The effect of goodwill impairment on cumulative abnormal return : are goodwill impairments associated with CAR for companies in the United States during 2009-2016?

1

The effect of goodwill impairment on cumulative abnormal return

Are goodwill impairments associated with CAR for companies in the United States during 2009 - 2016?

Name: Robbert Flens Student number: 11420784 Thesis supervisor: Dennis Jullens Date: June 25, 2018

Word count: 13689, 0

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

Statement of Originality

This document is written by student Robbert Flens 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

The association between goodwill impairment and cumulative abnormal return (CAR) is explored in this study for United States listed companies during the seven-year period from 2009 to 2015. The purpose of this study is to examine whether there is a significant relation between goodwill impairment and CAR. The CAR is a measurement of the concept. The relation is multiple tested by CEO change, risk rating and board independence to conclude if these variables change the outcomes of the association. Moreover, the market reaction for goodwill impairment is one of the objectives for this thesis. The research approach of this study is quantitative. The literature review which describes the Sloan (1995) model is used as a basis to explain the relationship. The results show that goodwill impairment is negatively associated with CAR. Also, the correlation analysis of CEO change showed a decrease of 1,22%. The risk rating decreases with 0,7% and the board independence 0,6%. The outcomes of these three variables are all significant. Based on this thesis we can conclude that goodwill impairments are informative in the determination of the stock price. Furthermore, the goodwill impairment information will be used by investors to evaluate the stock price. All in all, the relationships are relatively strong because of critical low probability value. The regression results resulted in the acceptance of the different hypothesis of this thesis.

Acknowledgements

I would like to thank my supervisor Dennis Jullens for all the reviews and supporting during this thesis. I also wanted to thank Lucas Trentelman from EY who gave me a lot of tips for writing this thesis. I received a lot of knowledge and guidance of them to successfully complete my thesis. Further, my parents help me with reviewing this thesis and know it is more readable. I am grateful for all the encouragement that everyone has given to me. Without all this support I was not capable of completing this challenging thesis.

5

Content

2.2 Interest from science 12

2.3 Variables 16

2.4 Association CAR and impairments 16

2.5 Research question and hypotheses development 19

3 Research design 23

3.1 Sample 23

3.2 Model and variables 25

3.2.1 Effect of CEO change 25

3.2.2 Effect of board of director’s independence 26

3.2.3 Effect of risk rating 26

3.2.4 Adjusted model 27

4 Results 28

4.1 Correlation 28

4.2 H1 test: relation between goodwill impairment and CAR 30 4.3 H2 test: relation between goodwill impairment, CAR and CEO change 31 4.4 H3 test: relation between goodwill impairment, CAR and risk rating 32 4.5 H4 test: relation between goodwill impairment, CAR and outside public board 34

4.6 Summary main findings 35

5 Conclusion and limitations 36

5.1 Discussion 37

5.2 Confounding events 39

1 Introduction

IFRS 3 was released in 2008 and was a project between the Financial Accounting Standards Board (FASB) and the International Accounting Standards Board (IASB). IFRS 3 includes the accounting regulation for goodwill impairment. The purpose of this new standard was to increase comparability between US GAAP and IFRS regulation. Due to this new standard there was no longer goodwill amortization but there are impairments made for goodwill. In IAS 36 is defined how to account for goodwill. This is to ensure that an entity’s assets are not carried at more than their recoverable amount. IAS 36 requires entities to conduct an annual impairment test where there is an indication of decreasing cash inflows. They succeed in improving the convergence between the standards and improve to clarify the accounting standard.

In 2012, Rio Tinto came in the news because the mining company was forced to take $8.8 billion impairment charge. The recent acquisition of the aluminum company Alcan leads to disappointments by investors. Rio Tinto paid $44 billion to buy the Canadases company Alcan in June 2007. This acquisition seeks to improve the performance of the division. The goodwill impairment resulted in a significant stock price decrease (Financial Times, 2012).

