A Comparison Between Impairment and
Amortisation in a Dutch Setting
Master Thesis
Justin Verbond
9413774
First supervisor: Dr. G. Georgakopoulos
Second supervisor: Dr. R. Boomsma
First draft: January 20, 2013
Final draft: January 30, 2013
Master’s in Accountancy & Control
Faculteit Economie en Bedrijfskunde
Universiteit van Amsterdam
Abstract 2
1. Introduction 4
2. Literature Review 6
2.1 Quality of accounting information 6
2.2 Goodwill accounting 8
2.3 Regulation on goodwill accounting 10
2.4 Amortisation of goodwill 14
2.5 Impairment of goodwill 16
2.6 Summary and contribution 20
3. Research Method 22 3.1 Hypotheses 22 3.2 Research Models 23 3.3 Data selection 25 3.4 Descriptive statistics 27 4. Results 29
4.1 Market Valuation Model 29
5. Summary & Conclusion 36
References 38
Abstract
Dutch
Dit onderzoek vergelijkt twee verschillende methodes waarop men moet omgaan met goodwill, in het bijzonder met betrekking tot het effect op de externe verslaggeving en de kwaliteit hiervan. Binnen de context van één van de kwalitatieve eigenschappen van financiële verslaggeving, zoals door de International Accounting Standards Board (IASB) gedefinieerd, wordt een model gebaseerd op waarde relevantie gebruikt om de effecten van de nieuwe impairment methode te onderzoeken. De dataset bestaat uit een aantal Nederlandse, beursgenoteerde bedrijven welke hun methoden hebben aangepast naar aanleiding van de introductie van de International Financial Reporting Standards (IFRS) in 2005. De resultaten lijken erop te wijzen dat de impairment methode minder waarderelevantie omvat dan de eerdere methode van afschrijving van goodwill. De verminderde waarderelevantie neemt verder af in tijden van crisis. Het lijkt erop dat de doelstelling van de IASB bij het uitbrengen van de nieuwe verslaggevingstandaard niet volledig is behaald. De implicatie van dit onderzoek is dat de IASB de impairment methode ten behoeve van goodwill zou moeten heroverwegen. De methode lijkt niet te leiden tot een hogere kwaliteit van financiële verslaggeving.
1. Introduction
The improvement of the quality of financial reporting is one of the objectives of the International Accounting Standards Board (IASB) issuing new accounting standards. IFRS 3 (IASB 2005, 2008), which became effective in 2005, is an example hereof. IFRS 3 radically changed the way in which companies have to account for goodwill. With the introduction of IFRS 3 it was no longer possible to amortise goodwill whereas companies had to, at least on an annual basis, measure the value of goodwill to see whether it needs to be impaired. The focus of this study is on this change in accounting treatment.
As one of the objectives of the IASB is to improve the quality of accounting information, it is interesting to test whether the change in accounting standard actually led to higher quality financial accounting. It is the goal of this study to find out whether the introduction of IFRS 3, impairment accounting, had the desired effect of increased quality of financial accounting. This empirical study is being performed in order to gain insights in relation to IASB’s formulated goals.
This study is based on value relevance, which is one of the qualitative characteristics of accounting information and uses a ‘value relevance model’ to find an answer to the research question(s). The study compares the periods 2001-2004, in which goodwill was amortized, and 2005-2012, in which goodwill had to be impaired and uses a selection of companies which are listed on the Dutch stock exchange and as a consequence applied accounting standard IFRS 3 since 2005. As the latter period contains data which is most recent data available this period was chosen.
The results from this study indicate less value relevant for the impairment method of goodwill accounting compared to the amortisation method. Hence, investors seem to perceive the impairment expense as less relevant information. The results of this study indicate that the introduction of IFRS 3 does not fully support the goals set by the IASB with the introduction of new accounting standards. Even though only one of IASB’s qualitative characteristics of accounting information is studied, it appears not all IASB objectives have been met. The results of this study imply that, in order to enhance the value relevance of goodwill impairments, the IASB might need to reconsider the content of IFRS 3.
The remainder of this study is organized as follows. Chapter 2 will provide the reader with a literature review which specifically addresses the qualitative
characteristics of accounting information. It further provides background on both methods of goodwill accounting and some other academic literature that is relevant for this study. The research method used to address the research question is introduced in chapter 3, whereas within chapter 4 the results of the models are presented. Chapter 5 ends this thesis with a summary and conclusion.
2. Literature review
This chapter will provide an overview of some of the extant literature on goodwill accounting. Due to its importance to this study some of the characteristics accounting information quality will be discussed. The different methods of goodwill accounting will be dealt with before the legislation surrounding this topic is touched upon. Subsequently, the sections thereafter focus on the methods of amortisation and impairment testing respectively. The last section of this chapter consists of a summary of the main findings of this study and the area in which it contributes to the existing literature.
2.1 Quality of accounting information
The ‘Framework for the Preparation and Presentation of Financial Statements’ was introduced by the International Accounting Standards Board (IASB) during 1989. It was eventually adopted in April of 2001 (IASB, 1989). The framework’s intent was to provide guidance for the preparation of accounting standards. During September 2010 the IASB approved an updated version of the framework, ‘Conceptual Framework for Financial Reporting 2010’ (the IFRS Framework). The conceptual framework lists a number of qualitative characteristics, which are the attributes that make the information provided in financial statements useful to users.
The four principal qualitative characteristics are understandability, relevance, reliability and comparability. Two fundamental qualitative characteristics of useful information listed within the framework are relevance and faithful representation (IASB, 2010). The IASB states that useful financial information must be relevant and a faithful representation what it purports to represent. They further state that items like comparability, verifiability, timeliness and understandability usually enhance the usefulness of financial information (IASB, 2010).
In order to be useful, information must be relevant to the decision-making needs of users. Information has the quality of relevance when it influences the economic decisions of users by helping them evaluate past, present or future events or confirming, or correcting, their past evaluations. In case it is capable of making a difference in the decisions users make information is deemed to be relevant. Based on it having predictive value, confirmatory value or both, financial information has the capability of making a difference in the decision-making process. Predictive value of
information is the degree in which it can be used as an input to certain processes predicting future outcomes. Financial information need not be a prediction or forecast to have predictive value. Users employ financial information comprising predictive value in their predictions. In case financial information provides feedback about (confirms or changes) previous evaluations it deems to have confirmatory value (IASB, 2010, p. 17). It is the relevance of accounting information this study focuses on.
According to Hitz (2007) the IASB puts more weight on fair-value based measures than on historical cost-based measures:
“The International Accounting Standards Board (IASB) has switched from historical cost-based measures to fair value-based measures with the intention to increase the decision usefulness of accounting information” (Hitz, 2007).
