MSc Accountancy & Control, track Accountancy Faculty of Economics and Business, University of Amsterdam
Empirical evaluation of the relationship between
goodwill impairment losses and stock price
Jeroen Schaapveld 10662391
Supervisor: dr. A. Sikalidis 11 June 2015
2 Statement of Originality
This document is written by student Jeroen Frank Schaapveld 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
1 Introduction ... 5
2 Literature review and hypotheses development ... 8
2.1 IFRS 3 – Intangible assets and goodwill ... 8
2.2 Estimates in the recoverable amount ... 10
2.2.1 Agency theory and the estimates in the recoverable amount ... 10
2.3 The reliability of goodwill numbers ... 11
2.4 The impact of goodwill impairments on stock performance ... 12
2.5 Moderator relationship of the use of discretion on the direct relationship between goodwill impairment and stock performance ... 15
2.6 Hypotheses development... 18
2.6.1 The effect of a goodwill impairment loss on the stock price ... 18
2.6.2 Proxies for managerial discretion ... 19
2.6.2.1 Management change ... 19
2.6.2.2 Board independence ... 20
2.6.2.3 Board activity ... 21
2.6.2.4 Financially literate audit committee members... 22
3 Research design ... 23
3.1 Sample ... 23
3.2 Variables and measurement ... 26
3.2.1 The Ohlson (1995) model ... 26
3.2.2 The effect of a goodwill impairment loss on the stock price ... 27
3.2.3 The effect of management change on the relationship between a goodwill impairment loss and the stock price ... 27
3.2.4 The effect of Board independence on the relationship between a goodwill impairment loss and the stock price ... 28
3.2.5 The effect of Board activity on the relationship between goodwill impairment losses and the stock price ... 28
3.2.6 The effect of financially literate audit committee members on the relationship between goodwill impairment losses and the stock price ... 29
4
4 Empirical results ... 29
4.1 Descriptive Statistics ... 30
4.2 Multivariate Results ... 35
4.2.1 The effect of a goodwill impairment loss on the stock price ... 35
4.2.2 The effect of management change on the relationship between a goodwill impairment loss and the stock price ... 39
4.2.3 The effect of Board independence on the relationship between a goodwill impairment loss and the stock price ... 42
4.2.4 The effect of Board activity on the relationship between goodwill impairment losses and the stock price ... 46
4.2.5 The effect of financially literate audit committee members on the relationship between goodwill impairment losses and the stock price ... 51
5 Conclusion and Discussion ... 52
6 Further research and limitations... 54
7 Bibliography ... 55
8 Appendix ... 59
8.1 Appendix I - Variable Definitions ... 59
8.2 Appendix II – Scatterplot All Sample ... 60
8.3 Appendix III – Scatterplot Impairment Sample ... 61
8.4 Appendix IV – Scatterplot Non-Impairment Sample ... 62
8.5 Appendix V – Regression results by year ... 63
5
1 Introduction
In order to improve the quality of financial reporting the IASB issued a new accounting standard in 2005. The revised International Financial Reporting Standard (IFRS) 3 Business Combinations is part of a joint effort by the International Accounting Standards Board (IASB) and the US Financial Accounting Standards Board (FASB) to improve financial reporting while promoting the international convergence of accounting standards (IASB, IFRS 3 Business Combinations , 2015). The IASB states, in the introduction of IFRS 3, that the main objective of the IFRS is to enhance the relevance, reliability and comparability of the information that an entity provides in its financial statements about a business combination and its effects. Among other things, IFRS 3 determines how an acquirer has to account for goodwill.
Goodwill is an asset that represents the future economic benefits arising from other assets acquired in a business combination that are not individually identified and that are separately recognized (IASB, IFRS 3 Business Combinations , 2015). Since goodwill represents future economic benefits, an impairment of goodwill may signal a decline in those future economic benefits. This study will examine the effect of the impairment of goodwill on the stock market to test whether investors perceive the information about impairments as a reliable indication of the future economic benefits of the firm and whether investors act on the stock market with this knowledge in mind. Based on a linear robust regression of 1.121 firm-year observations I do not find support for the first hypothesis that goodwill impairment losses are negatively associated with a firm’s stock price.
Furthermore, this study will test whether investors estimate how discretion is used in determining the amount of the goodwill impairment loss. Since determining goodwill impairment losses involves making an estimation about the future profitability of cash-generating units there is managerial discretion involved. This discretion can be used by managers to convey their private information about future economic performance or it can be used opportunistically. If the discretion is used to convey private information of management it can be considered as value relevant for investors and a reaction on the stock market is expected. If the impairment is used opportunistically, it is not value relevant and therefore I expect that investors will eliminate it from their expectations if we assume that investors are sophisticated enough to detect this opportunistic behavior.
6 Prior studies have found evidence on the likelihood that managers will engage in earnings management in several situations. Companies that have experienced a recent change in CEO are associated with a significantly higher reported amount of goodwill impairment losses (Zang, 2008; Lapointe-Antunes et al, 2008; Abughazaleh et al, 2008; Masters-Stout et al, 2008). Since the impairment-only approach within IFRS 3 offers an opportunity to manage earnings and there is a higher likelihood that recently appointed CEOs use this opportunity to manage earnings, I expect that investors will place a lower weight on the value-relevance of the impairment loss when it is recognized in the transition period of the CEO.
Corporate Governance can constrain managers in acting opportunistically. One important factor is the independence of the supervisory board. Independent board members have a positive effect on the quality of financial reporting and lower the likelihood of earnings management (Abughazaleh et al, 2011; Davidson et al, 2005; Baysinger and Butler, 1985; Beasley, 1996; Peasnell et al, 2005). An impairment loss that better reflects underlying economics is more informative to investors and therefore I expect that the relationship between goodwill impairment losses and a firm’s stock price is stronger when a firm has a higher proportion of outside directors since that impairment gives a more accurate indication of a decline in future economic benefits.
Another important factor is the activity of the board. Since boards that meet more often can devote more time in their monitoring function they may be better in monitoring the decisions of management (Xie et al, 2003; Valfeas, 1999; Conger et al, 1998; Abughazaleh et al, 2011). With help of the increased monitoring of the board, I expect that investors will perceive the goodwill impairment loss as more value-relevant.
Finally, earnings management and opportunistic behavior can be controlled and/ or constrained by an audit committee with financially literature members. According to Xie et al (2003), this is because relevant experiences and training are required to understand earnings management. Therefore financially literature audit committee members are associated with lower levels of earnings management or restatements (Carcello et al, 2006; Abott et al, 2004; Bedard et al, 2004). When audit committee members, who are better able to detect opportunistic behavior, monitor decisions of management with regard to the impairment of goodwill this may result in impairments that better reflect underlying economics. Therefore, I expect that the relationship between goodwill impairment losses and a firm’s stock price is stronger when at least one of the firm’s audit committee members is financially literate. After performing several regressions I
7 conclude that none of the hypotheses and expectations in this sections are supported. Thus, I conclude that goodwill impairment losses are not value relevant to investors, even when indicators of good corporate governance are present.
This study contributes to the extant empirical research on goodwill and asset impairments in several ways. The main contribution of this study is that I examine whether investors perceive the impairment of goodwill as more value-relevant when several corporate governance characteristics are involved and as less value-relevant when managers are incentivized to engage in earnings management.