Furthermore, due to scandals of Enron and Worldcom the pressure on the regulatory institutions has increased. The trust of the economic sector is negatively affected by this events and the effects are still touchable during these days. For example, the new IFRS standard of 2008 regarding the goodwill impairment is released to rise comparability between US GAAP and IFRS regulations. The crisis from 2008 has serious dramatically consequences for the goodwill amounts on the balances. There are during this period a lot of goodwill impairments made to adjust the balance with the right economic future benefits value. Giner & Pardo (2015) concluded that goodwill impairment has a lack of relevance and was driven by opportunistic decisions during and prior to the crisis. From 2002 the goodwill impairment became mandatory to annually test for deviations. Goodwill impairment becomes more important to management, auditors and investors because of the new regulation and the increasing size of goodwill (Hayn & Hughes, 2006). The big disadvantage for investors is that companies are using the goodwill impairment for earnings management and therefore abuse this opportunity. The impairments are mostly subjective and this leads to general assumptions about the future economic benefits. Further, is it difficult for all of the concerned parties to estimate the right value.

This study focusses on the association on goodwill impairment on the stock price of a company for United States listed companies. The underlying assumption is that investors look

7 at the company’s financial statements and conclude that a goodwill impairment has negative consequences for the future economic benefits. This reaction of the investors is reflected in the stock price and leads to a decrease. To better understand this relation, the CEO change will also be included. It is expected that a CEO will increase the impairment amount in his first year to show a lower profit. In the next year the CEO therefore do not have to make this impairment. If there are lower profits shown in a particular year, the stock price will also be lower. The next hypothesis looks again at the relation between goodwill impairment and stock price, but this time with special focus to the risk rating of a company. The risk rating of Standard & Poor reflects if the company is capable of meeting his financial commitments to others. The best rating is an A and the lowest rating is a D. In this study we assume that investors weighted the goodwill impairment less strong with an A-rating, because these companies are likely to fulfill their financial commitments. The other hypothesis looks into the outside public board members of the board of directors.

This study contributes to prior literature and practices in several ways. First, the research creates a better understanding of the relation between goodwill impairment and stock prices. The valuation to goodwill impairment information will be examined. Secondly, the influences of CEO change, risk rating and outside public board members will be included to this association for listed firms in the United States. The sample is used from 2009 to 2015 to take the after crisis period. The crisis has led to bankruptcies and we do not want to include this data. During a crisis period there are expected to be more impairments compared to other periods. However, based on the model of Sloan (1995) the results will be developed. Due to the different hypothesis each of the different variables are examined and multiple conclusion will be made for a more comprehensive picture of the goodwill impairment association on stock price. This research covers all the United States listed firms and therefore the sample will contain enough observations to show significant results. This study can help investors for better investment decision by paying attention to certain components of the financial statements. Finally, with the knowledge of this study the regulation can be improved for goodwill accounting. The knowledge accumulation is important because goodwill is a very considerable item in the financial statements.

The results of this thesis show the negative association of goodwill impairment on the CAR. The relation is divided by three other variables and show significant results. The correlation analysis of CEO change showed a decrease of 1,22%. The risk rating decreases with 0,7% and the board independence 0,6%. The results will be further explained in the result

paragraph. We can learn from these results that investors evaluate all of the variables as important to measure the stock price.

This thesis starts with literature review of goodwill impairment studies. The different perspectives and variables will be described. The research question and hypothesis development will also be clarified. The next chapter depicts the research design where the sample, model and variables are explained. Chapter four shows the results with regard to all the different hypothesis, while chapter five gives the conclusion and finalize it with a discussion of the limitations and suggestions for next researches.

9

2 Literature review

In this chapter I make a reference to relevant literature from prior research. This chapter explains why this study is interesting and why it is valuable. Section 2.1 includes a further understanding about the goodwill and goodwill impairment process. In section 2.2 I explain the different perspectives of the influence of goodwill impairment. Section 2.3 explains the variables which are used. Section 2.4 includes the association of stock price and goodwill impairments. And finally in section 2.5 I describe my research question and hypotheses development.

2.1 Accounting perspective

Goodwill is an asset which represent the future economic benefits. The following definition of goodwill is originally formulated by the IASB (2005): “An asset representing the future economic benefits arising from other assets acquired in a business combination that are not individually identified and separately recognized.” In 2005 a new reporting standard was established for goodwill accounting. It is no longer possible to write off goodwill on a straight-line basis. Instead, there is a mandatory annual impairment test, because the straight-straight-line amortization fails to provide useful information to the users. It only adds noise and therefore it is harder to use earnings measures to predict future profitability. The impairment test may be carried out at any time during a reporting year, but it is necessary that the same date is used every year.

