Opportunism in Goodwill Impairment Decisions : IAS 36 versus SFAS 142

(1)

Amsterdam Business School

Opportunism in Goodwill Impairment Decisions: IAS 36 versus

SFAS 142

Name: Laura F. Paanakker Student number: 11161884

Thesis supervisor: Georgios Georgakopoulos Date: 19 June 2016

Word count: 17097

MSc Accountancy & Control, specialization Accountancy Faculty of Economics and Business, University of Amsterdam

(2)

Statement of Originality

This document is written by student Laura Paanakker who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

Abstract

Although the topic has been discussed for many years prior to it, the Norwalk Agreement in 2002 put the rules-versus-principles debate high on the agenda of financial statement standard-setters, preparers and users and thereby also on those of academics. This thesis provides an outline of the existing literature on the debate that reasons about whether rules-based standards (RBSs) or principles-based standards (PBSs) should be preferable as a basis for financial reporting and strives to provide for a contribution to it with regard to goodwill impairment accounting. Through a cross-sectional empirical analysis, a sample of firm-years reporting in accordance with IFRS and a sample of firm-years reporting in accordance with U.S. GAAP are compared in terms of their goodwill impairment opportunism as a result of (1) the contracting motive, (2) the reputation motive, and (3) the valuation motive. The results reveal that, when measured in goodwill impairment recognition decisions, both samples show evidence of opportunism, but in a different degree when analyzed according to different motives. When assessed in goodwill impairment measurement decisions, only the valuation motive seems to be observed and only in the IFRS sample. Albeit no clear-cut conclusion can be retrieved from these findings, this thesis adds to the current knowledge by providing insight into the relative exploitation of certain agency theory-related motives among financial statement preparers under a RBS and a PBS.

(4)

1 Introduction

“[…] the Lord of the Rules started to speak. Gollum retreated further into the darkness. He started to mutter incessantly. ‘Rules or Principles, which are better? Do I want

Principles or Rules? […]‘Master,’ he said ruefully, ‘what is the difference between Rules and Principles?’ ‘It is quite simple. A principle will tell you generally what is the right thing to do and then you must judge what to do in specific circumstances. A rule will tell you what to do in specific circumstances.’ ‘I see,’ beamed Gollum, ‘principles are like a compass which show you the direction and rules are like maps which show you exactly how to get there.’ After a few moments

of reflection, he added, ‘I prefer maps.’ ‘It is not like that, Gollum. Maps can be wrong or if

you are in uncharted territory or if you have no map, you will need a compass. Also, you cannot make precise maps for every part of Middle Earth” (Ong & Hussey, 2005, p. 4)1_.

This quotation from Ong and Hussey (2005) describes in a nutshell what the principles-versus-rules controversy in financial reporting standards is about. With the Norwalk Agreement, signed by the International Accountings Standards Board (IASB) and the Financial Accounting Standards Board (FASB) in 2002, both financial statement setters agreed that convergence of national accounting standards was desirable (FASB/IASB, 2002). However, with International Financial Reporting Standards (IFRS) as issued by the IASB and U.S. Generally Accepted Accounting Principles (U.S. GAAP) as issued by the FASB as the two most prominent sets of accounting standards (Willmore, 2015), either one of them has to give in. This is, IFRS have a principles-based nature, while U.S. GAAP generally are more rules-based standards. Advocates of rules-based standards (RBSs) emphasize that more detailed guidance lowers potential lawsuits costs, enhances comparability, and confines earnings management practices (e.g., Benston et al., 2006). Concurrently, the arguments in favor of principles-based standards (PBSs) challenge these positions by stating that RBSs create a compliance based mentality that favor the form of a transaction over its economic substance (e.g., Schipper, 2003). In turn, proponents of more PBSs assert that this focus on compliance actually invites lawsuits and earnings management practices and only creates a false feeling of comparability (Swanson, Singer & Downs, 2013).

A general tendency is observed that the FASB is adopting a more principles-based approach, but the debate surrounding this issue is far from over (Ong & Hussey, 2005). This is partly due to the ambiguousness of the “desirable and undesirable effects of such a paradigm shift” (Agoglia, Doupnik & Tsakumis, 2011, p. 748). To address this uncertainty, several requests

1_{The italic style highlighting the conversation between Gollum and his master is added to the initial quote for}

(7)

have been expressed in prior literature to analyze relative earnings management opportunities and practices among firms reporting in accordance with considerably PBSs and those preparing their financial statements under more RBSs (e.g., Schipper, 2003; Liu, Yip Yuen, Yao & Chan, 2014; McEnroe & Sullivan, 2013 and Lawrence, Sloan & Sun, 2013). Combining this with the increasing importance of intangible assets in the economy (Savickaite, 2014), the investigation of the relative exploitation of management discretion in goodwill accounting under IFRS and U.S. GAAP that is offered in this thesis strives to be a relevant contribution to the literature on financial accounting that has not been investigated in the financial accounting literature before. Through this analysis, I strive to contribute to the rules-versus-principles debate that still occupies but is not limited to financial statement setters.

Although accounting for goodwill has been subjected to a certain degree of convergence among IFRS and U.S. GAAP, it will become clear that, when taking a closer look at the technicalities, it can still be considered to represent some of the principles-based roots of the IASB under IFRS and some of the rules-based roots of the FASB under U.S. GAAP. Therefore, by taking a sample of firm-years adopting IFRS and a sample of firm-years adopting U.S. GAAP and looking at their relative goodwill impairment decisions, I should be able to draw a conclusion with regard to the relative performance of both of these sets of standards in the context of goodwill accounting. In order to be able to evaluate the relative exploitation of management discretion among IFRS and U.S. GAAP specifically, both samples will be limited to firm-years having market indications that a goodwill write-off can be expected. Additionally, a discrepancy between private information incentives and agency theory-related incentives is made to be able to isolate the undesirable exploitation of the management discretion offered in both standards. Coining solely the latter opportunistic accounting decisions, the research question can be formulated as: What is the influence of a more PBS or a more RBS on opportunistic goodwill impairment decisions? Put differently, will managers exploit more of their management discretion due to agency theory-related motives when faced with either a RBS or a PBS?

The results of this study should be of interest to standard setters, since they can be used in the prevailing rules-versus-principles discussion. Additionally, because financial statement preparers and users are extensively involved in the standard setting process and also have a lot of interest in the outcome of this process, they should find the results equally interesting.

The remainder of this dissertation will be structured as follows. The next section will elaborate on the background of the aim for convergence of the IASB and the FASB, the rules-versus-principles debate that resulted from this, and the specific technicalities of accounting for goodwill under IFRS and U.S. GAAP. In Section 3, the literature relating to managerial discretion

(8)

will be discussed and the relative subjectivity in goodwill accounting under the sets of standards of interest. This section will conclude with the formulation of the hypotheses. In Section 4, the methodology will be outlined in detail, thoroughly motivating and defining all variables of interest. Section 5 interprets the results and Section 6 and 7 subsequently summarize these results and give comments on them.