The IASB has favored the use of fair value measures for non-financial assets where market-based measures are derived from managements’ own expectations of future cash flows. It is this change from historical to fair value-based measures and its relevance to decision makers which will be the focus of the research. The qualitative characteristic relevance is often used within studies relating to the concept of value
relevance of (certain components of the) financial statements. Value relevance is all
about capturing the relation between (components of the) financial statements and certain market information (e.g. share price). More specifically, it measures the extent to which share prices have captured specific parts of information. By measuring value relevance users can determine if certain information is deemed to be of interest in calculating the value of a company. Based on the degree to which information is used for decision making purposes it can be determined whether said information is deemed to be relevant.
With respect to the quality of accounting information Siegel (1982) cited:
“Quality appears to be an important attribute of accounting information. However, accounting quality is neither a readily measurable nor a generally agreed upon characteristic of a firm.”
Accounting information quality depends on the usefulness of this information to outside parties, such as shareholders and investors and can be measured in a number of ways. Since relevance is considered to be a fundamental characteristic of accounting information I will examine the relevance of certain accounting
information. As the different paragraphs of this chapter will show, value relevance is just one way to measure accounting information usefulness and/or quality.
The two methods of accounting for goodwill will be the main focus of this study: amortisation method and impairment method. I will be using accounting information relevance to investigate which of the two methods result in more useful accounting information. Because of the mandatory introduction of IFRS in 2005 and the corresponding change in accounting regulation, goodwill is an interesting section of accounting information to involve in research. The next paragraph will be devoted to goodwill and the different methods to account for it in the financial statements.
2.2 Goodwill Accounting
The cost of the business combination exceeding the fair value of the net identifiable assets acquired in a takeover is usually called goodwill. From an accounting perspective, goodwill is thus part of the cost of a business combination and consists of the premium an entity pays on top of the book value of the entity at the moment it is acquired. The definition of goodwill applied by the IASB is as follows:
“An asset representing the future economic benefits arising from other assets
acquired in a business combination that are not individually identified and separately recognized.” (IASB, 2005, app. A,).
The IASB argues that goodwill satisfies the general definition of an asset and thus can be capitalized. However, there has been a long-lasting discussion with respect to the recognition of goodwill and the subsequent accounting treatment thereof.
Ample academic literature is available which investigated the question whether goodwill can be seen as an asset.
“Johnson and Petrone (1998) come to the conclusion that goodwill meets the definition of an asset. However, they state that in order to recognize it as an asset it should be measurable, reliable and relevant as well” (Van Hulzen et al., 2011).
Whether goodwill is perceived to be an asset can be measured through the research of the relation between a firm’s value and its capitalized goodwill. In case investors use goodwill in their valuations, which can be tested by performing a regression analysis on firm value and capitalized goodwill, they perceive goodwill to be an asset useful for the determination of a firm’s value. A positive relation between equity values and goodwill is found by Jennings et al. (1996), indicating investors perceive goodwill as
a valuable economic resource which creates future economic benefits and thus represents economic value. This justifies recognition (Van Hulzen et al., 2011).
More evidence on a relation between goodwill and equity value is provided by Chen et al. (2004) and Churyk and Chewning (2003).
Evidence of increased value relevance as the result of adopting SFAS 142 in 2002 is provided by Chen et al. (2004). They find that more value-relevant information is provided by reported goodwill after adjusting for impairments. Churyk and Chewning (2003) also found positive relations between equity values and recognized goodwill (Van Hulzen et al., 2011).
According to Spacek (1964), capitalization and subsequent amortisation are arbitrary and understate net income. Therefore a better treatment would be to write off goodwill immediately against retained earnings. More arguments are being brought forward for this treatment, such arguments emphasize the difficulty of measuring goodwill and thus suggest to expense goodwill immediately at the date of acquisition. Proponents of this treatment argue that goodwill would not meet the definition of an asset as future economic benefits, if there are any, cannot be reliably measured. However, it seems that goodwill is perceived to be an asset which needs to be recognized on the balance sheet by most accounting regulation bodies and investors.
Capitalizing and subsequently amortizing goodwill was applied by many companies and has been allowed by most accounting standard boards for a long period. Applying this method will lead to a fixed portion of expenses being charged to the income statement every reporting period over the estimated useful lifetime, with a maximum lifetime set at forty years. According to many critics this method is not appropriate as it is based on the assumption that goodwill decreases equally over time. Amortizing does not represent fairly the real decrease in the underlying economic value. In response to these critics, the accounting standards boards have implemented a different goodwill accounting method: the impairment method.
This study aims to investigate whether the amended goodwill accounting standard is related to an increase in accounting quality. The following paragraph will elaborate on the current law and regulation with respect to goodwill accounting.
2.3 Regulation on goodwill accounting
Standard setters are often targeted by industry and other affected parties to move them to prevent the imposing of an objectionable requirement (Zeff, 2002). In
this respect, the IASB issued accounting standard IFRS 3 Business Combinations in 2004, which became effective at the start of 2005. Most important part of the standard was the revised treatment of goodwill accounting.
With respect to subsequent measurement IFRS 3 states that the acquirer shall measure goodwill acquired in a business combination after initial recognition at cost less any accumulated impairment losses” (IASB, 2005, Par.54). As such, IFRS 3 prohibits the amortisation of goodwill. The IASB argued that the useful life of goodwill cannot be measured reliably. Moreover, the pattern in which it diminishes is often impossible to predict (IASB, 2005, BC 140). Due to the aforementioned arguments, the method of amortisation results in arbitrary amounts. Instead, the Board now requires performing an annual impairment test. The details of such test are outlined in IFRS 3, paragraph 55, which refers to IAS 36 Impairment of Assets. The purpose of an impairment test is to compare the amount of goodwill recognized on the balance sheet with the underlying economic value.
No such thing of a general consensus seems to exist within the academic literature regarding subsequent measurement. “Bugeja and Gallery (2006) did a study with respect to the relevance of goodwill over the course of time. They show that goodwill is perceived to be relevant in case it is acquired in the past two years only. Hence keeping goodwill on the balance sheet longer than two years does not provide information that is relevant. The aforementioned shows that a need seems to exist for an accounting treatment for goodwill that deals with this supposed declining
relevance seems to exist” (Van Hulzen et al., 2011).
Some even argue that the amount of goodwill that was initially recognized should not be revalued. This school of thought bases their arguments on the major points stated by Zeff and Dharan (1994):
It is over-conservative to write goodwill off the books when it has not depreciated in value below the purchase price;
When goodwill has actually depreciated, it is not necessary to record that depreciation in the operating account;
It is impossible to determine accurately the extent to which the goodwill has depreciated.
Bugeja and Gallery (2006) state that capitalized goodwill should be expensed over time as it appears to lose its value after a certain period.