Based on US GAAP, a company needs to perform a two-step goodwill impairment test if it wants to test goodwill impairment. The first step that needs to be established is that the fair value of the reporting unit is compared with its carrying amount including goodwill. If the fair value exceeds the carrying amount there would not be a second step and the goodwill is not impaired. The second step compares the implied fair value with the carrying amount. If the carrying amount exceeds the implied goodwill fair value, there is an impairment loss recognized (FASB, 2018).

Based on IFRS, a company needs to do an annual impairment test where they determine the recoverable amount, which is the highest of fair value and value in use. This value needs to be compared to the carrying amount of the goodwill. There is an impairment made when the recoverable amount is less than the carrying amount. The impairment is included as an expense in the income statement.

Goodwill impairment is a highly frequent studied topic. Goodwill is capitalized after a company acquisition, when the purchase price exceeds the assets of the company. The impairment of goodwill occurs when the fair value of the goodwill goes below its book value. So, goodwill is an asset that presents the future value of the economic benefits which arise from other acquired asset from a business combination. The impairment of goodwill gives a signal to the stakeholders that there is a decline in future economic benefits. When a manager decides to make an impairment of goodwill, there is always a certain information asymmetry. The managers need to decide if they make an impairment and needs to act with his/her managerial discretion. Sometimes the impairment process is used to manage the earnings in a certain way. An example of this situation could be when there is already a loss in the current fiscal year of the company and the managers decide that another goodwill impairment could easily be added to this loss. When a manager has a certain incentive to publish the impairment this has probably a negative impact. Because stakeholders expect that there is something bad going on.

The rules and regulation between IFRS and US GAAP are somewhere different. IFRS is principle based and US GAAP has rule based standards. The advantage of principle based standards is that it increases comparability, verifiability and less discretion. The advantage of rule based standards are general guidance, flexible and substance over form. For both approaches are different advantages. For this thesis there are companies from the United States selected and therefore US GAAP is applicable. ASC 350 of the US GAAP provides relevant guidance how goodwill impairment should be recorded. Goodwill is allocated to a reporting unit. Due to different facts and circumstances a reporting unit is either an operating segment or one level below as an operating segment. The recognition of a goodwill impairment applies when the carrying amount of the reporting unit is greater than its fair value and the carrying amount of goodwill is greater that its implied fair value. Step 1 of the impairment recognition should be followed when the qualitative assessment determines that it is more likely that the carrying amount exceeds fair value. When this is the opposite step 2 needs to be followed. A reversal of the impairment loss is prohibited. The rules and regulation of US GAAP are in several ways different compared to IFRS. For IFRS the goodwill is allocated to a cash generating unit.

Under US GAAP goodwill is managed in the SFAS 141 Goodwill and Other Intangible

Assets and SFAS 142 Business Combinations. The identifiable criterion distinguishes

intangible assets from other assets. There is only goodwill if there is a business combination. The reporting unit is very important for the US GAAP regulation, because the acquired assets and assumed liabilities have to be assigned to this reporting unit. This is mainly different for

11 the IFRS regulations. On June 30, 2001 the standards went into effect. The most important change for this standard was that the business transaction method was chosen by the management. The managerial ability to justify the particular decision for accounting can result that two identic situations of acquisition can be different because of the difference in perspective and motivation of the management. US GAAP regulation negatively affect the judgement of the investors by this standard. Later on, US GAAP came up with improvements for this standard. First, they increased comparability of financial statements due to uniform business combinations treatment and secondly the capture of current market values of intangibles through simultaneous application of purchase method and goodwill impairment tests (FASB, 2001).

The choice for United-States listed companies arises from the fact that the U.S. is comparable to Europe. Prior literature has mainly focussed on the Europe setting. However, this study helps to better understand the U.S. market. It is expected that goodwill impairment strongly correlates with the stock price as we have seen in European settings. The knowledge accumulation of this study can help European investors to better select companies for

investments. If European investors wants to invest in European and U.S. listed firms as well, they can look at the goodwill impairment developments for the concerned companies. This study will help investors for the choice and evaluations of the companies performance.

Purchase price allocation is the allocation of the purchase price of a company to different assets and liabilities acquired from the transaction. The purchase price needs to be allocated to all assets and liabilities that meets the applicable recognition criteria. The value of the acquired assets and liabilities needs to be at fair value. On this way the book value is converted into fair value. The recognition of goodwill is the remaining value that exist from these procedures. The reason behind the purchase price allocation is to increase transparency for the acquisition process. For the acquirer this is mostly a complex matter to account and divide all these amounts. There are four steps for the purchase price allocation: analyze, identify, valuation and accounting. This form of accounting helps investors to better understand what the financial consequences are for the company (KPMG, 2007).