(9)

2 Background

This chapter will start with a discussion on the need for a single set of global financial reporting standards and continues to elaborate on the resulting rules-versus-principles debate. It concludes with a description of the regulatory background of goodwill accounting under IFRS and U.S. GAAP, covering the technicalities of the initial recognition and subsequent measurement of goodwill.

2.1 The Road Towards Globally Applicable Financial Reporting Standards

With increasingly international competing companies and rapid improvements in technology, the notion that we should have one single set of global financial reporting standards has gained widespread recognition. Applying the same standards in all leading capital markets is argued to increase financial reporting transparency, in turn increasing investor confidence, eventually enhancing the ability to raise capital and also lowering the cost of that capital (Moser, 2014). This would be beneficial for society as a whole as it stimulates economic activity. Additionally, the greater degree of comparability that would result from elimination of the many different local reporting standards will likely assist investors when making economic decisions (IMA Research Foundation, 2014). In consonance with this line of reasoning, the objective of the International Accounting Standards Board (IASB), established in 2001 to replace the International Accounting Standards Committee, is “to develop International Financial Reporting Standards (IFRS) that bring transparency, accountability and efficiency to financial markets around the world” (IFRS Foundation, 2015). The IASB has been highly successful in striving for this with currently more than 100 countries requiring IFRS as the basis for the financial reports of their publicly traded companies (IMA Research Foundation, 2014).

However, the arguments in favor of one single standard setter prescribing one single set of standards are not unquestioned. Research that articulates the importance of locally set standards posits that “due to the substantial cross-country variation in political, regulatory, litigation, and enforcement environments, it is unlikely that a single standard setter will satisfy the needs of more than 100 IFRS-applying countries” (Kaya & Pillhofer, 2013). Notwithstanding it, the objective of IFRS is particularly to be the solution for global financial market, thereby inherently implying that they are not the perfect answer for individual nations (IMA Research Foundation, 2014). In addition, the elimination of competition between different standard setters is argued to impose too much political pressure on the potential emerged monopolistic standard setter (Kothari, Ramanna & Skinner, 2010). This will likely threaten the independence it would need in order to provide unbiased and understandable standards that are universally applicable.

(10)

One major capital market that still requires its own set of standards is that of the U.S. This is, the Financial Accounting Standards Board (FASB), founded two decades before the establishment of the IASB, prescribes for U.S. firms that they prepare their financial statements in accordance with U.S. Generally Accepted Accounting Principles (U.S. GAAP). Despite the varying arguments that have been expressed and examined on the matter of global financial reporting convergence, the IASB and the FASB have together agreed upon the importance of common global financial reporting standards. This has been formalized in the Norwalk Agreement that they both signed in 2002, stating that they “each acknowledged their commitment to the development of high-quality, compatible accounting standards that could be used for both domestic and cross-border financial reporting” (FASB/IASB, 2002). To accomplish this, they assigned high priority to “removing a variety of individual differences between U.S. GAAP and IFRS” (FASB/IASB, 2002). This implied that they would both consider revision of standards that currently differ substantially among both sets of standards. In their memorandum of understanding in 2006, they reinforced these intentions and additionally expressed their common goal to ultimately develop a single set of high-quality financial reporting standards that would be globally applicable (FASB/IASB, 2006).

2.2 The Rules-Versus-Principles Debate

With the agreement to converge their sets of standards, the IASB and the FASB faced one critical issue that needed to be solved. This is, both boards historically adopted two very distinct philosophies guiding their approach to standard setting. The standards produced by the FASB are generally understood as more detailed and thorough than the standards produced by the IASB, resulting in the former to be classified as rules-based standards (RBSs) and the latter as principles-based standards (PBSs). As described by Bradbury and Schoder (2012), RBSs “have more rules, more justification, acknowledge less judgment is required, have more bright-line thresholds, have more scope exceptions, and are more verbose and complex”. To the contrary, Carmona and Trombetta (2008) describe PBSs as being characterized by “fundamental understandings that inform transactions and economic events [that] dominate any other rule established in the standard”. As such, PBSs are less detailed an rely more on professional judgment. The so-called ‘rules-versus-principles debate’ that sprouts from this discrepancy has hindered the IASB/FASB convergence project from the very beginning (Willmore, 2015).

Prior research has expressed numerous arguments in favor of both ideologies. Those in favor of a RBS have first of all argued that having more detailed financial reporting standards will lower the potential lawsuit costs in case of an accounting dispute because they allow less room

(11)

for differences in interpretation (e.g., Benston et al., 2006 and Liu, Yip Yuen, Yao & Chan, 2014) and because they discourage negotiation about reporting issues (Gibbins, Salterio & Webb, 2001). Audit power is also perceived to be higher in a system with RBS, which in turn lowers potential lawsuit costs as well (Nelson, Elliot & Tarpley, 2002 and Willmore, 2015). Secondly, it is argued that RBSs enhance the comparability of financial statements (e.g., Benston et al., 2006). Through more specific guidance, the chances of application inconsistencies are argued to be lower (Deloitte, 2013). Third and finally, RBS are attested to lower the opportunities for managers to participate in earnings management (e.g., Schipper, 2003; Benston et al., 2006 and Lawrence, Sloan & Sun, 2013). To the extent that more room for discretion in reporting standards lead to more opportunistic earnings manipulation2_{, this presents an argument in favor}

of a rules-based approach to financial reporting standard setting.

RBSs have also been scrutinized. One claim is that a RBS result in too much focus on the form of a financial reporting account, foregoing on the intent of the standard (Kershaw, 2005). This substance over form accusation is argued to stimulate a compliance mentality among financial statement preparers, inviting them to structure financial transactions such that they result in a predetermined desired outcome in line with the rules (Schipper, 2003 and Nelson et al., 2002). Such an approach to financial reporting may also result in erroneously misleading figures, for example by faulty aggregation of ostensibly comparable transactions (Swanson et al., 2013). Additionally, the room for professional judgment associated with PBS is argued to be necessary in order to present some critical accounting issues faithfully (Schipper, 2003; Healy & Whalen, 1999; and Deloitte, 2013). Taking a broader perspective, Coffee (2007) expresses concerns about the overall degree of regulation of the U.S. capital market, as it affects its competitiveness in terms of “the ability to attract foreign listings and offerings” and “the ability to minimize the cost of capital” (Coffee, 2007, p. 309). In line with this, he argues that more PBSs promote the former while impeding the latter.

The great variety of different arguments in (dis)favor of both sides of the debate has resulted in an extensive body of empirical research that investigates the relative performance of RBS and PBS, using a broad range of accounting quality proxies. Table 1 summarizes some of the most important contributions with IFRS and U.S. GAAP as proxies for respectively PBS and RBS.