As part of a project performed together with the FASB the IASB amended IFRS 3 in 2008. The revised IFRS 3 became effective as of July 1, 2009. Paragraph 54 of the revised standard does not explicitly refer to goodwill anymore; it prescribes that all assets and liabilities, acquired in a business combination, are subsequently accounted for in a manner in accordance with other standards on the specific item (IASB, 2008, Par.54). Despite the many differences that still exist between both standards; the general concept converged. The amendments to the revised standard do not substantially change the way goodwill is accounted for. IAS 36 Impairment of
Assets and IAS 38 Intangible Assets contain the rules with respect to goodwill
recognition and its measurement in subsequent periods.
Van Hulzen et al. (2011) elaborate on the reason of existence of the impairment method. They state that goodwill is considered to be an intangible asset falling under the scope of IAS 38 Intangible Assets. This standard provides for the regulation on the recognition of the asset and its subsequent measurement.
“IAS 38, paragraph 89 prescribes the making of a distinction between intangible assets with a finite and indefinite useful life” (IASB, 2004b, Par. 89).
This distinction is an important one as the method of subsequent measurement, which has to be followed, is determined by it. As the IASB considers estimating the useful life of goodwill to be very difficult goodwill is considered and treated as an asset with an indefinite useful life. Amortisation of goodwill is explicitly prohibited following IAS 36 Impairment of Assets, paragraphs 107 and 108 which prescribe application of
the impairment method (IASB, 2004b, Par. 107/108)”.
Further, it is IAS 36 detailing the moment an impairment needs to be made.
IAS 36 Impairment of Assets is all about the regulation on the impairment test (IASB, 2004a).
In the situation that the recoverable amount of the asset is below the carrying amount on a company’s balance sheet, an impairment needs to be made. Generally, whenever a company has reasons to assume that there is an indication of an impairment of the asset a company, it needs to check for a possible impairment loss. However, companies have to, at least, test for possible impairments each year (IASB, 2004a, Par.10b). Research has been undertaken to look into the concept of impairment tests. Carlin and Finch (2009) find that Australian firms tend to use lower discount rates in conducting goodwill impairment tests and conclude that management chooses the
discount rate opportunistically. However, they use the Capital Asset Pricing Model (CAPM) for the calculation of the discount rate and their findings critically depend on the validity of CAPM for calculating the discount-rate. Other scholars (Gallery, 2009) dispute the validity of CAPM in goodwill impairment testing. The CAPM can be used as a starting point which needs to be adjusted subsequently by the specific risks associated with the asset’s (or the cash generating unit’s) cash flows.
The recoverable amount of goodwill is determined as the higher of the ‘fair value less costs to sell’ and the ‘value in use’. The fair value less the cost represents the fair value decreased by the incurred expenses which the company would have to absorb in order for it to sell it. The present value of future cash flows represents value in use. No impairment is required to be taken into consideration in case one of these amounts exceeds the carrying amount of the asset. Determining the recoverable amount can be quite difficult. In case of a lack of an active market or no similar asset exists it might be impossible to determine the fair value. In cases where the assets themselves do not generate cash flows it will be impossible to determine the value in use.
The paper of Wines et al. (2007) shows it is not always an easy task to determine the recoverable amount. They find that the identification and valuation of cash-generating units and goodwill require numerous assumptions to be made in estimating fair value, value in use and recoverable amount. Considerable ambiguity and subjectivity are inherent in the IFRS requirements. IAS 36, paragraph 66 provides guidance for situations in which an entity experiences difficulties determining the recoverable amount of an asset. If this applies the company should determine the recoverable amount for the smallest cash-generating unit (CGU1) to which the asset belongs (IASB, 2004a, Par. 66).
Carlin and Finch (2007) have conducted a study of the largest New Zealand listed firms making voluntary disclosures relating to goodwill impairment testing under IAS 36. They find that the transition to IFRS has led to increased complexity in the nature of disclosures required in relation to goodwill and its impairment. IAS 36 prescribes the requirements for disclosure for a company with respect to their
1
A CGU is the smallest identifiable group of assets that generates cash inflows from continuing use that are largely independent of the cash inflows from other assets or groups of assets (IASB, 2004a, Par. 6).
financial reporting regarding the impairment test. A number of items which need to be addressed are:
1. the amount of the impairment; 2. the event that led to the impairment
3. the use of value in use or fair value less costs to sell as the measure of the recoverable amount;
4. information on the calculation of either method;
5. the class of assets to which the impairment is related in the case of CGUs (IASB, 2004a, Par. 130).
As the management of a company can apply a lot of discretion within the context of the impairment method, these disclosures are of particular interest to investors:
1. discount rates; 2. future cash flows;
3. the events that have possibly led to a lower recoverable amount should be included in the disclosures.
By means of the mandatory disclosure of this information in the financial statements, users of the reports will be able to review the discretionary choices, assumptions and estimates that were made by management to determine the recoverable amount and consequently the impairment expense.
The discretionary power that management of a company now has with respect to their assumptions and estimates when determining the need for an impairment makes the method of impairment of goodwill an interesting topic for research. Investigation of the change in accounting method and whether it has led to better quality of accounting information is interesting. Goodwill amortisation and goodwill impairment will be subjects of research within this study. The following two paragraphs outline previous research performed in relation to both methods of goodwill accounting.
2.4 Goodwill amortisation
Numerous academic articles have been written on the accounting for goodwill. However, said literature is predominantly on research with respect to the standard issued by the FASB, SFAS 142. It appears that research on IFRS 3 is much less present.
Previous research on SFAS 142 will be used as well, despite the differences in the FASB and the IASB standards on goodwill accounting. The differences between both standards seem to be minimal and are not specifically relevant for the research questions addressed in this study. It is this specific area where the current study contributes to the existing accounting literature. Examination of the impact of the new accounting method introduced by the IASB is facilitated by comparing findings from US-based research since regulation on the amortisation and impairment method was generally the same. One of the arguments of the IASB for switching to the method of impairment testing is the conviction that the method of amortisation leads to a situation of arbitrary accounting (IASB, 2005, BC 140). The relevance of amortisation has been investigated by scholars in several ways.
Jennings et al. (1996) came to the conclusion the amortisation of goodwill is relevant information to investors. They have investigated the relation between goodwill and equity. They conclude that there is a negative relation between goodwill amortisation and equity value. It seems that goodwill amortisation is perceived to be relevant accounting information. Since the evidence on the results is somewhat weak goodwill may not always decline in value for all companies or not at the pace expected by its economic lifetime. Although this is a limitation of the study it does show that information regarding goodwill amortisation of goodwill is relevant to investors.
Moehrle et al. (2001) performed a study with respect to the information content of goodwill accounting numbers. Their study relates to the plans of the FASB to change its previous Exposure Draft Business Combinations and Intangible Assets (Moehrle et al, 2001). Within said Exposure Draft it was suggested differentiating between different earnings numbers. A model is used to compare the explanatory power of the different types of earnings numbers. The authors find no significant difference in the explanatory power of the models with and without goodwill amortisation. This finding does not change for a sample of firms with material goodwill amortisation charges. As the explanatory power of the model with goodwill amortisation is not significantly lower the writers conclude that amortisation is not considered to be a source of noise. On the other hand, the inclusion of goodwill amortisation does not increase the explanatory power of the model either, based on which an inference can be drawn that goodwill amortisation is not perceived to be an item increasing explanatory power of the model.
