision usefulness.
With all the interest of the IASB and FASB in decision usefulness and their plan to integrate FVA into all their future standards it is interesting to look at the connection between the two. Various studies have contributed to the value relevance of FVA literature and have tried to find if FVA is more relevant than historical cost accounting or if FVA could be value relevant on its own (Landsman, 2007). Through a literary review my study tries to answer the following question: to what degree does FVA provides financial information that is considered decision useful by the IASB and FASB?
This thesis finds that FVA does provide decision usefulness as found in many previous studies, but that this decision usefulness has its limits. Based on arguments given by the literature I argue that FVA only provides relevant but no reliable financial information. The reason that FVA can still be called decision useful is because of the FASB and IASB’s choice to call reliable information sufficient enough to produce decision useful financial information. According to the FASB and IASB financial information has to be relevant at all costs and the reliability of this information has only a complementing role. This study confirms this idea and concludes that FVA provides decision useful information to the degree of its relevance but not its reliability.
The contribution of this thesis is that it defines decision usefulness in a different way. Normally decision usefulness is seen as a single object, that does not consist of different elements, which is a completely false idea. SFAC No. 8 introduces a model for decision usefulness, introducing different elements of decision usefulness. This thesis shows that every element has it effects on FVA and that every element gives a different conclusion about FVA. With splitting up decision usefulness this study has another merit to other studies. The thesis introduces some new fields of research that can be studied for further research. It shows that some problems and contributions of FVA are far from having been researched.
To answer the question stated above I will introduce fair value accounting and decision usefulness in the next chapter. In part three the main elements of the model determine the structure. Various elements of the model are discussed with respect to their relevant literature. The elements that are not discussed, are not discussed because of their lack of relevant literature. For every element discussed, I will state the effects that the results have on the information produced by FVA. The fourth part of my study will address some general findings in all the different literature, general remarks by researchers et cetera. Next I will introduce a revised model of decision usefulness. This revised model is used to show that when decision usefulness is centred around reliability, instead of relevance, fair value accounting would fail in producing the correct information, showing the boundaries of how
decision useful fair value accounting is. I conclude with some remarks on this project and will introduce some research fields that could receive more attention.
2 Fair value accounting and decision usefulness
This section will first explain what fair value accounting is. The definition of FVA is explained as well as the fair value hierarchy. In paragraph two decision usefulness is explained, a model for decision usefulness is introduced and all the model components have a brief explanation.
2.1 What is Fair value accounting?
FVA is a way to measure assets and liabilities that appear on the balance sheet (Laux & Leuz, 2009). When assets or liabilities are obtained by a company there needs to be a value assigned to them. In accounting there are different ways to assign such a value. FVA makes use of the fair value of asset or liability. Fair value is 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 date1 (Laux & Leuz, 2009). This definition suggests that the value of an asset or liability is defined by its exit price: the price for which it can be sold. The problem is that for some assets (mostly financial) or liabilities there is no market. Without a market, there is no price (Benston, 2008). When there is no market available managers must make their own estimates and create the prices with the help of models. This requires a lot of subjective input and comparability decreases. For this reason the IASB and FASB introduced a hierarchy consisting of three levels that is based on the input used for the valuation (Laux & Leuz, 2009). Level 1 inputs are prices in active markets and this price is the basis for the fair value assigned to the asset or liability (Song, Thomas and Yi, 2010). An example of a fair value amount with level 1 input is a building you can buy in the house market (Benston, 2008). Level 2 inputs can be based on two kinds of input. There are the observable prices in an active market, but for comparable assets and liabilities or observable prices in an inactive market for identical assets and liabilities. Inputs could be prices that are based on market measures or yield curves. The input has a more subjective nature than level 1. So level 2 inputs create an estimated price rather than observational price (Song et al., 2010). An example of a fair value amount with level 2 inputs is a building internally built and used. The difference is that this fair value amount is based on prices of an active market but depends on prices of other buildings (Benston, 2008). Level 3 inputs are inputs that are not observable through the market and reflect management assumptions of how market participants would price the asset or liability. Inputs that are used could be discount rates that are set by managers to determine
1 Every time the words ‘fair value’ are used, this definition is meant with those words.
the correct present value of an asset (Song et al., 2010)2. An example of fair value amounts with level 3 input are harder to explain. Benston (2008) names a reporting unit, which is valued by a managers financial forecasts or a three year option on exchange traded shares.
The higher a level the more subjective the input gets and the less verifiable a fair value estimate is (Benston, 2008). This is the major problem with what the FASB and IASB have in mind with fair value. Other problems are that the estimation of fair value with level 2 or level 3 inputs can be costly or time consuming or that for some non-financial assets the assigning of fair value is very hard to because there simply is none and that fair value amounts can easily be manipulated (Benston, 2008). Other arguments are that FVA is irrelevant, misleading for long term financial assets and that prices could be distorted (Laux & Leuz, 2009). Arguments in favour of FVA are that it increases comparability, is more relevant than other ways of measuring, reflects real economic substance, it concerns with real value instead of costs or the fact that they are mostly market based and not firm specific (Penman, 2007; Cascini & DelFavero, 2011).
Examples of fair value application in standards are SFAS 107 or SFAS 140 for financial assets and SFAS 142 for goodwill. SFAS 107 was the first fair value standard that allowed estimates on the balance sheet. In the case of SFAS 107 this were fair value amounts of loans, securities (Landsman, 2007). SFAS 140 has the same content and requires managers to state financial assets at fair value (Dechow, Myers and Shakespeare, 2010). SFAS 142 requires a fair value estimation of the unit that needs to be impaired and an estimation of the eventual write-off (Jordan, Clark and Vann, 2007).