(12)

TABLE 1 – Summary of Research Evidence on the Rules-Versus-Principles Debate

Author(s) Main Research Question Research Method Main Finding(s) Leuz (2003) Which financial

reporting standard leads to more information asymmetry? Compares the information asymmetry of the two standard regimes using a sample of German New Market firms.

The choice between IFRS and U.S. GAAP appears to be of little consequence for information asymmetry and market liquidity. Barth,

Landsman & Lang (2008)

Are U.S. GAAP-based accounting amounts associated with less earnings management, more timely loss recognition, and higher value relevance than IFRS-based accounting amounts?

Comparison of

accounting quality metrics of IFRS firms to those of US firms.

The results suggest that although IAS accounting amounts are of lower quality than those of US GAAP reported by US firms, they are of comparable quality to reconciled U.S. GAAP amounts reported by cross-listed firms. Van der Meulen, Gaeremynck & Willekens (2007) Which financial reporting standard, IFRS or U.S. GAAP, leads to higher quality earnings figures?

Compare the financial statement quality (measured applying multiple earnings attributes) of the two standard regimes using a sample of German New Market firms.

Overall, financial statement quality is very comparable. Only regarding the predictive ability of accounting information U.S. GAAP is found to perform better.

Gordon, Jorgensen & Linthicum (2010)

1. Are IFRS earnings of higher or lower quality than US GAAP earnings? 2. Are IFRS earnings

more or less informative than US GAAP earnings?

Compare proxies for earnings quality and earnings informativeness of all firms that file Form 20-F reconciliations from IFRS to USGAAP with the SEC during the period 2004 to 2006.

1. U.S. GAAP and IFRS share many earnings attributes. However, U.S. GAAP was found to exhibit higher cash persistence and higher value relevance.

2. U.S. GAAP was found to exhibit a higher relative information content and its net income has incremental informativeness over IFRS earnings and cash flows.

Agoglia, Doupnik & Tsakumis (2011)

What is the influence of standard precision and audit committee strength on financial reporting decisions? Two experiments analyzing U.S.-based financial statement preparers in a lease classification context.

1. Financial statement preparers are less likely to report aggressively when applying a more principles-based standards then when applying a more rules-based standard. 2. When faced with more rules-based

standards, financial statement preparers are less likely to report aggressively in the presence of a strong audit committee than in the presence of a weak audit committee.

Liu, Yip Yuen, Yao & Chan (2014)

What are the differences between IFRS and U.S. GAAP in terms earnings management facilitation?

Empirical regression that measures the difference in income smoothing through discretionary R&D, discretionary deferred tax expense and discretionary accruals.

Earnings management through R&D investment is significantly higher for the IAS/IFRS firms. Earnings management through accruals is not found to be significantly different between U.S. GAAP and IAS/IFRS firms.

(13)

TABLE 1 – Summary of Research Evidence on the Rules-Versus-Principles Debate (cont.) McEnroe &

Sullivan (2013) Are PBS preferred over RBS in terms of achieving the qualitative characteristics put forth by the Conceptual Framework? Examination of perceptions of important stakeholder groups regarding various financial attributes under PBS versus RBS.

In some cases, the qualitative characteristics of the Conceptual Framework are likely to be better achieved by RBS rather than by more PBS. Hong (2008) Do PBSs or RBSs lead to higher quality financial reporting? Analysis of relative earnings quality of 4095 firm-year observations of Chinese firms that report under both a principles-based set of standard and a rules-based set of standard.

PBSs lead to higher quality financial reporting decisions than RBSs.

Nelson, Elliott & Tarpley (2002)

To what extent does standard precision affect earnings management practices?

Field-based questionnaire of 253 auditors from a Big 5 audit firm that recalled and described 515 specific experiences with clients they believed attempted to opportunistically manipulate earnings.

Managers are more likely to attempt earnings management, and auditors are less likely to adjust earnings

management attempts, which are structured (not structured) with respect to precise (imprecise) standards.

Collins, Pasewark & Riley (2012) Do more rules-based lease accounting standards influence the classification of leases as capital or operating leases?

Archival research that examines lease

classification decisions in the period 2007-2009 of a matched sample of IFRS and U.S. GAAP reporting firms.

U.S. GAAP firms are more likely to classify leases as operating leases than IFRS firms; evidence in favor of a more principles-based approach to standard setting.

Van Beest

(2009) To what extent are earnings management practices affected by application of more rules-based or more principles-based financial reporting standards?

Use of two methods: 1. Experiment with a controlled environment. 2. Survey to complement the experiment.

It is not found true that managers faced with PBSs engage more often in earnings management through accounting decisions.

Nobes (2005) Are RBSs or PBSs more appropriate as guidance for financial reporting?

Theoretical analysis of four critical matters of financial reporting.

Standards currently considered rules-based are partly consisting of detailed guidance because they lack proper principles. The author argues that if the underlying principles would be improved, less “rules” would be needed, enhancing the overall quality of financial reporting standards. Ewert &

Wagenhofer (2005)

Do RBSs reduce earnings management practices and do they lead to more relevant information to the capital market?

Analysis of a rational expectations equilibrium model.

RBSs lead to less opportunistically manipulated earnings figures and provide more relevant information to the capital market, but this comes at some costs that potentially outweigh these benefits.

(14)

As can be observed, the results are mixed. Some provide evidence that RBSs are a superior reporting scheme (e.g., Gordon et al., 2010; Agoglia & Tsakumis, 2011; Liu et al., 2014; McEnroe & Sullivan, 2013; Nelson et al., 2002 and Ewert & Wagenhofer, 2005), while according to others PBSs are leading to more high quality earnings figures (e.g., Nobes, 2005; Hong, 2008; Collins et al., 2012; add more). Nonetheless, the FASB and the IASB agreed in their joint meeting in April 2008 that “the goal of joint projects is to produce common, principles-based standards, subject to the required due process” (FASB/IASB, 2008). In line with this, the FASB has communicated considerations that envisage that “some or all U.S. public companies might be permitted or required to adopt IFRS at some future date” (FASB/IASB, 2008). Although a final decision on the actual adoption of IFRS by the FASB has not been made yet, the board has gradually been adopting the principles-based approach in its newly issued standards (Bradbury & Schroder, 2012). As put by Wood et al. (2008, p. 8):

“It is clear […] that the debate over whether principles or rules should be the underlying concept in creating effective financial reporting is over. Principles have won the day. But there are many shades of debate and opinion within that consensus”.

This statement implies that although the consensus in favor of PBSs may be reached, plenty of research opportunities still exist in order to sort out the “many shades of debate and opinion”. Collins, Pasewark and Riley (2012) make an attempt to contribute to this research gap in the context of lease accounting. They reinforce the importance of further research investigating differences in financial reporting outcomes among financial statements prepared in accordance with IFRS and U.S. GAAP by stating (Collings, Pasewark & Riley, 2012, p. 683):

“As the FASB and the IASB move toward the development of joint standards, and as the SEC considers adoption of IFRS, it is useful to understand how these different accounting standards might lead to different reporting outcomes”.