Jennings et al. (2001) performed a research with respect to goodwill amortisation and its effect on the usefulness of earnings. Investors might prefer earnings without amortisation of goodwill in their valuation of share prices. Two valuation models are compared to investigate this. The explanatory power of the model excluding goodwill amortisation seems to have more explanatory power. The results are quite strong as they are significant for all of the years examined. Jennings et al. (2001) also come to the conclusion that amortization of goodwill is not relevant for investors and disturbs their share price valuations. They came to this conclusion by testing a model in which goodwill amortisation is added back, distinct from net income, to test if goodwill amortisation is relevant. Their results indicate this is not the case. In line with the findings of Moehrle et al. (2001) but contrary to the findings of Jennings et al. (1996), the study shows goodwill amortisation is not relevant to investors. Hence, it reinforces the decision to change the goodwill accounting standard.
The conviction of the IASB and the FASB that the amortisation method of goodwill is not a proper method to account for the diminishing value thereof is supported by a study performed by Li and Meeks (2006). Their study regards value relevance of goodwill in the UK, a market which is important regarding goodwill, as it seems to be the second largest takeover market. In line with the findings of Bugeja and Gallery (2006) they find that goodwill has value relevance but in a way that the relevance decreases over time. They also come to the conclusion that the method of amortization is not a proper method to account for the diminishing value of goodwill as it has no value relevance.
Another study in the field of value relevance of goodwill amortization is performed by Churyk and Chewning Jr. (2003). Their sample consists of companies listed on stock exchanges in the United States of America. They find that goodwill amortisation is negatively associated with equity values. The market perceives goodwill as an asset which decreases in value over time. Their findings are in line with the findings of Bugeja and Gallery (2006). Churyk and Chewning Jr. (2003) state that FASB Standard 142 is built on invalid arguments. According to Churyk and Chewning Jr. (2003) goodwill should be amortized over its useful economic life and thus treated like any other economic asset. According to IAS 36 Impairment of Assets companies may have to record an impairment without an impairment event taking place. Goodwill should be reviewed for a necessary impairment annually. This
automatically implies that, according to the IASB, the arguments brought forward by Churyk and Chewning Jr. (2003) are not applicable to IFRS 3. A decrease of the value of capitalized goodwill is recognized, at least annually, in the form of an impairment expense.
Overall, there seems to be mixed evidence regarding the relevance of goodwill amortisation. Despite the mixed evidence the IASB and FASB took the decision to amend the accounting standards relating to goodwill accounting. The following paragraph sets out academic research with respect to the impairment method.
2.5 Impairment of goodwill
The accounting method with respect to goodwill impairment is only effective for a limited number of years. This paragraph outlines some of the existing literature which deals with the impairment of goodwill and some of the studies that compare the two methods.
Since this study looks into the relevance of both the amortisation method and the impairment method of goodwill accounting it is interesting to look at prior studies that investigated either or both concepts of goodwill accounting. Saastamoinen et al. (2013) examined perceptions of financial analysts of the accounting for goodwill under IFRS and more specifically the concept of goodwill impairment testing. During late 2012 and early 2013 the authors gathered data from financial analysts in Nordic countries and Austria by means of a survey questionnaire. Their most important finding suggests that there are two separate schools of thought among the population of financial analysts. First school of thought is having a trusting attitude towards IFRS, whereas the second school exhibits a critical view on the role of management, and their estimates, in the whole process of goodwill accounting. Saastamoinen et al. (2013) have also listed some of the effects surrounding the introduction of the impairment method of goodwill accounting:
“Tenured management has been detected to record less write-offs (Beatty and Weber 2006; Ramanna and Watts 2009; Hamberg et al. 2011) and correspondingly, recent CEO changes are connected to goodwill impairment losses (AbuGhazaleh et al. 2011; Saastamoinen and Pajunen 2012). Following the same pattern, goodwill impairment charges are larger in the companies where the tenure of the CEO is shorter (Masters-Stout et al. 2008). Furthermore, executive compensation incentives influence impairment
decisions (Beatty and Weber 2006; Watts 2009). Carlin and Finch (2010) find that companies apply low discount rates to avoid goodwill impairment losses. The earnings bath type of behavior has also been linked to goodwill write-down decisions (Walsh et al. 1991; Jordan and Clark 2004, 2007; Sevin and Schroeder 2005; AbuGhazaleh et al. 2011; Saastamoinen and Pajunen 2012).”
By using a value relevance model Chambers (2006) examines the effects of SFAS 142 on financial reporting. Within this study the relevance of impairments and the effects of the elimination of goodwill amortization are studied. In line with the goals of the FASB, financial reporting quality improved by the use of impairments. However, Chambers (2006) concluded that the elimination of the amortisation method resulted in lower quality reporting. Chambers (2006) created some sort of fictional accounting system, which includes both the amortization as the impairment method. The latter system resulted in accounting numbers which seemed to be the most relevant. Chambers (2006) concluded that, consistent with the goals of the FASB, the introduction of SFAS 142 has led to higher accounting quality.
A similar research is performed by Chen et al. (2004). In line with Chambers (2006), Chen et al. (2004) find that value relevance increases after the adoption of SFAS 142. Hence, Chen et al. (2004) come to the conclusion that with the issuance of SFAS 142 the objectives of the FASB have been met. Their study makes a distinction between the one-time impairment at the moment of adoption and the subsequent yearly impairments. This impairment at the moment of adoption is due to the difference between the carrying amount at the moment of adoption and after all historical amortisations over the past years and the recoverable amount. It could be the case that the change in accounting standard has led to an extra impairment in the year following the introduction. The authors also investigated the timeliness effects of the impairment. They conclude that share prices already partly incorporated the initial impairment and the impairment in the first year. The authors found that the accounting information is pretty timely as they found a delay of twelve months between the economic decline in value of the goodwill and the recognition of the impairment in the financial statements.
Lapointe et al. (2009) also examine an initial impairment after which they conclude that the retroactive method is in line with the FASB goals with respect to accounting information quality. Application of the retroactive method requires
companies adjusting retained earnings in their opening balances for the impairment amount. This is contrary to the cumulative method, in which the initial impairment is reflected in the income statement. IFRS 3 requires application of the retroactive method. Lapointe et al. (2009) investigate the timeliness and value relevance of these initial impairments in a Canadian setting. Their findings are in line with Chen et al. (2004).