2.2 Decision usefulness
“The objective of financial reporting is to provide financial information about the reporting entity that is useful to existing and potential investors, lenders, and others creditors in making decision about providing resources to the entity (FASB, 2010).” Otherwise decision usefulness is used to provide financial information that is useful for stakeholders to make decisions regarding an entity. Decision usefulness is the number one priority of financial information. When the FASB or IASB introduces a new standard or sets new rules this is to create more or better decision usefulness for the users. The FASB (2010) guarantees that this information cannot be perfected and that the information always needs to be considered. For financial information to be decision useful it must consist of some qualitive and enhancing
2 For the rest of the study when fair value amounts are called estimates, or subjective etc. these fair values will be created with level 2 or level 3 input.
qualitive characteristics. The reason for this characteristics is that the information it provides is more useful for the users of this information and helps in making decisions (FASB, 2010). For financial information to be useful it must be relevant and faithfully represented. The information is enhanced when it can also be comparable, verifiable, timely and understandable (FASB, 2010). Relevance means that the financial information is capable of making a difference in the decision by users. These differences can be made when the information has predictive value, confirmatory value or a combination of the two. Beside these two characteristics the information must not be material, which puts a restriction on the relevance of the information (FASB, 2010). Faithfully represented means that the information represents economic phenomena, for this to be perfect the information must be free from error, complete and neutral (FASB, 2010).
Figure 2-1:This model gives a clear overview of decision usefulness and its elements according to the FASB and IASB.
The FASB understand that to create a perfect representation is an illusion and the qualities must be maximized to the extent possible. This idea of decision usefulness is based on the idea that relevance is more important than reliability. Relevance has the more present spot on
the model and the FASB says that faithful represented information cannot be decision useful on its own, but needs to be relevant to be decision useful (FASB, 2010).
2.2.1 Relevance
For financial information to be relevant it must have both a predictive value, confirmatory value and not be restricted by materiality in the numbers. The predictive value means that the financial information can be used as an input to processes employed by users to predict future outcomes (FASB, 2010). Financial information has confirmatory value if it provides feedback about previous evaluations. The confirmatory and predictive value are interrelated, which means that if numbers have a predictive value they most likely also have a confirmatory value (FASB, 2010). Information is material if omitting or misstating it could influence decisions made by the user of that information. The FASB does not state a uniform quantitative amount of materiality that information can have to be material or immaterial.
2.2.2 Faithful representation
For financial information to be faithfully represented it must be complete, free from error and neutral. ‘Complete’ means that all the information must be included that is necessary for a user to understand the phenomena being described. Neutral financial information means that it is represented without bias in selection and presentation. There must be no weighing, manipulation or emphasizing with the information to increase profitability. And free from error means that there are no errors made in the process of providing, describing and presentation of the information. The FASB also states that a faithful representation by itself has decision usefulness.
2.2.3 Enhancing Qualitative Characteristics
The qualitative characteristics are there to enhance the useful information. The characteristics are timeliness, verifiability, understandability and comparability. Comparability enables users to identify and understand similarity and differences in items. Verifiability means different capable observers can reach a consensus and come to the same numbers. Timeliness means that the information must be presented in time so that the information can influence their decisions when it is not too late. Understandability means that the financial information is not too complex to understand. Users can understand all the information.
3. Decision Usefulness of fair value accounting
In chapter two the model for decision usefulness was introduced. The main elements and its relevant literature will be discussed. First the literature with regard to relevance will be discussed and analysed. The second part surrounds literature regarding representation faithfulness. In the third and last paragraph two enhancing characteristics are discussed: verifiability and comparability.
3.1 Relevance
Relevance consists of two elements: predictive value and confirmatory value. The literature about confirmatory value of FVA is little and for this article relevance will consist of predictive value. Confirmatory value is replaced with materiality because of it constraining effect on relevance. So this paragraph discusses the connection between FVA and the predictive value of the information it produces and the materiality of this information.
3.1.1 Predictive value
When researches study the predictive value of FVA, they most commonly study this with the revaluation of assets. Revaluations correct older amounts and assign them a fair value. Researches use this fair value amount to find a correlation between these amounts and the future performance of the firm (Abroody, Barth and Kasznik, 1998). Examples of this sort of research are Abroody et al. (1998) or Lopes & Walker (2012). Abroody et al. (1998) suggest that revaluations can predict future firm performance. The sample that they use is from UK and Australian companies. They also find that revaluations predict changes in operating income and cash from operations, but that these predictions are dependent on a firms debt to equity ratio or economic circumstances. The research of Abroody et al. is just an example of a dozen revaluation studies. The problem with these studies is that the samples are mostly UK or Australian based. The reason for this problem is that revaluation research is limited to the sample countries where they permit revaluations. This is a logical problem since you can only research a topic when the rules permit it, but still almost all the research has a UK or Australian sample. Lopes & Walker (2012) try to use a Brazilian sample and they come to the same conclusion as Abroody et al.: fair value amounts have a predictive value.
Critique on these findings is that Brazil has the same set of rules as the UK and Australia, so the result is something you would expect. When the rules are the same, you hope that the result is the same. Barlev, Fired, Haddad and Livnat (2007) support this idea. They divide the world into four different accounting zones. An accounting zone is based on its standard and the rules it has. The sample they use is from 35 different countries that accept revaluations in their rules and they find that there is a difference in the predictive ability in
different accounting zones (Barlev et al., 2007). This sheds a new light on the predictive value of FVA. When the predictive value is dependent on the accounting zone, all the results from the previous study cannot be generalized. All the results must be placed in an accounting zone context. It even goes a step further: the predictive value also depends on the economic, cultural and legal structure (Barlev et al., 2007). Barlev et al. found that a firm in the same accounting zone, but with a different national structure can have a slightly different predictive value (Barlev et al., 2007). So the predictive value is not only accounting zone dependant, but also locally dependant. These results indicate that FVA has a dependant predictive value. Only the results of Barlev should be interpreted with caution, since you could doubt if the sample they use really represents the country or accounting zone.
Because revaluations are just a small fraction of the balance sheet, you could doubt whether revaluations are a good representation of the predictive value of FVA. This is a reason to look in other directions for a predictive ability. Goodwill impairment also seem to be a good indicator of future performance, but is limited to a one to two year range (Jarva, 2009). Although the revaluation literature do not suggest a timespan, we can assume that their predictive timespan cannot be longer than three or four years. The results of Jarva are replicated in a Spanish biological asset sample. Biological fair value assets seem to be a good indicator of future firm performance (Arigiles, Garcia-Blandon and Monllau, 2011). The problem with these two studies is if the samples (Biological assets and goodwill impairments) are enough to call them the reason of the prediction. Do these samples prove that FVA really has a predictive ability? Can these studies be used to generalize the predictive value?