Secondly, I examine the relationship between goodwill impairments and stock prices in a setting where IFRS is obliged instead of SFAS. This is done on the basis of a larger sample consisting more years of observations after the transition to IFRS 3 than other studies on IFRS 3.
At last, I provide further evidence on the perceived ability, by the stock market, of the supervisory board and the audit committee in constraining opportunistic managerial behavior.
The remainder of this paper is organized as follows. The next section describes a literature review of prior studies related to the IFRS standards, goodwill impairments and managerial discretion. The section ends with the development of the hypotheses. Section 3 describes the data collection and research methodology of this study which is followed by an elaboration of the results in section 4. At last, in section 5 there is a conclusion and discussion of the study and in section 6 I will discuss the possibilities of future research and the limitations of this study.
8
2 Literature review and hypotheses development
In this section I explain why my study is interesting and what the current state of knowledge is with regard to the subject of my study. Thereby I make use of research that is performed on both SFAS 142 and IFRS 3 since they both require an impairment only approach in which goodwill can only be impaired and which prohibits the use of an amortization technique.
2.1 IFRS 3 – Intangible assets and goodwill
Since 2005, accounting for goodwill has changed for companies that report under International Financial Reporting Standard (IFRS). In the introduction of the revised IFRS 3, the International Accounting Standards Board (IASB) states that the main objective of the IFRS is to enhance the relevance, reliability and comparability of the information that an entity provides in its financial statements about a business combination and its effects. Therefore it determines, among other things, how an acquirer in a business combination recognizes and measures the goodwill acquired in the business combination or a gain from a bargain purchase (IASB, IFRS 3 Business Combinations , 2015).
A business combination is a transaction or other event in which an acquirer obtains control of one or more businesses (IASB, IFRS 3 Business Combinations , 2015). According to defined terms in appendix A of IFRS 3, goodwill is an asset representing the future economic benefits arising from other assets acquired in a business combination that are not individually identified and separately recognized.
According to the IFRS 3 standard, goodwill arising in a business combination is measured as a residual at the date of acquisition and is recognized in the statement of financial position as required by IFRS 3.32. This standards requires that goodwill is in principle measured as the difference between the purchase consideration and the fair value of the identifiable net assets (Hamberg, M., & Beisland, L. A., 2014). After initial recognition, an acquirer shall measure and account for the assets acquired, liabilities assumed or incurred and equity instruments issued in a business combination in accordance with other applicable IFRSs for those items, depending on their nature (IASB, IFRS 3 Business Combinations , 2015). Therefore, the IASB eliminates the use of the pooling of interest method and prohibits the amortization of goodwill. Instead of this, it requires an acquirer to test the goodwill for impairment in accordance with IAS 36. This should be done by comparing its recoverable amount with its carrying amount annually, and whenever there is an indication that the goodwill may be impaired at the end of reporting period (IASB, IFRS
9 3 Business Combinations , 2015). Furthermore, after recognition shall an impairment loss recognized for goodwill not be reversed in a subsequent period (IASB, IFRS 3 Business Combinations , 2015). IAS 38 Intangible Assets prohibits the recognition of internally generated goodwill. Since any increase in the recoverable amount of goodwill in the periods following the recognition of an impairment loss is likely to be an increase in internally generated goodwill, rather than a reversal of the impairment it is prohibited to increase goodwill after an impairment test. (IASB, IFRS 3 Business Combinations , 2015).
Because goodwill is an asset that represents future cash flow arising from other assets, it should be allocated to each of the acquirer’s cash-generating units that is expected to benefits from the synergies of the combination for the purpose of impairment testing (IASB, IFRS 3 Business Combinations , 2015). A cash-generating unit is the smallest identifiable group of assets that generates cash inflows that are largely independent of the cash inflows form other assets or groups of assets (IASB, IAS - 36 Impairment of Assets, 2015). An impairment loss shall be recognized for a cash-generating unit if the recoverable amount of the unit is less than its carrying amount.
The recoverable amount of a cash-generating unit is the higher of a cash-generating unit’s fair value less costs of disposal and its value in use (IASB, IAS - 36 Impairment of Assets, 2015). In some situations it is not possible to measure fair value less costs of disposal because there is no basis for making a reliable estimate of the price at which an orderly transaction to sell the cash-generating units would take place between market participants under current market conditions (IASB, IAS - 36 Impairment of Assets, 2015). In that case, the entity can use the cash-generating units’ value in use as its recoverable amount which is largely an entity-specific measure according to the KPMG International Standards Group (2014). To determine the value in use, the following elements shall be reflected in the calculation:
(a) an estimate of future cash flows the entity expects to derive from the cash-generating unit;
(b) expectations about possible variations in the amount or timing of those future cash flows;
(c) the time value of money, represented by the current market risk-free rate of interest; (d) the price for bearing the uncertainty inherent in the assets; and
10 (e) other factors that market participants would reflect in pricing the future cash flows the entity expects to derive from the cash-generating unit (IASB, IAS - 36 Impairment of Assets, 2015).
2.2 Estimates in the recoverable amount
As explained in the section above, when determining the value in use, an entity is expected to make estimations about the future economic benefits of the cash-generating unit. Since this requires managerial discretion this could incentivize managers to overstate, understate, or not recognize an impairment loss depending on the reporting incentives (Abughazaleh, N., Al-Hares, O., & Roberts, C., 2011). KPMG performed a survey to consider the impact of IFRS requirements for various line of business companies and, in some sectors, it is shown that most of the companies in that sector use the value in use measure to determine the recoverable amount of a cash-generating unit (KPMG International Standards Group, 2014). Standard setters suggest that managers will use the accounting discretion permitted by the impairment-only approach to provide their private information about future cash flows according to AbuGhazaleh and Haddad (2012) but the discretion can also be used opportunistically (Hamberg, M., & Beisland, L. A., 2014). As Ramanna (2008) points out, managers can use two earnings management techniques to manage reported earnings under SFAS 142 with help of the unverifiable discretion in the impairment rules. They may either avoid timely goodwill write-offs and thereby overstate current earnings and net assets or overstate goodwill impairment losses and thereby understate current earnings and net assets.
2.2.1 Agency theory and the estimates in the recoverable amount
Watts (2003), as cited by Jarva (2009), states that the agency theory predicts that managers will use discretion permitted by SFAS 142 in circumstances where they have agency-based motives to do so. An agency relationship is defined as a contract under which one or more persons (the principal) engage another person (the agent) to perform some service on their behalf which involves delegating some decision making authority to the agent (Jensen, M. C., & Meckling, W. H., 1979). In the case of an impairment of goodwill there is some reason to believe that managers will not always act in the best interest of the principal since an impairment could indicate a decline in future economic benefits or that too much has been paid for an acquisition. When managers use the discretion opportunistically and the impairment does not reflect underlying economics, I do
11 not expect the impairment of goodwill to be value relevant. On the other hand, in cases where managers use the accounting discretion to provide their private information about future cash flows, I expect the impairment of goodwill to be value relevant since it potentially signals a decline in future cash flows.