Churyk (2005) provides more evidence on the effects of SFAS 142. By comparing market valuation of goodwill she examines whether the elimination of the amortisation is appropriate. One of the findings of Churyk is that goodwill at the moment of acquisition is only occasionally overvalued which is why she argues that systematic amortisation is not required. However, she does find certain justifications for goodwill impairment (Churyk, 2005). She concludes that it is appropriate to eliminate the method of amortization and that the impairment method is justified.
Li and Meeks (2006) investigated the relevance of impairments in the UK. Companies were allowed to use the impairment method between 1998 and 2002 It appears that in 2002, the final year in the sample, impairments were most common. Probably companies were more familiar with the concept of impairment testing in that year. Additionally, an economic downturn struck us during that year. Impairments are perceived to be relevant, contrary to amortisations. As a result of the large coefficient on impairments the authors interpret this as a potential overreaction to bad news or a signal to the market of decreased future earnings. Despite the very convincing results, due to its limited sample size, only one year, this study has a serious limitation. Iatridis et al. (2006) used a sample of firm-years to investigate the timeliness of impairments in the UK. The question whether goodwill impairments are associated with decreasing market value in the year preceding or in the year of the impairment itself is investigated. The results show that impairments are timely. The authors find a relation between the negative stock return in the same or preceding year and the impairment which seems to be significant. A limitation of the study is the potential sample selection bias. An impairment was reported by all firms in the sample, however it is not certain all firms which should have reported an impairment did so. Companies that did not report impairments in a timely manner were left out of the sample, due to the limited number of sample years (2000 and 2001). This bias in the sample selection potentially affected the findings.
The European Union introduced the requirement for companies to report in accordance with IFRS as of 2005. Since companies were applying these standards for the first time during this year lots of opportunities arose for academic research to investigate the effects of IFRS 3. However, the number of such studies is limited due to the limited period since adoption of IFRS. This study contributes to academic literature, because it not only extends the existing research with respect to the effects of IFRS 3 it also endeavors to enlighten decision makers with the provision of specific ex post empirical evidence (Barth, 2006).
A study performed by Barksjö and Paananen (2006) investigate the value relevance and timeliness of the financial statement information before and after the implementation of IFRS. They compare impairment expenses with amortisation expenses, but they also make a distinction between companies with relative high and low amounts of intangible assets. The authors state that previous research on goodwill has proved the method of amortisation does not always result in financial statements reflecting the underlying economic value of goodwill. Hence, companies with substantial amounts of intangible assets will benefit from the change to the impairment method of accounting. This benefit will be in the form of experiencing increased value relevance. They conclude that only this hypothesis is supported. All other hypotheses are rejected however. According to the findings of this study, the goal of higher accounting quality was not achieved by the introduction of IFRS. The study has a limitation as only the years 2004 and 2005 are selected within the sample, comparing both methods of accounting. A number of additional years could lead to more reliable evidence on the comparison between the two methods of accounting for goodwill as it is possible that because firms were not yet very familiar with the impairment method of accounting, investors perceive the information disclosed as less reliable because of (possible) measurement errors.
Within the same setting as Barksjö and Paananen (2006), namely the Swedish stock market, Hamberg et al. (2006) examined the effects of the adoption of IFRS. Evidence is found of goodwill having a having a higher degree of persistence under the impairment method. As impairment expenses do not per se occur every year this finding is consistent with the expectations. Hence, goodwill persists longer on the balance sheet. According to Hamberg et al. (2006) the introduction of IFRS has led to accounting information being more relevant. However, it could be due to a sample bias as the sample selection is limited. It could well have been possible that
impairments were not very common in the sample year(s) and this potentially had an effect on the findings. Moreover, Hamberg et al. (2006) investigate the nature of relevance of goodwill by buying stocks with relative high amortisation costs and selling stocks with relative low amortisation costs. In case investors did not incorporate the changes already in prices abnormal returns could have been expected because of the longer persistence of goodwill under the impairment method of accounting. The findings show this, indeed, is the case. Despite the results not being significant, investors seemed to perceive the higher persistence of goodwill as new information.
2.6 Summary and contribution
Both standard setting bodies changed regulations on goodwill accounting in the past years. They both abolished the option to amortise goodwill. Companies now have to perform an annual impairment test on goodwill. They need to ascertain their selves that the recoverable amount of the goodwill is higher than the carrying amount. In case it is the other way round, then goodwill has to be impaired.
Since then academic research has investigated both methods of goodwill accounting. Despite the extant research no consensus has been reached on their effect on accounting quality. The amortisation method leads to arbitrary accounting (IASB, 2005, BC 140), which implies that, after having changed to the impairment method, the impairment method does not. However, the latter statement is not in full supported by academic evidence; since there is no irrefutable evidence that accounting quality has improved.
It is the goal of the IASB with the introduction of IFRS 3 to come to higher quality of accounting information. This study examines whether the introduction of IFRS 3 and the subsequent change in accounting method for goodwill has achieved this goal. This is investigated by testing the (value) relevance of the information regarding the treatment of goodwill. Relevance is identified by the FASB and IASB as an important qualitative characteristic of accounting information. This study contributes to academic literature for a number of reasons. First of all, both the amortization as the impairment method is compared by virtue of the introduction of IFRS 3 by the IASB, whereas most academic research has focused on regulation by the FASB. Second, previous studies regarding IFRS 3 used specific local stock markets as dataset. I will use a dataset from the Netherlands, discussed in section 3.3.
The findings from the data from the Netherlands may well complement findings from previous studies. Finally, previous studies were executed just after adoption of the new standard. As more data is available in the dataset this study has the potential for stronger findings. The research method applied within this study is discussed in the next section.
3. Research Method
This chapter will introduce the research method that has been used to examine the research question. First, the hypotheses that are tested will be described. This description will be based on previous accounting literature in conjunction with the goals of the IASB in their instructions regarding a different accounting method for goodwill. Hereafter I will explain the model that was used to test the hypothesis. Subsequently, I will explain why this model is used and why it is relevant for examining the research questions. In the last section of this chapter the data selection process will be discussed.
3.1 Hypotheses
As mentioned in the literature review, the improvement of quality of accounting information and enhancement of its usefulness for decision making is an important reason for the IASB to issue new accounting standards. The introduction of IFRS 3 and the accompanying amendments to IAS 36 and IAS 38 are examples hereof. The goal of this research is to investigate whether the change in goodwill accounting has led to the anticipated improvement in the quality of accounting information, all in the context of the specific sample companies. The sample consists of almost 50 companies listed on the ‘NYSE Euronext Amsterdam’ (Appendix A). Section 3.3 will further elaborate on the data selection process
One of the common methods within the academic literature to test the quality of accounting information is the concept of value relevance. The characteristics of accounting information, as captured in the framework developed by the IASB and FASB, are the fundamentals of this method of testing for value relevance (IASB/FASB, 2008). A number of other scholars have conducted research with respect to goodwill accounting (for example Chambers 2006; Lapointe et al. 2009; Chen et al. 2004; Iatridis et al. 2006 and Barksjö and Paananen 2006) uses, amongst others, this property to test the effects of goodwill accounting standards. Within the context of this study, I will use value relevance only to test the hypotheses.