The last industry to look for a predictive value is the banking sector. Evans, Hodder and Hopkins (2014), Cantral, McInnis and Yust (2014) and Barth, Landsman and Wahlen. (1995) all find evidence for a predictive value in this sector. The problem with these findings is that they do not fit with the idea of the predictive value according to the IASB and FASB. One study finds that fair value amounts can predict credit losses, which is more useful for credit agencies (Cantral et al., 2014). Another finds that fair value amount can predict regulatory capital violations, but that it cannot predict future stock prices of banks (Barth et al., 1995). Although this finding is contrasted in the results of Evans, where fair value gains and losses can predict future bank share prices (2014). The IASB and FASB want an accounting concept to predict future firm performance and I doubt if credit losses or regulatory capital violation fit this picture. Overall I conclude that FVA has a predictive value, because of the abundance of evidence for different part of the balance sheet.
3.1.2 Materiality
The reason to study for a connection between materiality and FVA is that FVA produces numbers that are based on estimates that could potentially carry a lot of materiality. Their subjective nature arises from managerial input and unreliable nature. Auditors are sceptical about fair value amounts, because of their unreliability and complexity (Early, Hoffman & Joe, 2012). A problem with the study of the connection between materiality and FVA is that the literature is limited. Auditors find it hard to assign materiality to fair value estimates and researches have a hard job to research this assigning.
Griffin (2011) finds that auditors require companies to adjust fair value amounts when these amounts consist of subjective estimates and could possibly carry a lot of imprecision. These results imply that auditors think that fair value estimates carry a lot of materiality and need to be adjusted (Griffin, 2011). Christensen, Glover and Wood(2012) support this finding on an a priori base. From public data they theorise that fair value amounts contain extreme measurement uncertainty and that changes in small model input like interest rates or discount rates can change estimates by more than materiality. The problem is that these ideas are not supported with empirical data and are mere thoughts. Although they support it with a good basis, they do not have real evidence that the input they suggest create a larger materiality. Although these two articles state that auditors assign materiality to fair value amounts, an investigation by the Public Company Accounting Observation Board (PCAOB) denies this (Bell & Griffin, 2012). In 2010 the PCAOB made a statement about this investigation. The result was that auditors fail to evaluate fair value assumptions and when auditors assumed materiality they assumed way to high amounts. They did not have the knowledge to assign this materiality. Auditors are not ready to audit fair value estimates (Bell & Griffin, 2012). This study has the same problem as the a priori study: there is no empirical evidence. A reason could be that materiality is hard to test.
To look for other evidence, we could look at a different research field: audit fee literature. Audit fees are based on the clients’ size, risk and complexity (Hay, Knechel & Wong, 2006). You could assume that when a firm has more fair value on the balance sheet it could be assumed to be more complex, which would result in an audit fee increase. From this assumption we could also assume that an increase in audit fees means more materiality. Goncharov, Riedel and Sellhorn (2014) and Ettredge, Xu and Yi (2013) give evidence for the relationship that more complex fair value amounts increase the audit fees, but that the total amount of fair value in the financial statement decrease the audit fee, a result that corresponds with the findings of Hay.
But is this literature enough to suggest that fair value amounts are material? All the studies in the first paragraph suggest that materiality is based on complexity and uncertainty, but there is no real hard evidence on the influence of FVA on the decisions investors make or how the deviation in fair value amounts change these decisions. Overall is it difficult to assign materiality to fair value amounts and auditors still have a problem with this.
3.2 Representation faithfulness
Representation faithfulness is split up in three parts: completeness, free from error and neutral. Since there is almost no literature about completeness this part will be omitted for the same reason as confirmatory value is omitted. There is no element that replaces completeness, so the other two will suffice to discuss the representation of faithfulness.
3.2.1 Free from error
It is hard to provide any numbers without errors. Information is rarely without errors. There never is a perfect market to calculate the numbers and with an incomplete market estimation errors are made (Barth & Landsman 1995). In compliance with the SFAC model Barth & Landsman (1995) say that numbers without errors are a mere illusion, but what information must do is contain the least amount of errors. FVA is known now for the subjective input and because of this always has some mistakes. The most important benchmark are the investors, since they are the most notable users of financial statements.
Investors see the SFAS 157 and IAS 13 hierarchy as a reliable way to rank fair value estimates (Kolev, 2011). Investors see level 1 estimates as perfectly reliable because the input is market based. Investors also surprisingly see level 2 and level 3 as reliable enough to represent firm value, even though they are model based. But the findings are not strong for level 3 estimates (Kolev, 2011). Song et al. (2010) refute these findings and come to the conclusion that investors place less weight on level 3 estimates and see them as less reliable. The reason for this is that these estimates have an information risk, estimation errors and there is possible reporting bias. The reliability of these numbers can be increased when there is higher corporate governance in a company, because investors assign higher reliability to fair value amounts because of the drop in information risk etc. (Song et al, 2010). The conclusion from these results is that investors see the less subjective numbers as reliable and free from error enough, but when subjectivity increases investors rely on mechanisms that are beyond FVA.
Corporate governance is not the only thing that attributes to the reliability of fair value amounts and reduction of errors. Another mechanism is the inclusion of more reliability
disclosures in the financial statements (Chung, Goh, Ng, & Yong, 2011). Managers understand that fair value amounts are complex and contain a lot of subjectivity. When the subjectivity increases managers tempt to increase the number of reliability disclosures in the financial statements. Investors or analysts can use these disclosures to understand and perhaps correct fair value estimates (Chung et al., 2011). These reliability disclosures admit that FVA produces information with errors. Reliability disclosures help stakeholders understand the errors and correct them to reduce the amount of errors. Paananen (2008) complements the findings of Chung et al. about disclosures. Disclosures help to give more information, but when these disclosures do not have a high enough quality they are useless.
Besides corporate governance and reliability disclosures there are external factor that influences the reliability of fair value estimates. Riedel & Serafiem (2011) call this the quality of the information environment which consists of four inputs: market capitalization, forecast errors, analyst following and analyst forecast dispersion. For example a small firm has low market capitalization and is less attractive to follow for analysts and investors, because of this the company has low analyst following. Therefore analyst do not correct the financial statements and there will be high forecast errors. Overall the reliability of the financial statements decrease and so the errors in these statements are high. This also works for fair value amounts (Riedel & Serafeim, 2011). But the problem with the Riedel & Serafeim study is that they use a model to determine beta’s of companies and not the reliability, although they make the claim that beta’s reflect information risk of a company and information risk is negatively correlated with the quality of the information environment.