2.3 The reliability of goodwill numbers
The FASB believes that investors are better able to assess future profitability and cash flows with help of the impairment only approach (Jarva, 2009). This suggests that whether or not goodwill impairments are informative to investors depends on the relationship of the impairment with expected future cash flows. Jarva (2009) states that goodwill impairment losses could have a predictive ability for expected future cash flows or could lag behind the economic impairment of goodwill in which case it does not reflect underlying economics. This opinion of Jarva is shared by the KPMG International Standards Group. Although they did not perform a statistical survey, KPMG performed a study in which they interviewed stakeholders from various backgrounds and geographies about the relevance and effectiveness of impairment testing. The interviewees in their study said that although goodwill impairment testing is relevant in assessing how well an investment has performed, its relevance to the market is in confirming rather than predicting value because of the degree of subjectivity in goodwill impairment testing. Furthermore, they found that analysts indicated that they would like to see enhanced disclosure about goodwill impairments and more consistency in impairment-related disclosure to make companies more comparable (KPMG International Standards Group, 2014). As indicated before, on the contrary of this survey, much academic research have found evidence on the reliability of the goodwill balances and the impairment loss of goodwill.
Jarva (2009) studied whether firms manage fair value estimates in determining goodwill impairment losses. Thereby he assessed the reliability of unverifiable estimates by examining the relationship between goodwill impairments and the realized future cash flows in a sample of firms listed on the NYSE, AMEX, and NASDAQ between the years 2002 and 2005. He found strong evidence that the goodwill impairments required under SFAS 142 are related to one- and two-year ahead cash flows and therefore have significant predictive ability for those cash flows. The higher the goodwill impairments in his sample, the lower the expected future cash flows of that company and this suggests that firms do not manage the unverifiable fair value estimates. The three-year-ahead cash flow however is insignificantly associated with the goodwill write offs. The
12 insignificance suggests that the impairments of goodwill do not reflect underlying economics and therefore that markets, at least partially, seem to anticipate impairments before the announcement of the impairment. Although this is consistent with the view that the goodwill write-offs lag behind the economic impairment of goodwill, the coefficient is likely to contain significant measurement error. Jarva furthermore tests for a lower association between goodwill impairments and expected future cash flows for firms with contemporaneous restructuring, as a proxy for big bath accounting, but did not find any results that are significantly different from zero. This indicates that goodwill write-offs from restructuring firms contain very little or no information about future cash flows (Jarva, 2009). On the basis of the evidence Jarva concluded that the goodwill impairment losses are on average more closely related to economic factors than opportunistic behavior which could signal an increase in the value-relevance of the goodwill impairment losses.
Jarva (2009) and both the IASB and FASB suggest that the impairment only-approach with regard to goodwill offers managers an opportunity to convey their private information to the public. In the next section, I will elaborate on the perceived reliability of the goodwill impairments by investors as I discuss several studies that examined the effect of goodwill impairments on stock performance.
2.4 The impact of goodwill impairments on stock performance
As indicated in the defined terms of IFRS 3, goodwill represents future economic benefits and an impairment does therefore signal a decline in future economic benefits. As described in the previous section, goodwill numbers are informative to investors under SFAS 142. To test the value relevance of the impairments and to determine whether managers use the accounting discretion opportunistically or to convey private information about future cash flows, several studies examined the relationship between the goodwill impairment under IFRS 3 and SFAS 142, which are both impairment-only regimes, and the stock performance.
To extend the study of Jarva (2009) on whether firms manage fair value estimates in determining goodwill impairment losses, Ahmed and Guler (2007) test the effect of SFAS 142 on the reliability of goodwill balances and goodwill impairments. They state that SFAS 142 could either improve or deteriorate the reliability of goodwill impairments and goodwill balances. Compared to SFAS 121, under SFAS 142 goodwill impairments should be more reliable and informative because requiring firms to use fair values, instead of undiscounted cash flows, lowers the threshold for recognizing an impairment loss. Furthermore the reliability should increase
13 because of the higher frequency of impairment testing. On the other hand, the considerable subjectivity in assigning goodwill to reporting units, the subjectivity in determining the fair value or the value in use, and the complexity in determining such fair values are critics that suggest a lower reliability of goodwill write-offs under SFAS 142 (Ahmed, A. S., & Guler, L., 2007). The authors found a significant negative association between stock returns and goodwill impairments. This indicates an improvement of the reliability of goodwill impairments. Moreover, they found a positive and significant association between goodwill balances and stock prices in the post-SFAS 142 period. This suggests that investors assign higher valuation weights to goodwill after SFAS-142 relative to before SFAS SFAS-142. This is evidence is consistent with the view of Jarva (2009) that goodwill numbers are informative to investors under SFAS 142.
To examine the assumption of the IASB that managers will use the accounting discretion to provide their private information on future cash flows, Li et al (2011) examined among other things the information content of the impairment losses on goodwill. They expected that managers will use the goodwill impairments to convey private information to the public and that investors will use these announcements. Using a sample of firms that announced goodwill impairment losses during the period of 1996-2006, they found evidence that suggests that both investors and financial analysts revise their expectation downward on the announcement of an impairment loss. Furthermore, they state that the impairment losses serve as a leading indicator of a decline in future profitability. In their research, they compared the price impact of the impairment losses for three periods, the pre-SFAS-142, post-SFAS-142, and the transaction period. Their results were weaker for the post-SFAS-142 period but they also indicated that the weaker results are less likely the result of the market perceiving the use of greater managerial reporting discretion after SFAS 142 and more likely due to the smaller magnitude of post-period losses and their repeated occurrence during the post-SFAS-142 period. (Li, Shroff, Venkataraman, Zhang, 2011). This provides further evidence that suggests that the goodwill impairments are informative to investors and financial analysts after SFAS 142.
AbuGhazaleh et al (2012) also tested the assumption that managers will use the accounting discretion to provide their private information about future cash flows. They used a sample, within an IFRS setting, of 528 firm-years observations drawn from the top 500 UK listed firms for financial years 2005 and 2006. They found a significant negative association between equity market values and goodwill impairment losses. This suggests that these impairments are perceived
14 by investors to be value relevant and that they are incorporated in their firms valuation assessment. Moreover, the evidence indicates that managers are more likely to convey their private information about future cash flows through IFRS 3 (Abughazaleh, Al-Hares & Haddad, 2012).
This value relevance is also studied by Xu et al (2011). They studied firms that recognized goodwill impairment charges between 2003 and 2006 and they also found a negative stock price reaction on the goodwill impairment loss. They based their research on the thought that if markets are efficient, stock prices incorporate information about impairment but if the charge reveals additional information price reactions will occur. In their research they separated the sample into groups according to financial health and their findings suggest that goodwill impairments convey a value-decreasing signal. However, there is a difference in the perception of goodwill. For profitable firms, the goodwill impairment is perceived as a strong negative signal but this is less for loss firms. Goodwill impairment is significantly negative for profit firms but insignificantly different from zero for the loss firms (Xu, Anandarajan, Curatola, 2011). This indicates that investors react more severely on an impairment loss when it involves a profitable firm and less severely when it involves a less profitable firm or a loss firm. The authors conjecture the moderator effect of financial health reflects a belief that management is taking action to revalue assets as part of an overall strategy to improve the operating performance of the firm and/ or that the market views the impairment as a signal that the firm is taking action to address value deterioration.