The current thesis strives to contribute to filling this void by investigating one particular context of different reporting outcomes; accounting for goodwill impairments.

2.3 Goodwill Accounting under IFRS and U.S. GAAP

When an acquisition takes place, the consideration paid for the acquired company is generally higher than its book value. This premium is recognized as goodwill on the balance sheet of the acquiring company and is justified on the basis of managements’ expectations of the newly

(15)

acquired company or of the synergies the newly created business combination will generate (Giuliani & Brånnström, 2011). As such, goodwill is one of the most subjective items that can be found on the balance sheet. Given this ambiguous nature, initial and subsequent accounting for goodwill has been discussed extensively over the course of time, with several changes in the approach under both U.S. GAAP and IFRS taking place (Rees & Janes, 2012). As it is not the purpose of this thesis to elaborate on the development of the current accounting treatment of goodwill under both sets of standards, this section will be limited to an outline of the current practice.

Upon its formation, the Business Combinations project was already part of the IASB agenda (IASB, 2015). Up to then, IAS 22 Business Combinations prescribed any goodwill that arose from an acquisition to be recognized as an asset and subsequently amortized over its useful life. The FASB had a similar project taking place earlier, which resulted in the issuance of SFAS 141

Business Combinations and SFAS 142 Goodwill and Other Intangible Assets in June 2001 (FASB, 2001a

and FASB, 2001b). These two standards changed the accounting treatment of goodwill from the amortization approach that was also prevalent under IAS 22 into an impairment-only approach that required goodwill to be capitalized and subsequently tested for impairment at least annually, without a structural periodical write-off taking place. SFAS 142.18 defines a goodwill impairment as “the condition that exists when the carrying amount of goodwill exceeds its implied fair value” (FASB, 2001b, p. 12). Under U.S. GAAP, goodwill is initially allocated to the reporting units3

that are expected to benefit from the business combination (FASB, 2001b). Subsequently, SFAS 142 requires goodwill to be tested for impairment through a two-step procedure. In the first step, the fair value of a reporting unit is compared with its carrying amount including the carrying value of the goodwill allocated to it. The second step will only be performed if this reporting unit carrying value exceeds its estimated fair value. If this is the case, in the second step the impairment loss to be recognized will be measured by comparing the implied fair value of the reporting unit goodwill with the carrying amount of that goodwill. This implied fair value is calculated by first “allocating the fair value of a reporting unit to all of the assets and liabilities of that unit (including any unrecognized intangible assets) as if the reporting unit had been acquired in a business combination and the fair value of the reporting unit was the price paid to acquire

3_{Defined by SFAS 142.30 (FASB, 2001b, p. 15) as “an operating segment or one level below an operating}

segment”. SFAS 131.10 defines an operating segment as a component of an enterprise that has three characteristics (FASB, 1997):

1. It engages in business activities from which it may earn revenues and incur expenses.

2. Its operating results are regularly reviewed by the enterprise’s chief operating decision maker to make decisions about resources to be allocated to the segment and assess its performance.

(16)

the reporting unit” and then subtracting this from the fair value of the reporting unit (SFAS 142.21) (FASB, 2001b, p. 13). The FASB justifies this impairment-only approach in the introductory paragraph of SFAS 142. They argue that, as intangible assets became increasingly important economic resources for business entities and substantiated for an increasingly large proportion of the transactions taking place, better information about these intangibles was needed. Since they observed that financial statement users did not perceive goodwill amortization expenses to be useful information in analyzing investment, the shift to an impairment-only approach was a logical step to take (FASB, 2001b).

The IASB project on Business Combinations reassessed IAS 22 “with the objective of improving the quality of, and seeking international convergence on, the accounting for business combinations” (IASB, 2015). Among the concrete outcomes of this project were the issuance of IFRS 3 Business Combinations and a revision of IAS 36 Impairment of Assets and IAS 38 Intangible

Assets in March 2004, aligning the goodwill accounting treatment of IFRS broadly with that of

U.S. GAAP. That is, IFRS 3 and the revised version of IAS 36 also require firms to adopt an impairment-only approach to writing off goodwill (IASB, 2005a and IASB, 2005b). Despite the fact that this increased the overall convergence of both sets of standards with regard to their treatment of business combinations, substantial differences persisted. First of all, upon initial recognition goodwill is allocated to the cash generating unit (CGU)4_{or groups of CGUs that are}

“expected to benefit from the synergies of the combination, irrespective of whether other assets or liabilities of the acquiree are assigned to those units or groups of units” (IASB, 2005a, p. A1195). Consequently, IAS 36.88 stipulates that goodwill shall be tested for impairment at the CGU level as opposed to SFAS 142.37 which prescribes goodwill impairment testing to take place at the reporting unit level.

Secondly, under IFRS, the carrying value of the CGU (or group of CGUs) – including the carrying value of the goodwill allocated to it – is compared to its recoverable amount as opposed to its fair value under U.S. GAAP. The CGU recoverable amount is calculated as the higher of its fair value less costs to sell and its value in use (VIU). The latter is computed as the present value of all expected future cash flows to be generated from the CGU (or group of CGUs). This implies that the entity’s management needs to make several estimations (i.e., with regard to the discount rate, future cash flows, etc.). If the calculated recoverable amount is lower than the carrying value, an impairment loss will be recognized. In practice, it appears that companies

4_{Defined by IAS 36.06 as “the smallest identifiable groups of assets that generates cash inflows that are largely}

(17)

reporting in accordance with IFRS most often use this VIU option rather than the fair value less costs to sell option to determine the recoverable amount of a (group of) CGU(s) (Storå, 2013).

A third major difference concerns the measurement of the goodwill impairment loss to be recognized. As explained earlier, U.S. GAAP measures the impairment loss as the difference between the carrying value of the goodwill allocated to a particular reporting unit and its implied fair value. Contrarily, IFRS measures the impairment loss directly as the difference between the carrying value of a particular CGU and its recoverable amount, without consideration of the implied value of its goodwill and the carrying value of goodwill separately. This is, whereas SFAS 142.18 requires the two-step impairment test that was explained above, IAS 36.90 only instructs a one-step test.

Fourth and finally, the allocation of the goodwill impairment loss to be recognized is also dealt with differently among both reporting standards. U.S. GAAP allocates the impairment loss to the carrying value of the goodwill allocated to the reporting unit, with an impairment loss exceeding this carrying value simply reducing the goodwill account to zero without further actions. Contrarily, IFRS does allow for goodwill impairment losses that exceed the carrying value of goodwill allocated to the CGU, by allocating the residual to the other assets allocated to the CGU on a pro rata basis (IASB, 2005a).