The impairment method of goodwill accounting has been introduced by the IASB and has been motivated by a conviction that the previous method of goodwill amortisation will lead to arbitrary accounting measures (IASB, 2005, BC 140). As the information reported in the situation of goodwill amortisation will, by definition, not
reflect the underlying financial position of the firm it is deemed to be irrelevant in terms of the decision making process. One of the things that may be expected from new accounting standards is that they serve the goals of the IASB. One of these goals is to make accounting information more relevant.
The relevance of goodwill accounting numbers are expected to increase following the introduction of IFRS 3 and the amendments to IAS 36 and IAS 38. One may expect the goodwill impairment expense to be more relevant to the users of this information than the amortisation expense. Lapointe et al. (2009), Chambers (2006) and Chen et al. (2004) provide evidence for this statement regarding SFAS 142. Barksjö and Paananen (2006) cannot find evidence for the statement. They have examined the effects of IFRS in Sweden. The first hypothesis examined in this study is:
H1: The goodwill impairment expense value relevance is higher than the value relevance of goodwill amortisation expense.
Impairments are expected to occur more frequently in periods of economic downturn. The most recent economic crisis started in 2008 and lasted until the year 2013. This might have an impact on the results from our analysis. Hence, I will examine whether the relation between goodwill impairment expense and the value of the sample firms during the years 2008-2013 differs from the pre-crisis years. Also, I will compare the ‘crisis years’ (2008-2013) with the period when goodwill was still to be amortized. The second hypothesis examined in this study is therefore:
H2: The value relevance of goodwill impairment expense during crisis years is higher than the value relevance of goodwill impairment expense during ‘non-crisis years’ and/or goodwill amortisation expense.
3.2 Research Model
The hypotheses will be tested using a model often used in academic literature. In order to examine the relevance of accounting information I use a market valuation model, based Ohlson’s (1995) work.
The relation between market information and accounting information has earlier been described as the defition of value relevance. Accounting information is deemed to be value relevant when investors use this information in the valuation of a specific company and thus in their decision making process. It was Ohlson (1995) who developed a market valuation model that related the market value of a company with accounting information. The model describes the market value of the company as a function of book value of equity and net income.
with MVEit being the market value, BVEit being book value of equity and NIit being net income. The explanatory power of the model approximates the value relevance. In case the accounting information is useful, one may expect a strong relation with the market value of the company. This in its turn will lead to high explanatory power of the model. In case accounting information is not useful, the relation with the market value will be weak and the explanatory power of the model will be low.
One of the major advantages of using the model developed by Ohlson (1995) is the ease with which additional variables can be added to the equation. As a consequence of the aforementioned, two models and two explanatory powers can be compared and assessed. Based on this comparison a conclusion can be drawn regarding which accounting information is deemed to be more value relevant. The Ohlson-model (1995) is very appropriate for the testing of the research questions as defined in the previous paragraph as it is able to compare the value relevance of a model containing an amortisation variable with a model containing an impairment variable. In order to test the first hypothesis, the basic Ohlson-model (1995) with some necessary adjustments is used. To compare and assess the value relevance of goodwill impairments against goodwill amortisation, one needs two models, each of them containing one of the two variables. The first equation, being an extension of the traditional Ohlson-model (1995) is set out below:
(1)
where AMORTit represents the annual amount of goodwill amortisation and all other variables are as previously defined. Based on the assumption that financial markets
it it it it BVE NI MVE 0 1 2 it it it it it BVE NI AMORT MVE 0 1 2 3
are efficient (in the strong form) the value of the company will be measured at year-end. In order to control for size differences all variables will be deflated by the number of shares outstanding. It is expected that the extended model has a higher R square (degree of explanatory power) in case investors view the amortisation variable as relevant information. As already mentioned earlier, Jennings et al. (2001) find that, by adding a variable for the goodwill amortisation, the explanatory power decreases which indicates that it adds noise to the valuation. Moehrle et al. (2001) find that there is no significant difference between the equation with and without the goodwill amortisation expense.. The next section will explain the second equation which includes a variable for the impairment charge.
A comparison will be made between equation (1) and a model that contains an impairment variable:
(2) where IMPit is the amount of annual goodwill impairment and all other variables are as previously defined. Market value is measured at year-end and all variables will again be deflated by the number of outstanding shares. Equation (2) should have a higher degree of explanatory power than the basic Ohlson-model (1995) in case investors perceive impairment of goodwill as relevant information. Although they use a slightly different method, Barskjö and Paananen (2006) find no evidence for this, whereas Lapointe et al. (2009) find that adding the impairment variable, indeed, adds explanatory power compared to the equation without the impairment variable.
According to hypothesis the value relevance of goodwill impairments is higher than the value relevance of goodwill amortisation. The relevance will is measured using the explanatory power of equations (1) and (2). Following the hypothesis and the related critics regarding the arbitrary character of the amortisation of goodwill, it is expected that the value relevance of equation (2) is higher than that of equation (1). Depending on the significance of the variables in the model information on the magnitude of the variables AMORTit and IMPit could give more evidence on the comparison of their value relevance.
3.3 Data selection it it it it it BVE NI IMP MVE 0 1 2 3
In order to make a comparison between the two methods of treatment of goodwill, data from 2001-2004 is used for the amortisation method and from 2005-2012 for the impairment method. In order to test whether the results are different in times of financial downturn, when more impairments are expected to occur, the years 2008-2012 are isolated and compared with the amortisation method. One of the contributions of this thesis is the incremental sample size of eight years on the impairment method compared to other studies. It also contributes in a sense that it measures the effect of the change in accounting treatment during times of economic crisis, with an expected increased effect on the number of impairments. Although the periods for both accounting methods are of different length/duration, this does not imply, by definition, incomparability, because of different reasons. These reasons could possibly different economic conditions, the number of companies recording an amortisation or impairment expense on goodwill etc.
The financial database Datastream has been used to extract the data with which this study tries to examine the research question. The market index ‘Euronext’ is selected in order to extract the data. The database comprises equities from France, Portugal, Belgium and The Netherlands. Euronext contains two major equity indices: Euronext 100 and The Next 150 (Euronext). Jointly, data is available from the 250 largest companies listed at the participating markets. From this dataset a further selection is made in order to isolate the data required for the investigation of the companies traded on the Dutch market2. As this market index is large European equity index, including data for the Dutch, listed companies, I use it. Also, all of the companies included in the index are required to report under IFRS since 2005. However, a bias towards large companies exists in the sample selection because of the use of the largest index in Europe. This can also potentially affect the results and has to be taken into consideration in discussing them.