To assess the value relevance of goodwill balances instead of the afore mentioned goodwill impairment losses, Jennings et al (1996) examined the relationship between accounting goodwill numbers and equity values in a sample of firms in the United States during the period 1982-1988. They found evidence that the purchased goodwill represents expected cash flows in the view of investors. This is consistent with the defined terms of IFRS 3 that goodwill is an assets that represents future economic benefits and that it can be seen as an economic resource. Therefore, if the amount of goodwill is impaired, it could signal a decline in the expected future cash flows which could result in a fluctuation in stock performance as shown by prior research mentioned above.
To extend the work of Jennings (1996), Henning et al (2000) examined the valuation of the components of purchased goodwill. Consistent with Jennings et al (1996) the authors found evidence of the association between goodwill balances and stock prices in a sample of firms in
15 the period between 1990 and 1994. Furthermore the authors found that investors attach different weights to different components of goodwill.
Godfrey and Koh (2001) extended the study on the relevance of capitalized intangible assets to firm valuation by examining the explanatory power of goodwill on the market value of equity. They found, consistent with Henning et al (2000) and Jennings (1996), a significant relationship between the market value of equity and goodwill. The significance of the coefficient indicates that investors seem to incorporate capitalized goodwill in their valuation of firms and that goodwill is at least measured with some reliability (Godfrey, J., & Koh, P. S. , 2001)
The prior studies indicated that goodwill balances and goodwill impairments may contain relevant information to investors since it could indicate a decline in future economic benefits. Less reliable on other hand is the important role of managerial discretion in determining the amount of the goodwill impairment loss. In the following section, I will elaborate on the effect of certain characteristics on the timing and amount of the goodwill impairment loss.
2.5 Moderator relationship of the use of discretion on the direct relationship between goodwill impairment and stock performance
As the agency theory predicts, managers may use accounting discretion opportunistically in determining the recoverable amount of a cash generating unit in circumstances where they have agency-based motives to do so (Watts, 2003).
The technical summary of IFRS 3 Business Combination states as its core principle: “An acquirer of a business recognizes the assets acquired and liabilities assumed at their acquisition-date fair values and discloses information that enables users to evaluate the nature and financial effects of the acquisition”. Since this is the core principle, managers’ failure to impair goodwill can attribute to the information asymmetries between managers and stakeholders, particular in situations in which managers have favorable private information on future cash flows (Ramanna, K., & Watts, R. L. , 2012). Some academic studies examined the relationship between goodwill numbers and several characteristics of a firm which predicts managers’ use of discretion in determining the goodwill impairment loss.
Beatty’s and Webers’ (2006) study examined the effect of several potentially important economic incentives that firms face when making the accounting choice to recognize a goodwill impairment loss in the SFAS 142 transition period. Specifically they investigated the outcome of
16 the managers’ goodwill allocation and reporting unit decisions by examining the determinants of the SFAS 142 write-off decisions (Beatty & Weber, 2006). This is done by investigating several situations in which firms prefer above-the-line versus below-the-line accounting treatment and in which firms prefer to delay or accelerate expense recognition. According to Watts and Zimmerman (1990), accounting choices are affected by firms’ debt contract. Beatty and Weber (2006) continue on that thought and examined whether the accounting choices differ when debt covenant calculations are affected by mandatory accounting choices, as in this case SFAS 142. They found that firms with debt covenants that will be affected by accounting changes are less likely to record goodwill impairment charges and that they will record relatively smaller chargers than firms with debt covenants that are not affected by accounting changes. Moreover, the authors examined whether equity market valuation affects the goodwill impairment choice. Since markets react differently on impairments that are caused by accounting changes and impairments that are caused by a decline in fair values, the authors expected that high risk firms with great differences between undiscounted and discounted cash flows are more likely to record goodwill impairment charges and will record relatively larger charges. This choice is motivated by the expectation that markets are more likely to accept the charge caused by the accounting change of SFAS 142. Furthermore, since an impairment of goodwill may signal that the acquisition price was too high, the accounting choice to impair goodwill will also depends on whether the CEO made the original acquisition decision. The authors found that recently appointed CEOs are more likely to take impairment charges and will record relatively larger charges. Finally firms with listing requirements have incentives to take accounting actions that help them avoid being delisted. The authors found that firms listed on an exchange with financial-based listing requirements are less likely to record a goodwill impairment charge and record relatively smaller charges.
To conclude, the results indicate that firms’ equity market considerations affect their preferences for above-the-line versus below-the-line accounting treatment, and firms’ debt contracting, turnover, and exchange delisting incentives affect their decisions to accelerate or delay expense recognition. This suggest that the recognition of impairment losses are affected by both contracting and market incentives (Beatty & Weber, 2006).
Ramanna and Watts (2012) extended the work of Beatty and Weber (2006) by studying firms’ agency based motives to delay goodwill impairment losses between the transition year and subsequent years. To test for contracting motives, the authors used a proxy for the cost of violating
17 debt covenants as the product of the ratio of current period debt to prior period assets and found significant results. Moreover, they tested for managers’ accounting-based compensation as an indicator for whether a firms’ CEO received a cash bonus during the year. Ramanna and Watts state that such a bonus is usually tied to net income. Since the net income of the year decreases by recognizing an goodwill impairment loss, the bonus could incentivize managers to delay or decrease the impairment of goodwill. Also for this this hypothesis they found that the proxy for CEO compensation is a significant negative predictor of goodwill impairment. In this line, managers could also have reputational motives to delay or decrease the goodwill impairment loss (Ramanna, K., & Watts, R. L. , 2012). To proxy for reputation loss Ramanna and Watts used, consistent with Beatty and Weber (2006), CEO tenure since CEOs with longer tenures are more likely to have been involved in the acquisitions that resulted in goodwill. The results indicate that CEOs with a longer tenure are less likely to take goodwill impairment losses in order to avoid reputation loss.
Overall, the evidence suggests that managers, on average, use the unverifiable discretion in SFAS 142 to avoid timely goodwill write-offs in circumstances where they have agency-based motive to do so (Ramanna, K., & Watts, R. L. , 2012). This study indicates that managers, in some circumstances, use the accounting discretion in determining the recoverable amount opportunistically. Since this does not improve the relationship between goodwill impairment losses and future cash flows, the value relevance of the goodwill impairment is expected to be lower in these circumstances and this results in a lower association between goodwill impairment losses and stock performance.
AbuGhazaleh et al (2011) also examined managers’ use of the discretion in determining goodwill impairment losses offered by IFRS 3. They conducted their study in the UK with a sample of 528 firm-year observations over the period of 2005 and 2006 and their results suggests that managers are exercising discretion in determining goodwill impairment losses. The results furthermore indicate that the impairment losses are strongly associated with effective governance mechanisms and thereby suggesting that managers are more likely to be exercising their accounting discretion to convey their private information and expectations about the underlying performance of the firm than to be acting opportunistically. The results collectively suggest that, in recording goodwill impairments within IFRS 3, managers are likely responding to changes in economic circumstances and real declines in the value of the firm. Effective governance
18 mechanisms are likely to restrict managers’ ability to report goodwill impairments that differ from predicted economic losses, resulting in the recognition of more timely impairments that better reflect the firm’s underlying economics (N. Abughazaleh et al., 2011).
On the basis of these studies, I can conclude that managers are in some situations incentivized to use the goodwill impairments opportunistically depending on their agency-based motives. However, as Abughazeleh et al (2011) state, corporate governance mechanisms are likely to constrain managers in this opportunistic behavior.