What do all these mechanisms have to do with free from error? The results of these studies imply that FVA produces too much information with errors and firms need to create tools to reduce these errors. Although these results do not directly talk about errors, we can imply this because they are not talking about neutrality and representation faithfulness consists of neutrality and free from error. The conclusion is that FVA is wholly responsible for the reliability of its amounts, but firms need to protect itself from the errors that FVA can produce.
The Enron case gives a good analysis of the mechanisms. Enron created fair value estimates that were based on unreliable models, which created errors in these estimates. The estimates were unreliable because the accountants that made them stated them too optimistic and were not independent (Gwilliam & Jackson, 2010). As said before there are always errors in fair value estimates but the mechanisms introduced above can reduce these estimates. If we analyse the Enron case we see that if there was higher corporate governance
the estimates would have never been accepted, they would have been seen as unreliable or managers would have stated more reliability disclosers because of the uncertainty in these amounts. As there is a lack of internal mechanisms the analyst were unable to react. They did not have a good overview of how the estimates came to be and correction was impossible, which contributes to a low quality information environment. So the errors that came because of FVA was not only because of the accounting, but also because of Enron’s own failure.
Benston (2006) disagrees with this. He blames the demise of Enron wholly on FVA. Because the mandatory adoption of FVA provided managers with the opportunity to overstate fair value amounts and gave an unfair representation of Enron’s overall financial statements. The reason for the managers to do this was because of a monetary incentive. Their bonus was connected to the amount of fair value in the statements (Benston, 2006). We see again that FVA produces errors on its own and it needs help from companies to reduce this. Of course in this case managers would not even consider reliability disclosures because they act out of own interest and maybe analysts would find out about the overstatement if there were better internal mechanisms
The relationship that Benston described has been researched by Muller, Riedel and Sellhorn (2011). They find that mandatory adoption of FVA lead to less reliable fair value amounts, which the case study of Benston suggests. The problem with this study is that the study is incremental. It only compares the mandatory adoption with the voluntary adoption, which means that mandatory adoption still produces reliable numbers, but they produce less reliable numbers than voluntary adoption.
To summarize this paragraph: research shows that investors think level 2 and level 3 estimates have reliability (Kolev, 2011; Song et al., 2010; Chung et al., 2011). The trust in these numbers means that investors see them free from error enough to be reliable. The literature also states that the amount of errors in financial statements can always be decreased by firms, since complementing research suggests that FVA cannot do this on its own. Firms can use corporate governance or providing more reliability disclosures and thereby getting help from analysts to increase reliability and decrease the errors. For FVA to be free from errors is therefore impossible, it will always have some subjective input and mistakes are easily made. Inventors though see it as free from error enough, but think that firms have a responsibility in creating this reliability. They must accept FVA and try to use it to produce more reliable amounts. This complies with the results of Muller et al. (2011) who state that when FVA is more accepted and not a no-choice for a company it will result in more reliable amounts and there will be less errors.
3.2.2 Neutral
The overstatement of fair value amounts by Enron is just one example of non-neutrality created with the help of FVA. Benston (2006) argued that managers used the space created in the fair value standards to make these overstatements. The overstatements were made to let the statements seem more favourable, so the managers could receive their bonus. Different fair value based standards are criticised for the smoothing of income or earnings management and thereby creating non-neutral fair value numbers. There are a few standards that will be evaluated.
The first standard that will be examined is IFRS 3, which contains rules about goodwill write-off and impairment testing. IFRS 3 gives room for managers to report in an opportunistic way. The flexibility in standards gives managers the room to report write-offs on the moment that serves the managers interest best (Abughazaleh, Al-Hares & Roberts, 2011). Just like with the amount of errors in 3.2.1 flexibility is limited by strong governance mechanisms. When a company has strong governance mechanisms managers have far less incentive to report opportunistic, they will report more in line with the expected economic performance of the firm and the write-offs make economic sense (Abughazaleh et al., 2011). The overall result is that IFRS 3 succeeds in its task to let firms reflect their relying economic attributes and help give neutrality. Lhaopadchan (2010) suggests in his literature review that IFRS 3 and SFAS 142 do not produce neutral information. According to the studies he evaluated IFRS 3 and SFAS 142 give room for earnings management, because of the growing amount of goodwill in the financial statements and the uncertainty & management discretion that is used for estimating them. Still when you compare their results then Abughazaleh et al. make a better point. He tested for the effect of governance and when we take this into account the standards seem to be more neutral creating. The conclusion is that the neutrality is not guaranteed by the standard itself. IFRS 3 needs corporate mechanisms (in this case governance) to guarantee its neutrality.
The GAAP counterpart of IFRS 3 is SFAS 142. The content of this standard is also mostly about goodwill write-offs. When SFAS 142 was introduced in 2002 there was a lot of commotion about its effects on earnings management (Zang, 2008). In the year of introduction managers did use the flexibility in SFAS 142 to smooth income (Jordan et al., 2008). Larger write-offs were created than the years prior to the adoption of SFAS 142. Different incentives could drive this choice, but for SFAS 142 two stood out. Managers use the flexibility of the standard to overstate write-offs for creating higher future earnings. They do this to secure their position for later (Zang, 2008). The overstatement happens when there
is a manager rotation in the company. Jarva (2009), also finds this result but he comes to the result in a different way. Jarva looks at the write-offs versus the expected write-offs and looks if the two match. When the actual write-offs are bigger than the expected write-offs than there is earnings management. The understatement of write-offs is found not only used out of self-interest. Managers understate write-offs when firms reach their debt covenant constraints and managers try to boost earnings to reach the minimum set out in the constraints (Zang, 2008). These results gives a clear example of how flexible SFAS 142 can be. You could overstate when you need high future income and you could understate when you need high present income. As Lhaopadchen already made clear, managers could always defend their reason for over- or understatement because it are their own estimates. This supports the findings of Zang. Managers can bend the rules in exactly their own interest. The conclusion for SFAS 142 is that it can leave a lot of room for flexibility and therefore creating numbers that do not seem very neutral. It does not matter if these numbers come out of self- or company-interest. The numbers somehow show a picture of the company that is not neutral.