Merger and acquisition is a general term that refers to the consolidation process of companies. There are always two companies involved by a merger and acquisition process. If a company wanted to growth they can do two things. They can do an organic growth or they can growth through acquisitions. This is related to the strategy of the company which of the two is the best option. Trautwein (1990) defines seven different motives for merging. The theories which Trautwein explained are: efficiency, monopoly, raider, valuation,

empire-building, process and disturbance theories. In his event study he found different evidence for all of the acquisitions. Based on this grouping he was able to get better insight into the decisions of the companies acquisitions. During the past 30 years the number of transactions has significantly increases. For example, in 1985 there were approximately 500 transactions. Compared to 2000 there were at least 4,000 transactions. This is a very huge increase which can be linked to the increasing GDP. But not only the transactions have increases, the amount of each transaction have also raised. This resulted in rising demand for attention and caution to successfully handle these transactions.

In summary, goodwill represents the future economic benefits from a company and goodwill impairment is the charge when a goodwill fair value is below its book value. The next paragraphs provide different perspectives from prior researches about this topic.

2.2 Interest from science

Studies which examine the effects of impairments mainly focusing whether or not there is a goodwill impairment. Further, they classify this impairment scaled by the total assets. So the comparability between observations can be better estimated. This classification improves the understandability of the results of prior research.

The stock price of a company could react in three ways on a goodwill impairment. The stock price could increase, decrease or stay the same. In prior research we see this distinguish of the three possible reactions. It is important to understand to what extent a goodwill impairment gives new information. The management of the company could have insight information which can affect the goodwill impairment. Furthermore, the information needs to be timely so it adds value to the investors. Timely information means that the distribution of the information is sufficient in advance so the investors have enough time to review the information. The stock price reflects the present value at discount rate. The discount rate is a percentage which reflects the risk of the company.

The majority of prior empirical studies examining the association of goodwill impairment and stock price, but this results provide little evidence that it is of significant value to users. For example, Henning (2000) examine the association of goodwill impairment and stock price. This research conclude that goodwill impairment is one of the main important factors for the determination of the stock price. He found that there is a decline in the stock price when a recent goodwill impairment has occurred.

On the other hand, AbuGhazaleh (2012) did a research in a IFRS setting where there was no relation found between goodwill impairment and stock price. This means that there is no

13 significant association found between these two variables. This study suggest that the impairments are perceived by investors to reliable measure a goodwill impairment decline and incorporate this in their valuation of firm assessments. Furthermore, managers are more likely to use their accounting discretion to carry out privately held information about the firm underlying performance. Lapointe-Antunes (2009) also found a negative relation between goodwill impairment and stock price. This finding is consistent with investors perceiving losses as being sufficiently reliable measurements of the decline on goodwill. This information is used by investors for their valuation assessments. When a firm is expected to record a loss, investors put a higher valuation weight on the reported losses.

Hirschey and Richardson (2003) found evidence that stock holders anticipate on the goodwill impairment. In the preannouncement period the stock prices drop for approximately 41%. This means that stock holders already take a percentage of the loss before the actual loss is published. In the post announcement period there is further decline for at least 11%. This implies that stockholder strongly react on a goodwill impairment. The Basu (1997) paper, where ‘bad news’ reflect earnings quicker than ‘good news’, is in line with the findings of Hirschey and Richardson. Basu explains that stockholders react stronger on negative news, which can be linked to the findings of Hirschey and Richardson. It simply means that there is a further stock price drop in the post announcement period, because stockholders suspect that the current stock price is overvalued. This leads to a further decline until the stock price reaches market value, which reflects the right price of the stock. This theoretical view can help to understand why stock prices are in the post announcement period affected by the goodwill impairment.

In the case of goodwill impairment there is always negative attitude towards the news. Thus, this implies that stock holders rank bad news higher than good news. Stock holders pays attention to goodwill impairments and therefore there shows up a drop in stock price. If the stock holders think that the current stock price reflects the current market value of the company, the stockholders stop with downgrading the stock price. This theory could possibly be linked to the post announcement period result of the Hirschey and Richardson paper. When the stock holders are not convinced enough about the stock price, they go on with selling their shares until it reaches the right market value. The Basu paper provides fundamental theoretical arguments which can help for a better understanding about the goodwill impairment effect on stock price.