2.6 Hypotheses development
In this section, I will elaborate on the hypotheses. These are based on the findings of the studies discussed in the previous sections as well as complemented with additional literature. First, in section 2.6.1, I will hypothesize the direct effect of goodwill impairment losses on stock prices. In section 2.6.2, I will integrate the use of discretion in the hypotheses to examine the effects of several characteristics on the direct relationship between the goodwill impairment losses and stock prices.
2.6.1 The effect of a goodwill impairment loss on the stock price
As mentioned in section 2.1, goodwill is an asset representing the future economic benefits and therefore several academics (Jennings et al, 1996; Henning et al, 2000; Gofrey and Koh, 2001) have studied the relationship between stock price and goodwill. They all found that goodwill is significantly associated with the stock price and goodwill balances are value relevant to investors. Within the setting of SFAS 142, under the impairment-only regime, Xu et al (2011), Ahmed and Guler (2007), and Li et al (2011) studied a direct relationship between the goodwill impairment losses and stock price. They all found that investors deteriorate the expectations of the firm on the impairment of goodwill by means of a decline in stock price. Furthermore, Lapointe-Antunes et al (2009) also reported a negative and significant association between transitional goodwill impairments losses and stock prices. AbuGhazaleh et al (2012) examined the same relationship but in a setting of IFRS and found a significant and negative relationship as well. This suggests that investors act on the goodwill impairment loss and that it is integrated in their valuation assessment of the firms in an European setting as well. This evidence leads to the following first hypothesis.
19 H1: Goodwill impairment losses are negatively associated with a firm’s stock price.
2.6.2 Proxies for managerial discretion
Although several studies, Abughazaleh et al 2011; Beatty & Weber , 2006; Ramanna & Watts, 2012; Jarva, 2009, examined an effect of the managerial discrection on future cash flows or the timing and magnitude of goodwill impairment losses, none has examined the effect on the stock price. In this section, I hypothesize that several characteristics moderate the relationship between goodwill impairment losses and stock prices.
2.6.2.1 Management change
If new managers overstate the goodwill impairment losses in the transitional period of management change they can reduce the possibility of having to recognize a future impairment losses that declines operating earnings. Goodwill impairment charges may be perceived by the market as a signal of overpayment for the acquisition and that could harm the reputation of existing management. Therefore they are reluctant to accept goodwill impairment charges (Zang, 2008) and (Ramanna, K., & Watts, R. L. , 2012). On the other hand, new managers have incentives to overstate the goodwill impairment because this reduces the probability of having to recognize future goodwill impairments at the cost of their operating earnings. Furthermore, the impairments during the transition period can be assigned to the poor decisions of the former management (Zang, 2008), (Lapointe‐Antunes, P., Cormier, D., & Magnan, M. , 2008). To test these predictions, Zang (2008) conducted a study with a US sample during the SFAS 142 transition period. This provided evidence that the amount of the initial goodwill impairment loss is significantly greater for firms that experienced a management change than for firms that did not.
Lapointe-Antunes et al 2008 examined reporting incentives and constraints associated with the magnitude of transitional goodwill impairment losses in a Canadian setting. Consistent with Zang (2008), they found that firms that experienced a change in CEO recognize higher transitional goodwill impairment losses. Although both studies examined the effect of management change during the transition period to a new standard, they both found that these changes are associated when higher goodwill impairment losses. Still the accounting rules for goodwill in IFRS and SFAS offer an opportunity for managers to manage earnings.
Abughazaleh et al (2011) also expected and examined whether firms that experienced a recent change in CEO are more likely to report higher amounts of goodwill impairment losses.
20 Consistent with Zang (2008) and Lapointe-Antunes et al (2008), they found evidence that supports this expectation.
Masters-Stout et al (2008) extended the study on the subject of CEO change by not constraining the period to only the transition period. They examined the relationship between the tenure of CEOs of publicly held companies and goodwill impairment losses. Specifically they examined whether the impairment-only approach under SFAS 142 provides a mechanism for earnings management. They found that both the new internal and external CEOs, that became CEO in the last 2 years prior to reporting, impaired significantly more goodwill than their senior counterparts. Moreover, the authors found that new CEOs are more inclined to take a big bath since their results indicate that the impaired goodwill reduces when the net income is positive and that it increases when the net income is negative.
Since IFRS 3 offers the same opportunity to manage earnings with regard to the impairment of goodwill, I expect that investors will perceive impairment losses made by newly appointed CEOs as less value-relevant as opposed to CEOs with a longer tenure. This leads to the following hypothesis:
H2: The relationship between goodwill impairment losses and a firm’s stock price is weaker when a firm has a recent CEO change.
2.6.2.2 Board independence
Independent board members may improve earnings quality by reducing managerial self-interest and by monitoring and controlling decisions managers make in the production of financial statements (Alves, 2014). Beasley (1996) adds that, from an agency perspective, the ability of the board to act as an effective monitoring mechanism is dependent upon its independence of management.
Abughazaleh et al (2011) and Davidsion et al (2005) state that a higher proportion of independent directors results in a lower likelihood that a firm engages in accruals-based earnings management. In their studies they found a significant positive relationship between board independence and reported non-opportunistic goodwill impairment losses. This indicates that a higher proportion of independent board members results in a higher likelihood that firms report higher amounts of non-opportunistic goodwill impairment losses. This is consistent with the idea
21 that shareholder welfare is enhanced by boards of directors which are capable of monitoring management by means of independent judgement on the performance of management. Ceteris paribus, firms with more independent boards should perform better (Baysinger, B., & Butler, H., 1985). Furthermore, since independent and outside directors are perceived as valuable for the external markets when they perform well in their function, they have incentives to develop reputations as experts (Beasley, 1996). Beasley examined whether the proportion of independent board members is lower for firms that experienced financial statement fraud. He found a significant negative relation between the proportion of outside board members and financial statement fraud indicating that outside directors are important monitors of management and that they constrain management in fraudulent actions. Peasnell et al (2005) also tested whether the influence of outside directors extends to the financial reporting process and they found that firms with a higher proportion of outside board members are associated with less income increasing earnings management, consistent with the view of the other academics in this section.
Since a more non-opportunistic goodwill impairment loss, free from earnings management, is more relevant to investors, I expect that investors will place a higher weight on the goodwill impairment loss when a company has a higher proportion of outside directors. This results in the following hypothesis:
H3: The relationship between goodwill impairment losses and a firms’ stock price is stronger when a firm has a higher proportion of outside directors.
2.6.2.3 Board activity
In the previous section, I discussed the importance of the monitoring function of the board of directors on management with regard to the reporting of non-opportunistic goodwill impairment losses. The board activity is also an important dimension of the monitoring function of the board of directors.
Xie et al (2003) state that a board that meets more often should be able to devote more time in their monitoring function. Therefore this may be an important factor in constraining the tendency of managers to engage in earnings management. To proxy for earnings management, the authors use discretionary accruals and in their study they found that the number of board meetings is negatively associated with discretionary accrual. Vafeas (1999) tested the association between board activity and corporate performance. His view is that more board meetings are more
value-22 relevant for shareholders since directors in boards that meet more frequently are more likely to perform their duties in accordance with shareholders’ interests. He found that operating performance rises following years of abnormally high meeting frequency. This view is shared by Conger et al (1998) who states that to make effective decisions, directors need sufficient, well-organized periods of time together as a group. At last, Abughazaleh et al (2011) report that board activity, measured by the number of meetings by the board of directors, is related to a higher amount of non-opportunistic goodwill impairment losses. Therefore the frequency of meetings is an important resource of improving the effectiveness of a board. The evidence of prior studies leads to the following hypothesis:
H4: The relationship between goodwill impairment losses and a firm’s stock price is stronger when a firms’ board is more active.