As asserted in IAS 36.124 and SFAS 142.20, reversal of (part of) a goodwill impairment loss previously recognized is prohibited under both sets of financial reporting standards. Similarly, the recognition of internally generated goodwill subsequent to a business combination is not allowed under both IFRS and U.S. GAAP (IAS 38.49 and SFAS 142.B85) (IASB, 2005a and FASB, 2001b).

(18)

3 Hypothesis Development

This chapter will first discuss management discretion inherent to goodwill accounting in the current impairment-only approach of both IAS 36 and SFAS 142 and the incentives managers face to use this discretion. Subsequently, Section 3.2 analyzes the differences between the IASB and the FASB approach to goodwill accounting in terms of the room for management discretion and the incentives for it that are found in Section 3.1. Lastly, the hypotheses that are tested in order to provide an answer to the research question are extrapolated from this relative subjectivity.

3.1 Management Discretion in Goodwill Accounting

While the goodwill to be recognized upon the formation of a business combination is measured relatively straightforward – essentially the premium paid for a company above the carrying value of its net assets – its subsequent value is not as clear-cut (Lhaopadchan, 2010). The goodwill impairment tests under both US GAAP and IFRS are often criticized because they involve significant use of estimates. This is, the IASB and the FASB have left ample room for management discretion in their standards on (subsequent) goodwill accounting, which makes that IAS 36 and SFAS 142 are widespread convicted to facilitate earnings management practices by financial statement preparers (e.g., Ji, 2013; Wines, Dagwell & Windsor, 2007;Caruso, Ferrari & Pisano, 2016 and Seetharaman, Sreenivasan, Sudha & Ya Yee, 2006).

First of all, management may define reporting units and (groups of) CGUs at a higher level in order to cover up possible impairments at lower levels (Jerman & Manzin, 2008). As explained by Wines et al. (2007, p. 868), managers could include or exclude specific “subsidiaries, divisions and/or branches” ad lib.

Additionally, with regard to estimation of the fair value of the reporting unit under U.S. GAAP and the recoverable amount of the (group of) CGU(s) under IFRS, managers are often faced with a lack of comparable entities with publicly observable market value (Caruso et al., 2016). Therefore, a valuation model needs to be applied, on which managers in turn have a significant degree of influence. Correspondingly, they need to make several estimations about for instance future cash flows, discount rates, growth rates or market multiples (Avallone & Quagli, 2015 and Sevin & Schroeder, 2005). As put by Rees and Janes (2012, p. 31), “a market approach to computing the fair value of a reporting unit has proven to be very challenging, because the unique characteristics of a reporting unit make it difficult to find a comparable entity”. Obviously, the same can be said about the estimation of the recoverable amount of the (group

(19)

of) CCG(s). By manipulating these estimated values, managers can influence the necessity of a goodwill impairment and its measurement if one is deemed unavoidable.

Furthermore, Caruso et al. (2016) hypothesize that restructuring activities also provide an excellent opportunity for earnings manipulation through goodwill accounting. This is, although both the IASB and the FASB do not allow a restructuring of a reporting unit or (group of) CGU(s) with goodwill allocated to it, they both permit this under specific circumstances (see SFAS 142.27 and IAS 36.72). Through this venue, goodwill can be reallocated such as to influence the outcome of a goodwill impairment tests in a predetermined direction.

Another source of discretion is related to the prohibition to recognize internally generated goodwill under both sets of standards. Since it is difficult to distinguish the internally generated goodwill subsequent to the business combination taking place from its initially acquired value (Storå, 2013 and Caruso et al., 2016), managers could avoid recognizing a goodwill impairment through filling it up with newly generated goodwill that was initially not recognized.

Beatty and Weber (2006) find evidence that decisions to accelerate or delay goodwill impairment decisions are affected by several discretionary factors. Lhaopadchan (2010) even claims that managerial incentives are the most important determinant of goodwill write-down decisions. On the other hand, research by Ramanna and Watts (2012) and Bond, Govendir and Wells (2016) shows for a sample of firms reporting in accordance with subsequently U.S. GAAP and IFRS combined with at least one externally observable indicator of goodwill impairment loss recognition that this is often postponed or even denied altogether. Other research on the topic investigated and found that goodwill impairment decisions are also being opportunistically manipulated for various other earnings objectives (e.g., Hilton & O’Brien, 2009; Jahmani, Dowling & Torres, 2010; Zang, 2012; Ji, 2013; Szczesny & Valentincic, 2013; and Caruso, Ferrari & Pisano, 2016).

Moreover, is argued and found true by Ramanna (2008) that the development of the current impairment-only treatment of subsequent goodwill accounting prevalent under both U.S. GAAP and IFRS is at least partially due to pressure from executives. This is, prior to SFAS 142, U.S. GAAP allowed two options with regard to subsequent goodwill accounting; the pooling-of-interest method and the purchase method (FASB, 1970). In practice, most managers used the pooling-of-interest method to keep goodwill off the books. When it became clear that ABP Opinion 175_{was going to be replaced with a standard that only allowed one option, the}

(20)

impairment-only approach provided a second-best alternative that avoided the amortization approach executives wanted to refrain (Ramanna, 2008). Since the IASB revised IAS 36 in response to the issuance of SFAS 142 by the FASB, this standard was indirectly also affected by this process. According to this reasoning, the very nature of the current impairment-only approach prescribed by both sets of standards is inherently connected to the exploitation of management discretion.

In general, managers use the discretion provided to them in order to either signal private

information about expected (future) company performance or to exploit opportunities consistent

with incentives supported by agency theory (e.g., Lhaopadchan, 2010 and Filip, Jeanjean & Paugam, 2015). In terms of the latter rationale, Dye (1986) distinguishes internal agency theory-related motives from external agency-related motives. He describes internal motives to be related to earnings being perceived as one of the main indicators of managers’ performance. In the goodwill accounting context, this can be interpreted as an impairment loss that lowers earnings will in turn will lead to a lower evaluation and subsequently to less compensation in the form of bonuses. Data yielded by Healy and Wahlen (1999) supports this by providing convincing evidence that indeed earnings management practices are related to management compensation incentives. With regard to external agency-related motives, Dye (1986) posits that managers need to answer to the expectations of analysts and investors. These external parties regard a goodwill impairment write-off as an indication that management has made an improper and expensive acquisition or did a poor job exploiting the potential synergies, which incentivizes managers to postpone or avoid those impairments whenever perceived to be possible. Healy and Wahlen (1999) also find confirmatory evidence related to these external agency-related motives.

With regard to the private information motives, AbuGhazaleh, Al-Hares and Roberts (2011) show that goodwill impairments in an IFRS sample are more strongly associated with effective governance mechanisms than with agency theory-related proxies and they interpret this as managers conveying private information about future firm performance expectations to the public. Abuaddous, Hanefah and Laili (2014) find similar results, also in the context of companies reporting in accordance with IFRS.