I have used the following variables in the dataset: book value of equity, earnings for the year, share price at year end and either the amortisation or impairment amount, depending on the year of the sample. All variables were divided by the number of shares outstanding in order to control for size differences. After having collected the data from the Datastream database, all observations with missing
2
data were manually updated in case data was available from the company websites. The next chapter will discuss the results of the empirical study.
3.4 Descriptive statistics
This section covers the implementation of the data sample as described in the chapter into Microsoft Excel. First the descriptive statistics of the data will be described in paragraph 3.4.1 Then the power and significance of the model will be described in paragraph 3.4.2
.
3.4.1. Descriptive statistics
The first part of this section will look into the descriptive statistics of the sample that was described in the previous section. These descriptive statistics will be used to gain insight in the data set that is used.
Tables 3.1 up till and including 3.3 provide an overview of the variables used in the regression. The table shows that the sample included 193 available firm years during the period prior to the introduction of IFRS on January 1, 2005. The “post-IFRS” sample included 392 firm years whereas the sample consisting of company crisis years consists of 245 firm years.
Table 3.1 Results Pre‐IFRS
Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
BV (per share) 193 0.08 69.41 9.77 0.06
EPS 193 -3.57 10.40 1.10 0.41
Amort Goodw (per share) 193 - 0.001320 0,00004 3,315.79
Table 3.2 Results Post‐IFRS
Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
BV (per share) 392 0.74 5,366.00 13.69 0.04
EPS 392 -9.14 16.53 1.31 0.23
Table 3.3 Results Post‐IFRS ‐ Crisis
Descriptive Statistics
N Minimum Maximum Mean Std. Deviation
BV (per share) 245 0.74 83.74 14.08 0.04
EPS 245 -9.14 5.39 0.78 0.33
4. Results
This chapter will present the results of the statistical analyses made on the dataset which was extracted from Datastream. The results of the statistical analyses will be discussed, interpreted, explained and compared with all of the hypotheses described in Chapter 3. Microsoft Excel and an add-in which is called AnalysisToolPak have been used to perform the statistical analysis. The first paragraph below presents the results of the market valuation model in which the value relevance of the two methods of goodwill accounting are compared. It further presents the results of the analysis between the amortization method and the impairment in times of financial crisis, as it is expected that impairments are more present in such period.
4.1 Market Valuation model
Hypothesis 1 is examined by the comparison between equations (1) and (2). Table 1 below contains the results with respect to equation (1).
Table 1
(1)
N Intercept BVE NI AMORT
2001 48 9.19 0.62 1.46 16.50 (p-value) <0.01 <0.01 0.10 0.20 (t statistic) 5.21 4.16 1.70 1.30 2002 48 5.46 0.59 1.26 0.00 (p-value) <0.01 <0.01 0.07 #NUM! (t statistic) 4.98 5.43 1.86 65535.00 2003 48 6.51 0.57 2.36 0.00 (p-value) <0.01 <0.01 0.03 #NUM! (t statistic) 6.46 4.99 2.18 65535.00 2004 49 4.30 0.56 4.52 8.36 (p-value) <0.01 <0.01 <0.01 <0.01 (t statistic) 5.46 7.50 7.13 3.64 Pooled 193 6.51 0.63 1.88 10898.60 (p-value) <0.01 <0.01 <0.01 <0.01 Adj. R² (t statistic) 10.34 11.00 4.59 3.29 0.732
For the years 2001-2004, there were 193 observations. The results on the variable BVEit are consistent and thus very conclusive.since allof the results are significant at the 1% level over all firm-years and as well in the pooled sample. This implies that, within the sample, information regarding the book value of equity is used to determine the value of the companies. However, this result is not surprising since the book value of equity represent the value of the company that is attributable to its equity holders.
The results on the variable NIit and the amortisation variable are less consistent as they only appear to be significant at the 1% level during the year 2004 and within the pooled sample. During the years 2001-2003 the results for NIit, with p-values ranging from 0.03 to 0.1, are somewhat significant, implying the relation between the variability in firm value and the variability in earnings to be less present compared to the relation with BVEit.
Interesting are the results on the amortisation variable. There does not seem to be a relation between the amortisation variable and firm value for the period 2001-2003. In 2004 and in the pooled sample the amortisation variable is significant, with p-values below the 1% level. This may have been caused by the fact that the treatment for accounting purposes of goodwill amortization expense changed between 2004 and 2005. Investors could potentially have been reevaluating their perspective on goodwill amortisation, because of the fact that the expense might not occur in the next year. Therefore investors might have scrutinized this figure. The coefficient on goodwill amortisation for the pooled sample is significant at the 1% level. This may have been caused by the scrutiny with which this number has been regarded during the year 2004.
It is at striking that the coefficients of the variable for amortisation are all positive. This would imply that an increase in expenses related to the amortisation of goodwill leads to an increase in value of the company. This assertion contradicts the idea that higher amortisation expenses diminish the value of a specific firm. Since the coefficients are mostly insignificant, no very strong conclusions can be drawn from this observation. The explanatory power of the model is the adjusted R² of 73.2% meaning that approximately seventy-three percent of the variance in share prices can be explained by the variation in the combination of the three variables in the model.
Table 2
(2)
N Intercept BVE NI IMP
2005 49 8.27 0.29 5.15 23504.22 (p-value) <0.01 0.153 <0.01 0.09 (t statistic) 5.61 1.45 4.52 1.76 2006 49 11.55 0.42 4.20 22958.08 (p-value) <0.01 0.123 <0.01 0.57 (t statistic) 5.76 1.57 3.05 0.57 2007 49 12.01 0.67 0.86 112920.14 (p-value) <0.01 <0.01 0.319206338 0.04 (t statistic) 7.00 4.82 1.01 2.10 2008 49 4.74 0.61 0.93 -2612.87 (p-value) <0.01 <0.01 0.025218587 0.02 (t statistic) 4.65 13.71 2.32 -2.47 2009 49 6.24 0.98 2.76 -1176.58 (p-value) <0.01 <0.01 <0.01 0.83 (t statistic) 4.10 11.97 4.27 -0.21 2010 49 5.77 0.28 8.98 -82846.02 (p-value) <0.01 0.015 <0.01 0.04 (t statistic) 3.55 2.53 6.63 -2.14 2011 49 6.88 0.36 4.70 3740.89 (p-value) <0.01 <0.01 <0.01 0.10 (t statistic) 4.19 4.01 5.69 1.70 2012 49 7.85 0.93 2.59 -2969.65 (p-value) <0.01 <0.01 <0.01 0.42 (t statistic) 2.78 6.48 3.17 -0.82 Pooled 392 8.97 0.68 2.30 -2411.26 (p-value) <0.01 <0.01 <0.01 0.07 Adj. R² (t statistic) 13.62 19.45 10.21 -1.82 0.676
The results on equation (2) are presented in Table 2. With the exception of the year 2010 BVEit is significant for at least at a 1% level for all the years in the sample as of 2007. This is in line with the results prior to the change in accounting method shown in table 1. The years 2005 and 2006 were those immediately after the change in the accounting method, which might be related to the insignificant results of our sample.