SFAS 107 sets out rules about loan fair values. Just like SFAS 142 the rules are very flexible and leave room for accounting discretion. The difference with the results of SFAS 142 is that the flexibility is priced positively, which means that investors perceive the information through discretion to be useful for them (Beaver & Venkatachalam, 1995). The investors see the discretion as a way of signalling the firms performance by managers (Beaver & Venkatachalam, 1995). The problem is that this study is very old and managers have become smarter with the use of the flexibility. Another problem is that this result is found in loan fair value amounts and cannot be transferred to other standards like SFAS 142 or SFAS 157. This signalling could be positive, but could also be used to mislead users. In the case of banks, they overstate loan fair values to try and give a positive signal about their risk and performance (Nissim & Penman, 2007). The problem is that there is not any other study done on the connection between earnings management and SFAS 107 and that these results are based on very old samples (1994, 1995).
SFAS 140 leaves room for flexibility because managers can set their estimates of default rates, prepayment rates or discount rates to calculate fair value. These fair value amounts can be used to create gains or losses because securitizations are treated as sales (Dechow et al., 2010). This leaves room for managers to make gains higher when a firm has lower earnings than the previous year. Managers also use the discount rate of their estimates to reduce the size of losses (Dechow et al., 2010). Not only the discount, default or prepayment rate leave room for flexibility, but also the timing in SFAS 140 can be used to
manage earnings. Dechow & Shakespeare (2009) find that managers time manage the time of reporting securitizations to meet earning thresholds. Securitizations are mostly reported at the third month of the quarter or even more in the last five days of the quarter. These results are too easily called earnings management. Barth & Taylor (2009) criticize the findings of Dechow et al. They think that their results do not imply earnings management. These results can only imply choices made by higher management or choices made by another interested parties, but the results do not imply any earnings management (Barth & Taylor, 2009).
The SFAS 140 literature gives the best example of the problems concerning fair value neutrality. When managers have the option to estimate their own rates they will use this in their own interest. Managers will use these rates so that they can collect their bonus or out of fear for losing their jobs because of higher losses. If this is compared to the small literature of SFAS 107, banks use rates to estimate loan fair value that serve their own interest. As Nissim & Penman (2007) found that smaller banks have more need to use the flexibility than large banks because they bear higher risk. The same counts for SFAS 142 and IFRS 3 in a different sense. The literature suggest that standard setters leave to much flexibility in their standards to ever create neutrality. When the FASB or IASB want real neutral numbers they need to give benchmarks about what proper rates are. Every opportunity a manager gets to use flexibility in his own interest gives him an incentive to use it. With that path the creation of neutral number is impossible, because there will always be bias and FVA will always will be a tool for earnings management.
The problem with all these studies is that they study one standard. They put their focus on just one standard and never look at the broad picture of FVA. The results all these studies bring is not enough to call FVA non-neutral. There are some studies that provide evidence on the non-neutrality of SFAS 157 (Chung et al., 2011; Danbolt & Rees, 2008), but these studies use samples like real estate to determine that bias is created with the rules from these standards. Although this problem is not easily resolved, I assume that the standards and articles studied in this paragraph are just a small fraction of all the non-neutrality in fair value standards.
3.3 Enhancing Qualitative Characteristics
The decision usefulness model describes four enhancing qualitive characteristics that complement relevance and representational faithfulness. In the literature review the literature of two of these characteristics will be described: verifiability and comparability. Not only because these are the most important of the four, but also because these two are the best choices for discussion. One reason for this is that the literature on verifiability and
comparability is by far the most. Timeliness is hard to study and understandability lies very close to verifiability. Another reason is that in the old model of SFAC No. 2 verifiability and comparability had a more prominent spot (FASB, 1980). The FASB saw the importance of these elements but denied them an important spot in the new model.
3.3.1 Comparability
Comparability consists of three factors: 1) having similar input, 2) similar procedures 3) the same system of classification (Barlev & Hadded, 2007). Although the article of Barlev & Hadded is published in 2007 the article itself is older. According to them FVA was the procedure to create comparable numbers. They admitted that to create fair value numbers different input was needed, but that would get compensated with the use of similar procedures and all firms using the same system of classification (Barlev & Haddad, 2007). The two were not completely wrong. At this time FVA is implemented in different standards that firms have to use: similar procedures to create different fair value amounts, and with the introduction of SFAS 157 and IAS 13 there is created a system of classification that is the same for every company: the fair value hierarchy. So perhaps Barlev & Haddad were right that FVA could eventually lead to comparability was it not that the problem lies with having similar input. They thought that the implementation of fair value meant that the time of measurement was the same for all the parts of the balance sheet. This was not only the same for one company but for all the different companies in the world. Perhaps it sounded right at the time the article was written, but it sounds naïve now.
For example the results of Dechow et al. (2010) suggest that managers use different amounts of discount rates to calculate the present value of their income. To calculate fair value amounts, different kinds of input are used. All the literature of goodwill suggest that managers manage the write-offs and can determine the size of these write-offs. This helps to overstate or understate the goodwill amount on the balance sheet (Zang, 2008; Jordan et al., 2008; Jarva, 2008). The result is that different amounts of input are used resulting in different outputs, suggesting that comparability is low. The flexibility of fair value that is discussed in section 3.2.2. could have a negative impact on comparability. Managers use the room to choose their own input and the input of one manager is almost never the same as the input of another resulting in bad comparability.
A way to solve this comparability problem is to see comparability from a different perspective: not as an input output relationship (to compare if the same input creates the same output), but as a way to distinguish firms from one another. Barth & Venkatachalam (1995) found that discrete fair value amounts could be used to signal investors and other financial
statement users. Comparability is then the way to see one firm as firm A and one firm as firm B and to examine why the differences occur when the same procedures are used. If you look at FVA from this perspective, FVA succeeds in its role of comparability.
The main role that SFAC No 8 assigns to comparability is that it can help the users of information compare one company to another and not to look at the differences. Although QC21 states that it is not only about similarities, but also about differences the focus looks to be mainly on comparing similar items that are produced through similar input (FASB, 2010). This is hard for FVA to achieve because of the subjective input. A solution for this problem has already been offered in section 3.2.1: increasing reliability disclosures could help increase comparability between similar items (Paananen, 2008; Chung et al., 2011). Reliability disclosures give information about the way numbers are created and can be used to compare firms when same producers are used but different kinds of input. This conflicts with the model of Barlev & Hadded, but it can result in adjustment by analysts and investors resulting in comparable input.