2.6.2.4 Financially literate audit committee members
There are several reasons to believe that the effectiveness of the audit committee is influenced by the relative status of the audit committee (Badolato, Patrick G., Donelson, Dain C., & Ege, Matthew , 2014). They state that, first of all, audit committees need both ability and authority to gain respect of managers and to influence financial reporting outcomes. To be able to monitor such a complex issue as a goodwill impairment, certain financial expertise is required. Secondly, audit committees with higher relative status are likely to be more active monitors of management since they have the ability to ask the right questions and to be more effective in questionable financial reporting (Badolato, Patrick G., Donelson, Dain C., & Ege, Matthew , 2014). Since determining goodwill impairment losses requires managerial discretion, an audit committee with a higher relative status may be more effective in assessing the accuracy and reliability of the impairment.
Xie et al (2003) also state that the audit committee may have a direct role in controlling earnings management since its function is to monitor a firms’ financial performance and financial reporting. They state that an audit committee member should be financially literate because an active, well-functioning, and well-structured audit committee may be better able to prevent earnings management. Since relevant experience and training is required to understand earnings management, financial literate audit committee members may be better in monitoring a firms’ financial performance and financial reporting. In their US sample the authors found a significant negative relationship between their proxy for earnings management and the proportion of outside
23 investment bankers in the audit committee. Therefore they concluded that an audit committee that has members with some financial background may better serve as a financial monitor. Carcello et al (2006) also examined whether audit committee financial expertise mitigates earnings management. They believe that financial experts are trained and experienced in recognizing accounting manipulation and that they are therefore better able to mitigate earnings management. In their US sample they found evidence that the presence of at least one financial expert in the audit committee is associated with a lower level of earnings management. Consistent with Carcello et al (2006), Abbott et al (2004) and Bedard et al (2004) found that audit committees with at least one financial expert are associated with less restatements and earnings management.
The evidence of the prior studies suggests that audit committees with at least one financial expert are more successful in their monitoring function of management and the accounting decisions of management. When those audit committees monitor the decisions of management with regard to the impairment of goodwill, this may result in impairments that better reflect the underlying economics instead of opportunistic behavior. The following hypothesis tests whether investors react stronger on goodwill impairments of firms with financially literate audit committee members.
H5: The relationship between goodwill impairment losses and a firm’s stock price is stronger when at least one of the firm’s audit committee members are financially literate.
3 Research design
At first I will explain how I selected the sample for the study. Thereafter I will explain the model that I used and I continue with a description of the variables used to test the hypotheses.
3.1 Sample
To conduct this study, I selected data from the financial database Datastream. Within, Datastream I collected data from firms listed in the Netherlands, Belgium, France, and Germany. Since these countries are all civil law countries they are comparable for the purpose of this study. Thereby, I only focused on equities that are traded in the currency Euro on the Euronext.liffe Amsterdam, - Brussels, - Paris, and Bourse Frankfurt during a period starting at 01/01/2005 and ending on 31/12/2014. The resulted in 24.090 firm-year observations. Table 1 provides an overview of the
24 sample selection. Following prior research (e.g. Abughazaleh, N., Al-Hares, O., & Roberts, C., 2011; Abughazaleh, Al-Hares & Haddad, 2012; Jarva, 2009; Jennings, R., Robinson, J., Thompson, R. B., & Duvall, L. , 1996; Ahmed, A. S., & Guler, L., 2007) 11.070 firm-year observations classified as observations in the Financial Services sector are excluded since their financial reporting processes tend not to match with those of other industries.
Firm- Year Observations
Equities traded in the currency EUR on Exchange Markets (with DS Mnemonic) of Euronext.liffe Amsterdam (H), Euronext.liffe Brussels (B), Euronext.liffe Paris (F), and Frankfurt (D) from the years 2005-2014
24.090
(-) Observations related to the Financials industry 11.070
(-) Observations with zero Market Value of Equity or missing information 1.916
(-) Observations with no positive goodwill balances 3.540
(-) Observations with Badwill 80
(-) Observations with not available goodwill impairment losses 6.299
Final Sample 1.185
(-) Outliers (See table 2 for a specification) 64
Final Sample for statistics 1.121
Goodwill impairers 695
(62%)
Non goodwill impairers 426
(38%)
Observations with regard to board independence 398
Observations with regard to the number of board meeting 380
Observations with regard to the Audit Committee 407
Observations of years with an effect by the change of CEO 131 Table 1: Sample Construction
With help of the database Datastream I derived an overview of all the equities traded on the above mentioned exchange markets and excluded those that belong to the sectors banks, financial services, life insurance, nonlife insurance, real estate investments, and real estate investment trusts. The market-value of equity was not available for 1.916 firm-year observation, 3.620 firm-year observations had no positive goodwill balance or a badwill balance, and there was no information
25 available on the goodwill impairment of 6.299 firm-year observations. Hence, this resulted in a final sample of 1.185 firm-year observations. As table 2 specifies, 64 outliers are detected and excluded from the sample. Observations are expected to be outliers when they deviate more than three standard deviations from the mean of the variable. This resulted in a final sample ready for statistical analyses of 1.121 firm-year observations. Although the initial sample is significantly narrowed down, it remains larger than the sample used in prior studies to evaluate the effect of an impairment loss on the share price (Abughazaleh, Al-Hares & Haddad, 2012; Abughazaleh, N., Al-Hares, O., & Roberts, C., 2011; Van Hulzen, P., Alfonso, L., Georgakopoulos, G., & Sotiropoulos, I., 2011). The final sample sample consists out of 695 impairment and 426 non impairment firm-year obeservations. 398 observations are available with regard of the independence of the supervisory board, 380 observations with the number of board meeting held during the year are available, and 407 observations provide information about whether or not the audit committee consists out of at least one financial expert as defined by the SEC. Finally, there are 622 firm-year observations available with regard to the CEO of the company and 131 of these observations are affected by a change of the CEO. Using database Orbis I selected information about the directors of firms in the Netherlands, Belgium, France, and Germany. Finally, the company Saint-Gobain S.A. (ISIN: FR0000125007) had no amount of outstanding shares available for scaling but had all other information available. Therefore, I hand-collected the number of common shares outstanding at year-end and the book-value of those shares from the annual reports of Saint-Gobain S.A..
Table 2: Descriptive Statistics before exclusion of outliers and process of excluding outliers
Variables M SD Lower bound Upper bound Exclusion
Market Value of Equity scaled
by total assets 0.6708847 0.5614277 -1.0133984 2.3551678 16
Book Value of Equity scaled by
total assets 1.59541 13.75364 -39.66551 42.85633 9
Net income scaled by total
assets 0.0424497 0.092017 -0.2336013 0.3185007 21
Carrying value of Goodwill
scaled by total assets 0.1907511 0.1522044 -0.2658621 0.6473643 9
Impairment of Goodwill scaled
by total assets 0.0108054 0.0392821 -0.1070409 0.1286517 9
26
3.2 Variables and measurement
In this section I will explain the variables that I used to test my hypotheses together with the model that I used. Since the number of observations differs for several variables, each hypothesis or construct is explained separately.