In sum, it can be concluded that goodwill impairment decisions based on the impairment-only approach currently prescribed by both U.S. GAAP and IFRS can be assumed to be affected by both private information motives and agency theory-related motives. For the purpose of answering the main research question of this dissertation, a distinction will be made between these two types of incentives. Although it is in general expected that more room for discretion will lead to more exploitation of this discretion by managers, manipulation because of

(21)

private information motives should be welcomed by investors while agency theory-related manipulation is undesirable. In line of this reasoning, opportunistic goodwill decisions are in this study only those manipulations that are driven by the latter motive. Taken altogether, it is assumed that goodwill accounting under IAS 36 and SFAS 142 is at least partially due to opportunistic exploitation of management discretion. The next subsection will elaborate on the relative expectation of this opportunistic exploitation among firms applying IFRS and those applying U.S. GAAP, based on findings of prior research.

3.2 Relative Management Discretion in IAS 36 and SFAS 142

Although the above described ambiguousness covers goodwill accounting under both sets of reporting standards, different degrees of it can be distinguished. These can be interpreted as a more principles-based approach to the issue under IFRS and a more rules-based approach under U.S. GAAP. The paragraphs below will support this statement.

The first difference pointed out in Section 2.3 – the difference regarding the unit level basis of the initial allocation and the subsequent impairment test of goodwill – provides for a first difference in scope of management discretion. Some authors have expressed a consensus that the (group of) CGU(s) level introduces more subjectivity in the impairment test than the reporting unit level. As explained by Swanson et al. (2013, p. 29), “a reporting unit is much broader in scope than a CGU to the extent that a responsibility center contains groups of assets each of which may generate separately identifiable cash flows or which synergistically operate to generate cash flows for the unit”. As such, they argue that the FASB approach normally leads to a higher level approach to goodwill impairment testing than the IASB method. Recognizing the lower level approach of the IASB, others argue that more detailed entity-specific estimations need to be made for CGU calculations, in turn increasing the degree of subjectivity associated with it (e.g., Wines et al., 2007 and Storå, 2013). Hamilton, Hyland and Dodd (2011) adopt a more fundamental attitude, reasoning that since the nature of goodwill implies that it is highly integrated in the other assets and liabilities of the business combination it would be economically more appropriate to allocate goodwill to higher level units.

However, counter-arguments have also been suggested. Swanson et al. (2013, p. 29) also provide a rebuttal to the arguments in favor of the reporting unit level approach of the FASB by stating “because CGUs require that assets or asset groupings be related to specific cash flows, which are independent of other cash flows of the company, it is more difficult to avoid [goodwill] impairments when conditions clearly indicate a deterioration”. From this, it can be concluded that performing the goodwill impairment test at the reporting unit level actually

(22)

introduces more subjectivity to it. Moreover, Swanson et al. (2013) identify another complication related to the more aggregated nature inherent to the FASB approach to goodwill impairment testing. They argue that this leads to the managerial discretion resulting from the earlier explained difficulty to distinguish internally generated goodwill from its initially recognized value is more precarious under U.S. GAAP, thereby implying that the IFRS approach offers relatively less management discretion with regard to goodwill impairment decisions.

With respect to the second difference that was stipulated in Section 2.3 – the difference regarding the valuation method of the unit to be tested for impairment – it is said that compared to the fair value alternative, the VIU option as allowed under IFRS invites more opportunistic outcomes because it takes more entity-specific elements into account, which are in turn easier to be manipulated (e.g., Caruso et al., 2016 and Storå, 2013). Besides being more specific, the estimations related to the VIU option are further complicated by additional considerations such as the selection of the appropriate discount rate, the risk assessment for each (group of) CGU(s) on each other, tax deductibility issues, depreciation treatments and some financial costs (Abbuadous et al., 2014). Camodeca, Almici and Bernardi (2013, p. 22) conclude that “all [this] raises well-grounded doubts on the usefulness and overall reliability of the current estimation model for the recoverable value of goodwill”.

While the VIU option is by some understood to be more prone to opportunistic exploitation of management discretion, it is argued by other authors that under specific market conditions it still may be the more appropriate approach (e.g., Cramer, 2006 and Comiskey & Mulford, 2010). Comiskey and Mulford (2010) note that sometimes market values are not rational or considered too timely to faithfully reflect the true and fair value. Additionally, they argue that sometimes market values are simply not readily available. As explained by Hamilton et al. (2011, p. 57) the evidence available for estimation of an asset’s value can be ranked with “a price in a binding sale agreement in an arm’s length transaction” representing the best option, “an active market for the asset where similar assets would be traded” as second-best alternative, and “the best information available to reflect an amount that the entity could obtain in in arm’s length transaction between willing and knowing parties” as least reliable basis. In cases such as the latter, whilst recognizing its increased subjectivity, the more principles-based approach of IFRS would be preferable over its U.S. GAAP counterpart.

Regarding the two-step versus one-step method, it is posited by Willmore (2015, p. 6) that by not considering the implied fair value of goodwill, IFRS “measures an impairment without being sure that one exists”. Hamilton et al. (2011) analyze the IASB approach in a similar manner. This can be interpreted as U.S. GAAP to provide for a better procedure, since it

(23)

considers the implied fair value of goodwill for the measurement of the impairment loss rather than simply writing goodwill off with the excess carrying value of the unit under scrutiny. However, Schultze (2005) fights this interpretation by noting that because of the implied fair value determination, the goodwill impairment decision under U.S. GAAP is not only dependent on the reporting unit fair value6_{calculation, but also on the change in the fair value of its net}

assets. Further, since “both internally developed intangible assets and newly generated hidden reserves increase the fair value of net assets for purposes of the impairment test” they positively influence the goodwill impairment loss (Schultze, 2005, p. 279). This creates a so called form over

substance approach, because those increases in the value of the other assets in the unit being

tested are not shown on the balance sheet. According to his reasoning, the two-step approach “creates misleading accounting” (Schultze, 2005, p. 295) and the one-step approach of the IASB is – although more principles-based – more relevant for financial statement users.

The last difference that was mentioned in Section 2.2 concerns the allocation of a goodwill impairment loss when it is determined that one exists and its value has been estimated. It was explained that under U.S. GAAP the impairment loss is limited to the value of goodwill recognized at that moment, while under IFRS a potential excess is allocated to the other assets of the unit on a pro rata basis. This can again be interpreted as to reflect a more principles-based approach under IFRS relative to the U.S. GAAP conduct, since it allows larger losses to be recognized when deemed economically funded. Similarly, Cramer (2006, p. 15) puts forward the view that “IFRS […] allows for rational methods to be used as allocation if the impairment of goodwill which allows for flexibility, one of the effects of implementing principles-based standards”.