3.3.2 Verifiability
Like materiality there is no evidence on the question how unverifiable fair value amounts are. The only way to study verifiability is with indirect relationships. Assumptions are needed to test whether fair value amounts are verifiable or unverifiable. Ramanna and Watts (2012) assume that SFAS 142 has an unverifiable nature and that managers use this nature to manage earnings. They assume a relationship between unverifiable estimates and earnings management. Section 3.2.2 has already set out the relationship between earnings management and FVA and these results can help us with the verifiability of FVA.
The results of section 3.2.2 suggest that the estimates that different standards require help managers to manage their earnings and bias amounts (Dechow et al., 2010; Chung et al., 2011; Jarva, 2009). These estimates all have an unverifiable nature because they require input of a manager that only the manager can verify and this is not what the FASB and IASB see in verifiability. They explicitly state that there must be different observers that come to the same conclusion with the financial information (FASB, 2010). And can observers come to the same conclusion? This is a hard question to answer. We can assume that when managers manage their earnings they try to hide this from observers, because they want no one to discover. This makes fair value amounts unverifiable according to the FASB. The only problem is that this assumption is never empirically supported. The evidence that section 3.2.2 gives is that managers use flexibility to manage earnings, but not that this flexibility comes from creating unverifiable numbers.
Although the assumption is never empirically supported (except maybe by Ramanna and Watts (2012)), while reading the articles on the relationship between FVA and earnings management you could see that researchers accept the assumption. When they talk about the problems of fair value estimates, they recognise that one part of the problem one is how unverifiable these estimates are because of the subjective input (Dechow et al., 2010; Chung et al., 2011; Jarva, 2009). Theoretical articles also say that the production of unverifiable financial information is one of the major problems of FVA (Penman, 2007; Benston, 2008). All these articles do not solve the problem of how weak empirically supported all these claims are. This lack of evidence does not make the claim strong that FVA produces financial information with weak verifiability. The problem can only be solved with more empirical research instead of a priori arguments. Although I support the a priori arguments on the verifiable nature of FVA. The subjective input and the need of estimation for level 2 or level 3 amounts is so high that to verify these numbers can be very costly and hard. This connection gives FVA a very unverifiable nature and is logically made by the researchers.
4 Discussion
This section gives a brief overview of the general findings in the literature about decision usefulness. After that I will give an answer to the research question. This answer will be given in two different ways, first the original model of decision usefulness will be used to answer the question. Then I will introduce a revised model which in my eyes complies more with the results of the literature. These two answers will be compared and a final answer will be given.
In researching the literature on FVA, I have found that FVA is generally seen as decision useful. But as always the literature makes some remarks about this decision
usefulness. The decision usefulness depends on the level of input that a fair value amount or estimate has. As stated many times before, the higher the level of subjectivity the less
researchers trust its decision usefulness. The literature indicates that the introduction of a real fair value definition and a way to classify the fair values is appreciated. SFAS 157 helps researching fair value now and it gives FVA a possibility to be researched better in the future, this being so because of the data that become available in a few years’ time.
The literature also gives some remarks with the introduction of a definition and a classification system. First they see that the introduction of the levels gives managers the opportunity to justify vague financial asset values and give them a place on the balance sheet with the only clarification being that it is made out of level 2 or 3 input, making it easier for managers to justify their possible use of earnings management. The second problem is that FVA is a catalyser for measurement error. In contrast with other forms of accounting (in particular historical cost accounting) FVA could not easily assign a value to an asset or liability, even in the lowest level of input FVA could already have measurement problems. How can the price in an active market at the exact time X be said to be the ‘fair value.’ Eventually these problems do have their impact on the decision usefulness of FVA, they can limit it or they can, when fully understood, increase it.
The question of this research was to what degree FVA can be seen as decision useful. Decision usefulness is seen as the most important objective of financial accounting according to the IASB and FASB. The information that any accounting concept must produce, must have the power to let stakeholders make decisions. To describe this decision usefulness section two introduced a model, the literature surrounding the most important components of this model was analysed in section three. The results from this analysis show that according to the model FVA is indeed decision useful. Underlying the model is the IASB’s and FASB’s idea that relevance is more important than reliability (replaced in the model with
information lies in their possibility to change the thought of a stakeholder and try to make a difference when the information is perceived by the stakeholder. The IASB and FASB give little attention to the verifiability, or if the numbers give a true and fair view of the economic reality the financial statements want to represent In the following tables I will summarize my findings on every element.
Element Conclusion
Predictive value - FVA has a strong predictive value. Research shows that FVA can predict future cash flows and help investors make predictions and forecasts on their own. - Because of the interrelation between predictive
and confirmatory value, I assume that FVA also possess the confirmatory ability.
Materiality - Materiality can become a huge problem for FVA. Analysis of the research indicate that fair value numbers can possess a lot of materiality.
- Auditors need to be trained in auditing fair value amounts and learn the process of the creation and estimation of fair value numbers.
- The overall problem is that the study of materiality is very young and still needs to grow.
Relevance - The conclusion is that FVA produces relevant
financial information according to the model of decision usefulness.
Table 4-1: The relevance of FVA
Element Conclusion
Free from error - Just like every other accounting concept FVA can be impossible to be free from error.
- The errors can be reduced with the help of some firm specific mechanisms
Neutral - FVA leaves a lot of room for management bias,
because of its flexibility.
- Managers abuse this flexibility to increase profitability or decrease losses.
Representation faithfulness - FVA cannot be called representation faithful according to the model. It does not succeed in the most important element: that of neutrality.
Element Conclusion
Comparability - The comparability of FVA information is hard to determine, but the problem lies mostly in the different input that is used. This input differs for almost every estimated amount.
- Because of this problem FVA does not produce very comparable numbers, but they have the potential to become comparable if reliability disclosures are
included.
Verifiability - The verifiability of fair value numbers is just as hard as their comparability. Because of subjective input it is hard to verify them, since subjective truth is hard to verify.
- The problem with the study of verifiability is that just like materiality it is not studied much and I have found the results out of indirect studies.