3.2.1 The Ohlson (1995) model
The following model is used to implement the analysis:
𝑀𝑀𝑀𝑀𝑀𝑀 = 𝛽𝛽0+ 𝛽𝛽1𝐵𝐵𝑀𝑀𝑀𝑀𝑖𝑖𝑖𝑖+ 𝛽𝛽2𝑁𝑁𝑁𝑁𝑖𝑖𝑖𝑖+ 𝜀𝜀𝑖𝑖𝑖𝑖 where:
MVE = firm i’s market value of equity at the end of the year in which the goodwill impairment test is performed scaled by the total assets at the end of that year; BVE = firm i’s book value of equity at the end of the year in which the goodwill
impairment test is performed minus the carrying value of goodwill at the end of that same period scaled by the total assets outstanding at the end of that year; NI = firm i’s pre-tax profit at the end of the year in which the goodwill impairment test
is performed plus the reported goodwill impairment loss, recognized in that year, scaled by the total assets at the end of that year.
The model is based on the work of Ohlson (1995) and is called a market valuation model. Ohlson (1995) developed this model to evaluate whether investors use accounting information. This is based on the idea that investors use accounting information in the in their valuation of the company and in their decision making when the accounting information is relevant (Van Hulzen, P., Alfonso, L., Georgakopoulos, G., & Sotiropoulos, I., 2011). The basic Ohlson model formulates the market value of equity as a function of the book-value of equity and net income.
The market value of equity is measured by multiplying the share price with the number of ordinary shares issued at the company’s year-end. The book value of equity is calculated by multiplying the book value at the company’s fiscal year end by the number of common shares outstanding. Finally the pre-tax profit represents all income and losses before any federal, state, or local taxes.
27 3.2.2 The effect of a goodwill impairment loss on the stock price
In addition to the Ohlson (1995) model I added two variables that allow me to test the value-relevance of the carrying value of goodwill and the impairment of goodwill. The remaining variables are consistent with those in the Ohlson (1995) model in section 3.2.1. Therefore the following model is used to implement the analysis:
𝑀𝑀𝑀𝑀𝑀𝑀 = 𝛽𝛽0+ 𝛽𝛽1𝐵𝐵𝑀𝑀𝑀𝑀𝑖𝑖𝑖𝑖+ 𝛽𝛽2𝑁𝑁𝑁𝑁𝑖𝑖𝑖𝑖+ 𝛽𝛽3𝐶𝐶𝑀𝑀𝐶𝐶𝐶𝐶𝑖𝑖𝑖𝑖+ 𝛽𝛽4𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖+ 𝜀𝜀𝑖𝑖𝑖𝑖 where:
CVWG = firm i’s carrying value of goodwill at the end of the year in which the goodwill impairment test is performed plus reported goodwill impairment loss of that year, scaled by the total assets at the end of that year;
IMP = firm i’s reported goodwill impairment loss of the year reflected as a positive number scaled by the total assets at the end of that year. IMP is 0 for firms that do not report a goodwill impairment loss.
Abughazeleh et al (2012) and van Hulzen et al (2011) extended the model of Ohlson (1995) by including variables that measure the carrying value of goodwill at the end of the reporting year and the goodwill impairment loss recognized during the year. According to van Hulzen et al (2011) a major advantage of the Ohlson model is that extra variables can be easily added to the equation. When those variables are value relevant to investors, the variable will be significantly associated with the market value of equity and the model will explain more of the market value of equity.
The carrying value of goodwill represents the excess cost over the fair market value of the net assets that are purchased. The calculation of goodwill under IFRS 3 is explained in section 2.1. The impairment of goodwill represents all impairments of goodwill during the fiscal year. It excludes impairment losses for goodwill arising from a transitional impairment test since this is reported as a change in accounting principles.
3.2.3 The effect of management change on the relationship between a goodwill impairment loss and the stock price
To test whether the effect of a goodwill impairment loss on the stock price is weaker when the firm experienced a recent change in CEO I added a dummy for the change in CEO and an interaction variable in addition to the model used to test the effect of a goodwill impairment loss on the stock price. This resulted in the following model:
28 𝑀𝑀𝑀𝑀𝑀𝑀 = 𝛽𝛽0+ 𝛽𝛽1𝐵𝐵𝑀𝑀𝑀𝑀𝑖𝑖𝑖𝑖+ 𝛽𝛽2𝑁𝑁𝑁𝑁𝑖𝑖𝑖𝑖+ 𝛽𝛽3𝐶𝐶𝑀𝑀𝐶𝐶𝐶𝐶𝑖𝑖𝑖𝑖+ 𝛽𝛽4𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖+ 𝛽𝛽4𝑀𝑀𝑀𝑀𝑁𝑁𝐶𝐶𝑀𝑀𝑀𝑀𝑁𝑁𝐶𝐶𝑀𝑀𝑖𝑖𝑖𝑖
+ 𝛽𝛽5𝑀𝑀𝑀𝑀𝑁𝑁𝐶𝐶𝑀𝑀𝑀𝑀𝑁𝑁𝐶𝐶𝑀𝑀𝑖𝑖𝑖𝑖× 𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖 + 𝜀𝜀𝑖𝑖𝑖𝑖 where:
MANCHANGE = a dichotomous variable equal to 1 if firm i experienced a change in CEO in t, t-1 or t-2, and 0 otherwise;
MANCHANGE X IMP = an interaction variable which multiplies the dummy variable of management change for firm i with firm i’s reported goodwill impairment loss of the year reflected as a positive number scaled by the total assets at the end of that year.
3.2.4 The effect of Board independence on the relationship between a goodwill impairment loss and the stock price
In this section I will explain the addition to the model used to test the effect of a goodwill impairment loss on the stock price. To examine whether the relationship between goodwill impairment losses and a firms’ stock price is stronger when a firm has a higher proportion of outside directors, the following model and variables are used:
𝑀𝑀𝑀𝑀𝑀𝑀 = 𝛽𝛽0+ 𝛽𝛽1𝐵𝐵𝑀𝑀𝑀𝑀𝑖𝑖𝑖𝑖+ 𝛽𝛽2𝑁𝑁𝑁𝑁𝑖𝑖𝑖𝑖+ 𝛽𝛽3𝐶𝐶𝑀𝑀𝐶𝐶𝐶𝐶𝑖𝑖𝑖𝑖+ 𝛽𝛽4𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖+ 𝛽𝛽6𝐵𝐵𝑁𝑁𝑁𝑁𝐵𝐵𝑀𝑀𝐼𝐼 + 𝛽𝛽7𝐵𝐵𝑁𝑁𝑁𝑁𝐵𝐵𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖× 𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖 + 𝜀𝜀𝑖𝑖𝑖𝑖
where:
BINDEP = the percentage of independent directors calculated by dividing the number of independent directors by the total number of board members;
BINDEP X IMP = an interaction variable which multiplies the percentage of independent directors of firm i with firm i’s reported goodwill impairment loss of the year reflected as a positive number scaled by the total assets at the end of that year.