From the argumentation above it can be concluded that, while the move to the impairment-only approach by the IASB and the FASB is generally understood to embody convergence between both standards setters, prevailing differences in accounting for goodwill impairments continues to reveal the rules-versus-principles debate among IFRS and U.S. GAAP. With the exception of the first one, in terms of all differences as pointed out in Section 2.2 there appears to be consensus that IAS 36 still is considered a more PBS and SFAS 142 more a more RBS. This provides a research setting in which the implications of both sets of standards on goodwill impairments can be investigated, since, in light of the discussion on management exploitation of discretion offered in goodwill accounting under IFRS and U.S. GAAP, it can be inferred that application of IAS 36, adopting a more principles-based approach, will most likely

(24)

lead to more opportunistic goodwill impairment decisions than application of SFAS 142, which applies a more rules-based perspective (see also Willmore, 2015). On these grounds, the two hypotheses of this dissertation are, in the alternative form, formulated as:

H1 : Application of IFRS will, relative to application of U.S. GAAP, lead to more

opportunistic goodwill impairment loss recognition decisions.

H2 : Application of IFRS will, relative to application of U.S. GAAP, lead to more

(25)

4 Methodology

This chapter will describe the research method and correspondingly the sample and variables used. It will conclude with the regression equations that are used to test H1 and H2.

4.1 Research Method

As introduced in the introductory chapter, the main research question of this thesis is: in the context of goodwill impairment decisions, will application of IFRS lead to more opportunistic financial reporting decision when compared to application of U.S. GAAP? An answer to this inquiry will be sought after by, firstly, taking two samples of firms; one reporting in accordance with IFRS and one with U.S. GAAP. In order to be able to compare the opportunistic exploitation of management discretion in both samples, both samples only include firms that (1) have a substantial book goodwill account and (2) show market indications of a goodwill impairment loss to be recognized. Subsequently, based on the model of Ramanna and Watts (2012), proxies for both private information motives and agency-theory related incentives are used to analyze the recognition and measurement of goodwill impairments in both samples. This analysis starts with univariate frequency tests of association of all dichotomous independent variables on goodwill non-impairment. Thereafter, the Pearson correlations between the continuous explanatory variables are analyzed. Ultimately, a cross-sectional multivariate regression will be run for both samples of the proxies for the private information motive, the proxies for the agency theory-related motive, and some control variables on subsequently impairment recognition and impairment measurement.

4.2 Sample Selection

Based on the model of Ramanna and Watts (2012), two samples were gathered; one consisting of firms that prepare their financial reports in accordance with IFRS and one consisting of firms that prepare their financial reports in accordance with U.S. GAAP. The samples consist of firm-years covering the period from 2005 until 2010. Years before 2005 cannot be included because the IASB required the impairment-only method that year for the first time for its publicly traded firms. Further, only firms that had a book goodwill of at least one million were included so that the identified motives would be strong enough to be of significant effect and financial service organizations were excluded from the sample because of their specific nature. As argued by Ramanna and Watts (2012), firms with an equity-book-value greater than its equity-market-value (BTM>1) are likely to be overvalued. Therefore, it is assumed that for those firm-years, investors

(26)

expect net asset impairments to take place7_{. Since the sample firm-years all have substantial}

goodwill accounts, at least some of this impairment should be directed towards this account. However, Ramanna and Watts (2012) continue to recognize that “[this] condition can also be generated by certain GAAP rules on contingent losses, deferred taxes, and the impairment of (non-goodwill) long-lived assets”. To alleviate this potentiality, only firm-years with a BTM>1 of which the prior firm-year also had a BTM>1 are included in the sample. Thus, when referred to the year t, the second firm-year with a BTM>1 is contemplated. The Compustat database was consulted to retrieve this data, which resulted in a sample of 368 firm-years for the IFRS sample and 192 firm-years for the U.S. GAAP sample. A closer look at the samples reveals that the IFRS sample includes 157 unique firms and the U.S. GAAP sample 87. An overlap of 7 firms is detected who switched from U.S. GAAP to IFRS at some year included in the period under investigation. Assuming that this change did not affect the relative impact of adoption of the set of financial standards on goodwill impairment decisions, these firms were not excluded from the sample. Furthermore, the IFRS sample consists of companies from 28 unique countries and the U.S. GAAP sample includes data from 15 different nations. From these lists, an overlap of 11 countries is observed that are represented in the IFRS sample and also in the U.S. GAAP sample.

4.3 Variables Description

This section will motivate and describe the proxies used for the variables used in the regression analysis for the purpose of answering the research question of this thesis. The choice and definition of the variables is based on Ramanna and Watts (2012), who analyze opportunistic goodwill impairment decisions in a U.S. GAAP context.

4.3.1 Dependent Variables

It was introduced earlier that goodwill impairments can be analyzed by occurrence (recognition) and by magnitude (measurement). To test H1, a regression will be run with a dummy for

goodwill impairment recognition as dependent variable. This variable is defined as ImpR. If for a specific firm-year goodwill is impaired, ImpR = 1 and ImpR = 0 if it is not impaired. To test H2,

the same regression will be run, but with impairment measurement as dependent variable. This

7_{This claim is also put forward by Swanson et al. (2013, p. 36) who state that “firms with more growth prospects}

(i.e., low ratio, high market value) are less likely to have impaired goodwill and, therefore, are less likely to incur a goodwill impairment”.

(27)

variable is defined as ImpM and is proxied for by taking the impairment loss recognized by a firm in year t and dividing it over the total assets of that firm in year t-1.

4.3.2 Independent Variables

Based on the discussion in the hypothesis development chapter, proxies for both private information motives and agency theory-related motives for manipulating goodwill impairment decisions are taken as explanatory independent variables.

4.3.2.1 Private Information Motives

It was argued earlier that room for discretion in standards provides managers with opportunities to reveal information to investors about performance that would not be visible under more strict standards. In the context of goodwill accounting, this is reinforced by the FASB (2001, p. 7) who state that “the changes included in this Statement will improve financial reporting because the financial statements of entities that acquire goodwill and other intangible assets will better reflect the underlying economics of those assets”. From Ramanna and Watts (2012, p. 12) positive net share repurchase activity and positive net insider buying “can be interpreted as rational responses by managers to situations where their stock is undervalued”. On these grounds, firm-years that show either of these two are assumed to be firm-years in which managers have positive inside information. Positive net share repurchase activity is retrieved from the Thomson ONE database and is proxied by Repurchase = 1 if the share repurchase activity is positive and Repurchase = 0 if otherwise. Positive net insider buying is also retrieved from the Thomson ONE database and is proxied for by Insider = 1 if insider buying is positive and Insider = 0 if otherwise. Firm-years with either Repurchase = 1 or Insider = 1 are indicated by InfoAsym = 1.