Enhancing characteristics - FVA does not succeeds in fulfilling the two most important characteristics. The IASB and FASB will say that this is not that important, because they are just to enhance the decision usefulness but I think that these elements are important. This will be shown in the final answer.
Table 4-3: The enhancing characteristics of FVA
The final answer to the question is that FVA is decision useful as defined by the IASB and FASB, because of the reason that relevance is said to be more important than reliability. I strongly doubt this answer because of the fact that the literature puts a strong emphasis on the problems discussed in the first paragraph of this section. For this reason I want to introduce an alternative model of decision usefulness where the emphasis is on reliability. This alternative model gives, according to me, an idea of the thoughts that researches have concerning what is important for FVA.
Figure 4-1: The revised model of decision usefulness
Because of the problems the literature suggest I think that a revised definition of decision usefulness is needed to really test if FVA is decision useful. In this new definition as set out in the model, the financial information needs to be reliable and after that comes the relevance. I believe that reliability and relevance are interrelated. This idea is supported by the FASB in SFAC No. 2 where they still believed that relevance could not be useful on its own and needed reliability (and vice versa). In this old model verifiability was also an element of reliability and not a component to enhance the decision usefulness. The reason I put this in the model is because the literature suggest that the verifiability of fair value numbers is hard, but that without verifiability it is hard to believe the numbers can mean something, so they are irrelevant. The reason to put reliability before relevance is because of all the articles that are written about flexibility, earnings management and the errors in FVA. There were so many articles that study these problems in combination with FVA that they are a threat to the decision usefulness of FVA.
When this model is used to determine the decision usefulness of FVA, FVA fails the test. When reliability is put on a same or even higher level than relevance FVA cannot be decision useful. The reason for this is explained many times: subjective input, estimation errors, management bias. The literature results suggest that these problems do exist and the model of the IASB and FASB does not acknowledge that. By setting reliability last they think that relevance is enough to create decision usefulness but it is not. When numbers are not reliable it will create no problems on the short term but it will on the long term, when it is discovered that the numbers are not reliable enough.
We have seen from the literature that there are ways to solve the problems that are raised by the literature: corporate governance, reliability disclosures or creating a better information environment, but these mechanisms are only needed when the problems are acknowledged by standard setters. Companies can choose for themselves to create a higher corporate governance or to put more reliability disclosures in the financial statements, of course they are not obligated by any institution, but they should at least be given some sort of incentive. When an incentive is given in the form of standards that work more towards
reliability instead of relevance they will have a reason to create more reliable information. What does this change for the answer on the research question? It shows that if the settings are changed than FVA would not be decision useful at all, it could not be used because it would create misleading information. So FVA is useful to the degree that it produces relevant information, but lacks to produce reliable information. The revised model commits to this argument by showing that when reliability is more important than relevance FVA fails to be decision useful. When the components set by the FASB and IASB define decision usefulness FVA fails to create reliable information.
Limitations of this study are that not every component of the model has been
thoroughly researched. There is a lot of indirect research on the different components which means that the literature cannot always give full explanation of the relationship. Examples of this are the materiality that fair value amounts could carry or the real verifiability of fair value amounts. The second limitation is that most of the studies that are analysed are US based, although most standards are almost the same, but with respect to their content there could be some minor differences. The study could be more complete, with extra European based studies.
5 Conclusion
This study tried to answer the question on to what degree fair value accounting succeeds in being decision useful. The conclusion is that based on the model introduced SFAC No. 2 and arguments delivered by the literature, FVA succeeds in providing relevant financial
information, but fails to deliver reliable financial information. I came to this conclusion through studying and analysing different fields of literature, some containing many articles and some only represented in three or four articles. These articles provided evidence on the relevance part of decision usefulness, but also on the dangers of FVA. With these dangers in mind I formulated a different model with a different idea of decision usefulness. In this model decision usefulness still consists of relevance and reliability, but relevance was not as
important as before. I believe that this model showed what researches thought about
reliability: it is something important and portrays real problems for FVA. With this model I examined the reliability with the analysed literature and came to the conclusion that fair value amounts are not reliable enough. So with the literature I showed that FVA is relevant, but not reliable on its own.
There are some problems relating to this conclusion. There are some components of the model that are not researched thoroughly. So I suggest for future research to try avoiding studies of the general idea of value relevance or decision usefulness, but focus more on one component of decision usefulness. The reason is that some components like materiality or verifiability are researched on a low scale, which means that the connection between FVA and materiality/verifiability is open for research. The second problem is that the research is very US based. A way to solve this problem is obviously conducting more research in countries outside of the US. The third suggestion for research is to look more at the reliability-relevance relationship. Which of the two is really more important and what do investors perceive is the most important part of decision usefulness? By looking at the fundamental distinction between reliability and relevance, we could shed yet another light on the discussion of FVA.
6 References
Aboody, D., Barth, M. E., & Kasznik, R. (1999). Revaluations of fixed assets and future firm performance: Evidence from the UK. Journal of Accounting and Economics, 26(1), 149-178.
AbuGhazaleh, N. M., Al‐ Hares, O. M., & Roberts, C. (2011). Accounting discretion in goodwill impairments: UK evidence. Journal of International Financial Management &
Accounting, 22(3), 165-204.
Argilés, J. M., Garcia-Blandon, J., & Monllau, T. (2011). Fair value versus historical cost-based valuation for biological assets: Predictability of financial information. Revista De
Contabilidad, 14(2), 87-113.
Barlev, B., Fried, D., Haddad, J. R., & Livnat, J. (2007). Reevaluation of revaluations: A Cross‐ Country examination of the motives and effects on future performance. Journal
of Business Finance & Accounting, 34(7‐ 8), 1025-1050.
Barlev, B., & Haddad, J. R. (2003). Fair value accounting and the management of the firm.
Critical Perspectives on Accounting, 14(4), 383-415.
Barth, M. E., & Landsman, W. R. (1995). Fundamental issues related to using fair value accounting for financial reporting. Accounting Horizons, 9, 97-107.
Barth, M. E., Landsman, W. R., & Wahlen, J. M. (1995). Fair value accounting: Effects on banks' earnings volatility, regulatory capital, and value of contractual cash flows. Journal
of Banking & Finance, 19(3), 577-605.
Barth, M., & Taylor, D. (2010). In defense of fair value: Weighing the evidence on earnings management and asset securitizations. Journal of Accounting and Economics, 49(1), 26-33.