Contrary to the coefficients on the amortisation expense, the coefficients on the impairment expense aren’t all positive. They seem to be a mix of positive and negative numbers. As of the year 2008, which was the first year of the financial crisis, the coefficient is negative and with the exception of the year 2011 it remains negative up till 2012. The coefficient for the pooled sample is negative as well. Together with the (pooled) p-value of 0.07 it can be concluded that a somewhat significant negative effect can be attributed to the impairment amounts within the sample. This is in line with the general idea that increased impairment expenses leads to decreased equity values and stock values. Although it seems that there is inconsistency in the significance of the coefficients over the years, it can be concluded that, in general, the impairment expense is used by investors for their valuation of the company.
This is supported by the explanatory power of the model. Equation (2) has an R² of 67.6%. Compared with the explanatory power of 73.2% of equation (1), it can be concluded that the change from amortisation of goodwill to impairment of goodwill leads to a slightly decreased value relevance. There is a possibility however that the decrease in explanatory power is the result of a decrease in value relevance of one of the other variables in the equations, BVEit and NIit. With the goal of adding credibility to the conclusions, I’ve conducted an additional analysis (which is not incorporated in this thesis) by performing an additional regression without the amortisation or impairment variable. These analyses appear to show that decreased relevance is not per se caused by adding the impairment variable. The R² which was the result for the 2005-2012 sample decreased compared to the R² of the sample years 2001-2004. Hence, the decreasing explanatory power of the model can be caused by the impairment variable but there’s a definite effect within the value relevance of the other variables as well.
Hypothesis 1 states value relevance of goodwill impairment exceeds the value relevance of goodwill amortisation. Based on the decreased explanatory power of equation (2) compared to equation (1) and the significance of the amortisation
variable compared with the insignificance of the impairment variable, I conclude that this hypothesis is rejected. Goodwill amortisation are more value relevant than goodwill impairments.
By running a regression on equation (2) over the period 2008-2012 hypothesis 2 is examined. It states that the value relevance of goodwill impairment expense during crisis years is higher than the value relevance of goodwill impairment expense during ‘non-crisis years’ and/or goodwill amortisation expense. This expected increase in value relevance can be interpreted in several ways: an increase in explanatory power of the model for the impairment regression during this specific period, a movement in the amounts of the coefficients on certain variables for the period 2005-2007. In this subsection I will try to explain these interpretations and evaluate the results of the regression.
The results on equation (2) during the period 2008-2012 are presented in Table 3. Table 3 (2)
N Intercept BVE NI IMP
2008 49 4.74 0.61 0.93 -2612.87 (p-value) <0.01 <0.01 0.025218587 0.02 (t statistic) 4.65 13.71 2.32 -2.47 2009 49 6.24 0.98 2.76 -1176.58 (p-value) <0.01 <0.01 <0.01 0.83 (t statistic) 4.10 11.97 4.27 -0.21 2010 49 5.77 0.28 8.98 -82846.02 (p-value) <0.01 0.015 <0.01 0.04 (t statistic) 3.55 2.53 6.63 -2.14 2011 49 6.88 0.36 4.70 3740.89 (p-value) <0.01 <0.01 <0.01 0.10 (t statistic) 4.19 4.01 5.69 1.70 2012 49 7.85 0.93 2.59 -2969.65 (p-value) <0.01 <0.01 <0.01 0.42 (t statistic) 2.78 6.48 3.17 -0.82
Pooled 245 7.63 0.73 2.08 -2494.78 (p-value) <0.01 <0.01 <0.01 0.06 Adj. R² (t statistic) 8.64 17.70 6.40 -1.91 0.593
Obviously, the results per year are the same as presented in Table 2. However, the results for the sample on a pooled basis show differences. The coefficient for BVEit increased from 0.68 to 0.73 which implies an additional weight being attributed to this variable. The coefficient NIit decreased from 2.30 to 2.08 which may imply the contrary, less weight being attributed to the net income number. This could be caused by the number of firms that are affected by the financial downturn and investors relying less on the reported net income numbers when determining the value of the firm. The increase in the coefficient for BVEit might be related to the aforementioned situation in which investors seem to be taking other aspects into consideration when they are calculating a firm value. This statement is being reinforced by the decreased explanatory power of the model for the period 2008-2012, 59.3%.
Hypothesis 2 states that the value relevance of goodwill impairment expense during crisis years is higher than the value relevance of goodwill impairment expense during ‘non-crisis years’ and/or goodwill amortisation expense. Based on the decreased explanatory power of equation (2) for the period 2008-2012 compared to the periods 2001-2004 and 2005-2012 and the significance of the amortisation variable compared with the insignificance of the impairment variable, I conclude that this hypothesis is rejected. Therefore, goodwill impairments during the years 2008-2012 are less value relevant than goodwill impairments during the period 2005-2007 or the goodwill amortisation which were present during the period 2001-2004.
5. Summary and Conclusion
After the FASB (SFAS 142) the IASB (IFRS 3) has also changed its accounting standards with respect to the treatment of goodwill during past years. It is no longer accepted to amortise goodwill annually. Companies have to perform annual impairment tests now. In case the recoverable amount falls below the carrying amount on the balance sheet, the company has to recognize an impairment expense.
The increase of the quality of accounting information has been the objective of the change. It has been argued by the IASB that when information is of an increased quality and thus more useful for decision making purposes, investors will prefer this Relevance is pointed out as one of the fundamental characteristics in the framework that the IASB and FASB produced together..
This study investigated whether the aforementioned change in the accounting method for the treatment of goodwill has led to increased accounting quality. The study is a contribution to the current academic literature, since its main focus is on the introduction of IFRS 3 in 2005 as opposed to most previous research which examined the effects of SFAS 142. Further, this study has reviewed the effects of the introduction of IFRS 3 during an extended period (2005-2012) whereas previous research struggled with the problem of data availability over a longer time span. This study used an extended sample of eight years (2005-2012) and collected data for Dutch entities included in the biggest market index in Europe (Euronext). In doing so, the study was able to result in stronger evidence on the research question than research conducted in earlier years.
Based on this particular sample and the qualitative characteristic of relevance the results of this study do not show evidence of increased accounting quality. Overall, the amortisation expense of goodwill seems to be more relevant than the impairment expense on goodwill, implying that investors find the meanwhile abolished amortisation expense more useful for decision making purposes. Based on the aforementioned I conclude that the IASB objective has not completely been met. The results of this study imply that the IASB should reconsider the impairment method of goodwill accounting. For example, the IASB could issue clear guidelines and instructions of how and when to perform an impairment test. They can consider having the impairment test performed by certified auditors only in order to make it