4.3.2.2 Agency Theory-Related Motives

From the prior argumentation it can be concluded that in terms of goodwill impairment decisions, at least some of the manipulations will not be motivated by private information incentives, but rather on agency theory-related grounds. According to Ramanna and Watts (2012), the agency theory-related motives can be subdivided into contracting motives, reputation motives, and valuation motives. Each of them will be accounted for in the regression analysis.

4.3.2.2.1 Contracting Motives

Financial statement preparers might have contracting-related reasons not to impair goodwill or to accelerate goodwill impairments. Firstly, it is found that “managers make accounting choices consistent with bonus maximization” (Fields, Lys & Vincent, 2001, p. 299). Therefore, accounting-based bonus schemes have often been voiced as an important driver of earnings

(28)

management practices (e.g., Fields et al., 2001; Beatty & Weber, 2006 and Ramanna & Watts, 2012). Therefore, the first proxy that is used to indicate contracting motives is whether or not the CEO received an accounting-based compensation bonus in a given firm-year. This data is retrieved from Compustat and given the label Bonus. The dummy variable Bonus = 1 if the CEO received such a bonus and Bonus = 0 if otherwise.

Another factor related to contracting incentives for earnings manipulation is whether or not “the firm is being traded on an exchange with accounting-based delisting requirements” (Ramanna & Watts, 2012, p. 15). This is intuitive when one recognizes that these delisting requirements increase the cost of underperforming earnings. Potential delisting or even being subjected to a debate concerning this may cause substantial harm to a company, increasing the willingness for managers to manipulate the figures in order to avoid these troublesome scenarios. Beatty and Weber (2006) also find evidence confirming this assertion. Therefore, a dummy variable Delist is included that is set to 1 if the particular firm-year is listed on an exchange with such requirements and to 0 if otherwise. Based on Beatty and Weber (2006, p. 267), firm-years listed on either the NASDAQ or the AMEX “have objective delisting requirements based on numerous factors, including the firm’s net worth. […] The OTC [and NYSE] does not have such delisting requirements”. The variable Delist is set to 1 or 0 corresponding to these exchange listings as retrieved from Compustat.

A third and final proxy for contracting motives is the leverage ratio. Swanson et al. (2013) also include this in their analysis of goodwill impairment opportunism and defend this variable by stating that “more leveraged firms are expected to be more likely to impair goodwill because a higher level of debt in the capital structure may subject the firm to greater oversight due to contracting provisions of bond indentures. Accordingly, the firm is less likely to hide any deterioration in market value” (Swanson et al., 2013, p. 36). The leverage ratio is labeled Lev and it is computed – similar to Ramanna and Watts (2012) – by dividing for a specific firm-year total debt in year t over total assets in year t-1.

4.3.2.2.2 Reputation Motive

The longer a CEO is working for a company, the higher the likelihood that that specific CEO was also involved in the acquisition that led to the goodwill account. Simultaneously, to the general investors’ understanding, goodwill write-offs mostly indicate that either the consideration paid in the acquisition was too high or management underperformed compared to its initial expectations (Beatty & Weber, 2006 and Ramanna & Watts, 2012). Taken together, these reputation concerns imply that “firms whose CEOs have relatively longer tenures also are less likely to take write-offs” (Beatty & Weber, 2006, p. 259). Building on these grounds, an

(29)

independent variable, Tenure, is included in the regression analysis. This variable is proxied – similar to Beatty and Weber (2006) and Ramanna and Watts (2012) – by the number of years the CEO of a specific firm-year has been in that position.

4.3.2.2.3 Valuation Motive

Another incentive sprouting from the agency theory school of thought is related to valuation concerns. As explained by Fields et al. (2001), the effect of managerial choices on equity valuations is another driver of earnings management practices. Drawing on this, Beatty and Weber (2006, p. 265) hypothesize that “firms whose […] goodwill impairments are expected to be more highly capitalized by the market will be more likely to record SFAS 142 goodwill impairment charges and will record relatively larger charges when adopting SFAS 142”. Ramanna and Watts (2012) proxy for capitalization of earnings in returns by taking the earnings response coefficient (ERC), included as ERC in the regression equation. Based on this reasoning it is expected that a higher ERC will increase the incentive for managers to opportunistically manipulate earnings, and thus lead to less and lower goodwill impairments if possible. Similar to Ramanna and Watts (2012), ERC for each firm-year in the sample is computed by regression of the firm’s share price on its operating income. For the regressions in this thesis, quarterly figures have been used of 5 years prior to the particular firm-year.

4.3.3 Control Variables

Similar to Ramanna and Watts (2012), two proxies are used to control for the “economic necessity of a write-off” and two proxies are used to control for the “magnitude of goodwill write-offs” (Ramanna & Watts, 2012, p. 21). The first control variable used for recognition of goodwill impairments is the variable BnH, indicating the buy-and-hold return of the particular firm-year. This included possible dividends and share price increases (decreases). Secondly, the variable QBTM>1 indicates the number of quarters the particular firm-year in the sample had a BTM>1.

The control variables used for measurement of goodwill impairments are Size and

PropGW. The former controls for the particular firm-years’ firm size and is computed by taking

the natural log of its total assets at the end of fiscal year t-1. The latter controls for the particular firm-years’ proportion of goodwill and is determined by dividing total goodwill at the end of fiscal year t-1 over total assets at the end of fiscal year t-2. Besides Ramanna and Watts (2012), Swanson et al. (2013) also control for firm size and goodwill proportion in their analysis of goodwill impairment decisions.

(30)

4.4 Regression Equations

As explained in Section 4.1, H₁ and H₂ are ultimately tested by running two cross-sectional multivariate regressions. Although the independent variables stay the same for both hypotheses, the dependent variable concerns ImpR for H1 and ImpM for H2. Additionally, as was explained

earlier, different control variables apply for each regression. For each sample firm-year i, equation 1 represents the regression equation for H1 and equation 2 represents it for H2.

H₁: ImpR_i = Intercept + b₁InfoAsym_i + b₂Bonus_i + b₃Delist_i+ b₄Lev_i + b5Tenurei + b6ERCi + b7BnHi + b8QBTM>1i + εi

(1)

H2 : ImpMi = Intercept + b1InfoAsymi + b2Bonusi + b3Delisti + b4Levi +

b₅Tenure_i + b₆ERC_i + b₇Size_i + b₈PropGW_i + ε_i (2)

Because firm-years from the same firm will most likely show correlated observations, the regressions are clustered per firm. The remaining heteroskedasticity is corrected by running robust regressions. Equation 1 is estimated by running a binomial logistic regression, since ImpR is a dichotomous variable. As ImpM has a continuous character, the beta coefficients of these explanatory variables are estimated by an ordinary least squares regression.

Opportunism in Goodwill Impairment Decisions : IAS 36 versus SFAS 142