Beaver, W., & Venkatachalam, M. (2003). Differential pricing of components of bank loan fair values. Journal of Accounting, Auditing & Finance, 18(1), 41-68.
Bell, T. B., & Griffin, J. B. (2012). Commentary on auditing high-uncertainty fair value estimates. Auditing: A Journal of Practice & Theory, 31(1), 147-155.
Benston, G. J. (2006). Fair-value accounting: A cautionary tale from enron. Journal of
Accounting and Public Policy, 25(4), 465-484.
Benston, G. J. (2008). The shortcomings of fair-value accounting described in SFAS 157.
Journal of Accounting and Public Policy, 27(2), 101-114.
Cantrell, B. W., McInnis, J. M., & Yust, C. G. (2013). Predicting credit losses: Loan fair values versus historical costs. The Accounting Review, 89(1), 147-176.
Cascini, K. T., & DelFavero, A. (2011). An evaluation of the implementation of fair value accounting: Impact on financial reporting. Journal of Business & Economics Research,
9(1)
Christensen, B. E., Glover, S. M., & Wood, D. A. (2012). Extreme estimation uncertainty in fair value estimates: Implications for audit assurance. Auditing: A Journal of Practice &
Theory, 31(1), 127-146.
Chung, S. G., Goh, B. W., Ng, J., & Yong, K. O. (2013). Voluntary Fair Value Disclosures Beyond SFAS 157’s Three-level Estimates. Available at SSRN 1335848.
Danbolt, J., & Rees, W. (2008). An experiment in fair value accounting: UK investment vehicles. European Accounting Review, 17(2), 271-303.
Dechow, P. M., Myers, L. A., & Shakespeare, C. (2010). Fair value accounting and gains from asset securitizations: A convenient earnings management tool with compensation side-benefits. Journal of Accounting and Economics, 49(1), 2-25.
Dechow, P. M., & Shakespear, C. (2009). Do managers time securitization transactions to obtain accounting benefits? The Accounting Review, 84(1), 99-132.
Earley, C. E., Hoffman, V. B., & Joe, J. R. (2012). Are auditors skeptical of management's level 2 versus level 3 fair value classification judgments? Available at SSRN, 2119720 Ettredge, M. L., Xu, Y., & Yi, H. (2009). Fair value measurements, auditor industry expertise,
and audit fees: Evidence from the banking industry. Social Science Research Network,
Evans, M. E., Hodder, L., & Hopkins, P. E. (2014). The Predictive Ability of Fair Values for Future Financial Performance of Commercial Banks and the Relation of Predictive Ability to Banks' Share Prices. Contemporary Accounting Research, 31(1), 13-44.
FASB. (1980). Statement of financial accounting standards No. 2: Qualitative characteristics of accounting information. Stamford, CN: FASB.
FASB. (2006). Fair Value Measurements. Statement of Financial Accounting Standards No.157. Norwalk, CT: FASB.
FASB. (2010). Statement of financial accounting standards No.8: Conceptual framework for financial reporting. Stamford, CN: FASB
Goncharov, I., Riedl, E. J., & Sellhorn, T. (2011). Fair value and audit fees. Review of
Accounting Studies, , 1-32.
Griffin, J. B. (2011). The effects of uncertainty and disclosure on auditors’ fair value materiality decisions. Available at SSRN 1886512,
Gwilliam, D., & Jackson, R. H. (2008). Fair value in financial reporting: Problems and pitfalls in practice: A case study analysis of the use of fair valuation at Enron. In Accounting Forum 32(3), 240-259.
Hay, D. C., Knechel, W. R., & Wong, N. (2006). Audit fees: A Meta‐ analysis of the effect of supply and demand attributes*. Contemporary Accounting Research, 23(1), 141-191. IASB. (2011). Fair Value Measurement. International Financial Reporting Standard No. 13.
Retrieved from: http://www.iasplus.com/en/standards/ifrs/ifrs13
Jarva, H. (2009). Do firms manage fair value estimates? an examination of SFAS 142 goodwill impairments. Journal of Business Finance & Accounting, 36(9‐ 10), 1059-1086.
Jordan, C. E., Clark, S. J., & Vann, C. E. (2011). Using goodwill impairment to effect earnings management during SFAS no. 142’s year of adoption and later. Journal of
Business & Economics Research (JBER), 5(1)
Kolev, K. (2009). Do Investors Perceive Marking-to-model as Marking-to-myth?: Early Evidence from FAS 157 Disclosure (Doctoral dissertation, New York University, Graduate School of Business Administration).
Landsman, W. R. (2007). Is fair value accounting information relevant and reliable? evidence from capital market research. Accounting and Business Research, 37(sup1), 19-30. Laux, C., & Leuz, C. (2009). The crisis of fair-value accounting: Making sense of the recent
Lhaopadchan, S. (2010). Fair value accounting and intangible assets: Goodwill impairment and managerial choice. Journal of Financial Regulation and Compliance, 18(2), 120-130.
Lopes, A. B., & Walker, M. (2012). Asset revaluations, future firm performance and firm-level corporate governance arrangements: New evidence from brazil. The British
Accounting Review, 44(2), 53-67.
Muller III, K. A., Riedl, E. J., & Sellhorn, T. (2011). Mandatory fair value accounting and information asymmetry: Evidence from the european real estate industry. Management
Science, 57(6), 1138-1153.
Nissim, D., & Penman, S. H. (2007). Fair value accounting in the banking industry.
Paananen, M. (2008). Fair value accounting for goodwill under IFRS: an exploratory study of the comparability in France, Germany, and the United Kingdom. Germany, and the United Kingdom (September 30, 2008).
Penman, S. H. (2007). Financial reporting quality: Is fair value a plus or a minus? Accounting
and Business Research, 37(sup1), 33-44.
Ramanna, K., & Watts, R. L. (2012). Evidence on the use of unverifiable estimates in
required goodwill impairment. Review of Accounting Studies, 17(4), 749-780.
Véron, N. (2008). Fair value accounting is the wrong scapegoat for this crisis. Accounting in Europe, 5(2), 63-69.
Zang, Y. (2008). Discretionary behavior with respect to the adoption of SFAS no. 142 and the behavior of security prices. Review of Accounting and Finance, 7(1), 38-68.