3.2.5 The effect of Board activity on the relationship between goodwill impairment losses and the stock price
To test whether the relationship between goodwill impairment losses and a firm’s stock price is stronger when a firm’s board is more active, the following model and variables are used:
29 𝑀𝑀𝑀𝑀𝑀𝑀 = 𝛽𝛽0+ 𝛽𝛽1𝐵𝐵𝑀𝑀𝑀𝑀𝑖𝑖𝑖𝑖+ 𝛽𝛽2𝑁𝑁𝑁𝑁𝑖𝑖𝑖𝑖+ 𝛽𝛽3𝐶𝐶𝑀𝑀𝐶𝐶𝐶𝐶𝑖𝑖𝑖𝑖+ 𝛽𝛽4𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖+ 𝛽𝛽8𝐵𝐵𝑀𝑀𝐶𝐶𝐵𝐵𝑁𝑁𝑀𝑀𝑁𝑁𝐵𝐵𝐵𝐵𝑖𝑖𝑖𝑖+ 𝛽𝛽9𝐵𝐵𝑀𝑀𝐶𝐶𝐵𝐵𝑁𝑁𝑀𝑀𝑁𝑁𝐵𝐵𝐵𝐵𝑖𝑖𝑖𝑖
× 𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖 + 𝜀𝜀𝑖𝑖𝑖𝑖 where:
BACTIVITY = the number of meetings held by the supervisory board during the financial year;
BACTIVITY X IMP = an interaction variable which multiplies the number of board meetings by the supervisory board of firm i with firm i’s reported goodwill impairment loss of the year reflected as a positive number scaled by the total assets at the end of that year.
3.2.6 The effect of financially literate audit committee members on the relationship between goodwill impairment losses and the stock price
Finally, to assess whether the relationship between goodwill impairment losses and a firm’s stock price is stronger when at least one of the firm’s audit committee member is financially literate, the following variables are added to the model used to test the effect of a goodwill impairment loss on the stock price:
𝑀𝑀𝑀𝑀𝑀𝑀 = 𝛽𝛽0+ 𝛽𝛽1𝐵𝐵𝑀𝑀𝑀𝑀𝑖𝑖𝑖𝑖+ 𝛽𝛽2𝑁𝑁𝑁𝑁𝑖𝑖𝑖𝑖+ 𝛽𝛽3𝐶𝐶𝑀𝑀𝐶𝐶𝐶𝐶𝑖𝑖𝑖𝑖+ 𝛽𝛽4𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖 + 𝛽𝛽10𝐹𝐹𝑁𝑁𝑁𝑁𝐹𝐹𝑁𝑁𝐵𝐵𝑖𝑖𝑖𝑖+ 𝛽𝛽11𝐹𝐹𝑁𝑁𝑁𝑁𝐹𝐹𝑁𝑁𝐵𝐵𝑖𝑖𝑖𝑖 × 𝑁𝑁𝑀𝑀𝐼𝐼𝑖𝑖𝑖𝑖 + 𝜀𝜀𝑖𝑖𝑖𝑖
where:
FIN LIT = a dichotomous variable equal to 1 if firm i has an audit committee with at least three members and at least one financial expert within the meaning of Sarbanes-Oxley;
FIN LIT X IMP = an interaction variable which multiplies the dummy variable FIN LIT of firm i with firm i’s reported goodwill impairment loss of the year reflected as a positive number scaled by the total assets at the end of that year.
4 Empirical results
In this section the results are discussed. This is started by the descriptive statistics and continues with the multivariate results. In this section I also provide tables with results and in the appendix
30 I display scatterplots of the data, additional results that help to understand my data, and an overview of the variables definition.
4.1 Descriptive Statistics
Table 3 provides descriptive statistics for the continuous variables used in the multiple robust regression examining the Ohlson model (1995) and the value relevance of the carrying value of goodwill and the goodwill impairment loss of the entire sample. Panel A covers the sample of both the firms that reported a goodwill impairment loss and the firms that did not report an impairment loss. It consists out of 1,121 firm-year observations. Panel B provides descriptive statistics for the continuous variables used in the multiple robust regression of the impairment sample. This panel only includes firms that reported a goodwill impairment loss. Furthermore panel C provides descriptive statistics for the continuous variables used in the regression for the non-impairment sample and therefore only contains observations of firms that did not report goodwill impairment losses during the year. All variables are divided by the reported total assets.
The mean of the market value of equity in the all sample is 0.6285799. Initial evidence on a decline in market value upon a goodwill impairment loss is provided by a lower mean of the market value of equity in the impairment sample, which is 0.5642152. Furthermore, firms that did not report a goodwill impairment sample have an average market value of equity, scaled by total assets, of 0.7335880. This mean is higher than the mean of the big sample and the impairment sample.
The same pattern is shown for the book value of equity and the pre-tax income. The all sample provides a mean of the book value of equity (pre-tax income) of 0.3959171 (0.0475590), the impairment sample provides the lowest mean of 0.2491882 (0.0429050), and the non-impairment sample provides the highest mean of 0.6352988 (0.0551517). This suggests that firms that did not report a goodwill impairment loss perform also better in terms of pre-tax income than firms that did report a goodwill impairment loss.
Firms in the impairment sample report the highest carrying value of goodwill with a mean of 0.1900533. On the contrary, firms that did not report a goodwill impairment loss report the lowest carrying value of goodwill with a mean of 0.1700025. Since the carrying value is adjusted for the reported goodwill impairment loss of that year, this could be of interest for the effect of a goodwill impairment loss on the stock price.
31 The number of meetings by the supervisory board is approximately equal for the three samples with on average 8 meeting during the fiscal year. Probably of interest for hypothesis 3, the average percentage of independent board members is lowest in the impairment sample (60.82%) and highest in the non-impairment sample (81.08%).
In the impairment sample 81 of the 381 observations are affected by a change in CEO while in the non-impairment sample 50 of the 241 observations are affected. Finally, in the impairment sample, 218 firms that have an audit committee with at least one financial expert reported an impairment while 63 firms reported the impairment without having at least one financial expert in the audit committee. 97 of the 126 firms that did not report a goodwill impairment loss had an audit committee with at least one financial expert.
Table 4 provides a pairwise correlation between the variables used to test the models. This matrix provides additional evidence for a positive relationship between the market value of equity on the one hand and the book value of equity and pre-tax income on the other hand since both are significantly and positively correlated.
Although the carrying value of goodwill is positively correlated to the market value of equity it is in contrast to the predictions not significantly correlated. On the other hand, consistent with the predictions, the impairment of goodwill in significantly and negatively correlated to the market value of equity. This suggests that a higher impairment of goodwill is correlated with a lower market value of equity. Furthermore, a change in CEO is negatively correlated with the market value of equity but not significantly. Contrary to the prediction in the hypotheses development is that an increase in the number of meetings held by the supervisory board is significantly and negatively correlated to the market value of equity. Lastly, an audit committee with at least one financial expert is negatively correlated with the market value of equity but this correlation is not significant.
All variables used for a proxy of good corporate governance, independence of the board, activity of the board, and at least one financial expert in the audit committee are positively correlated with impairment but none of the correlations is significant. This suggests that a higher percentage of independent board members, a higher frequency of board meetings, and the presence of at least one financial expert results in a higher likelihood of a firm reporting a goodwill impairment based on this correlation matrix.