Nichols and Wahlen (2004) mentioned that there is a relation between the goodwill impairments and the stock prices. There is a model developed which includes three

steps/assumptions. The first assumption is that the current results and revenue give the investor information. With this information they can decide whether they want to invest in a company. The second step is that the future information can be reliable linked to the future results and future expectations. The last step is that the stock prices are going to react on the future expectations. When this is the case, we can conclude that the stock prices are going to react which new information for the investors. Prior literature has shown that goodwill impairment news gives new information to the investors. This eventually lead to stock price fluctuations. The future expected results and are going to be adjusted by the investor when the final goodwill impairment is released. When there is a goodwill impairment announcement the information is available for all stakeholders (Li, 2006). Li founds evidence of a significant negative market reaction to goodwill impairment. This negative reaction is greater for firms with a higher proportion of goodwill and also for companies which release little information prior to the impairment.

Duff & Phelps (2012) state that there is a relation between goodwill impairment and stock price during the key recessionary time period from 2008 through 2009. This study looked at 5.000 U.S. listed firms. During that period, goodwill impairment is over $188 billion for U.S. listed companies. Because of this extreme numbers, there was a big drop in 2009 for goodwill impairment for approximately $29 billion. These numbers have an impact on the stock prices of the companies, in particular IT-related companies. Companies like Microsoft are heavily affected by the financial crisis and therefore needs to write-off a high extent of their goodwill. This study contributed to better understanding the effects of the financial crisis on goodwill impairment.

In conclusion, there are two categories which examine the effects of goodwill impairments. Henning provides positive evidence about the association, while AbuGhazaleh found a negative association. Hirschey and Richardson found that stock holders are anticipating on the goodwill impairment. Nichols and Wahlen implies that there is a relation between goodwill impairment and the stock price. The next paragraphs provide the explanation of the above mentioned variables which are used in this thesis.

To summarize the most important findings of prior research I have made a table. In table 1 I have defined an overview of most impairment papers. This summary gives a quick overview of the most important aspects of each individual paper. This overview contributes to the knowledge of prior literature.

15 TABLE 1 - Overview of main findings prior literature

Nr. Authors Period Country Data Period Size Findings 1. Henning et al. 2000 United

States

U.S.-listed firms.

1990-1994

3,097 Goodwill is one of the main important factors for determination of stock price. 2. AbuGhazaleh et al. 2012 Europe Top 500 UK listed firms 2005-2006

746 There is no significant association found between goodwill impairment and stock price.

3. Lapointe-Antunes

2009 Europe Société des Bourses

Françaises Index (SBF 120)

2005 106 There is a negative relation between goodwill impairment and stock price. This finding is consistent with investors perceiving losses as being sufficiently reliable measurements of the decline on goodwill.

4. Hirschey et all. 2003 Europe U.S.-listed companies 1992-1996

80 There is evidence found that stock holders anticipate on the goodwill impairment. In the preannouncement period the stock holders already take a percentage of the loss before the actual loss is published.

5. Nichols & Wahlen 2004 United States NYSE, AMEX and NASDAQ 1988-2002

90,470 There is a relation between goodwill impairments and stock prices. To better understand this relation they developed a model.

6. Li et all. 2006 United States

UK listed firms 2004 2,034 Li founds evidence of a significant negative market reaction to goodwill impairment. This negative reaction is greater for firms with a higher proportion of goodwill and also for companies which release little information prior to the impairment.

7. Duff & Phelps

2012 United States

Duff & Phelps database

2007-2011

5,247 Duff & Phelps (2012) state that there is a relation between goodwill impairment and stock price during the key recessionary time period from 2008 through 2009.

2.3 Variables

The first variable for this thesis is goodwill impairment. Goodwill impairment includes the sum of all impairment of goodwill special items reported after taxes. Special items represent unusual and non-recurring items reported by the company above taxes and are presented as a net figure. This item includes the write-off of goodwill, impairment of goodwill and other intangibles when combined, impairment of unamortized intangibles and positive impairments of goodwill or unamortized intangibles that indicate that the company is reversing a part of the previous charge. The second variable which I use for this thesis is stock price. Stock reflects the current market value of a company. In this thesis I use the daily stock price from the North America database and Canadian fundamental and market information on active and inactive publicly held companies. For example, daily trading prices from the Nasdaq national market are included as well as stock prices from the New York Stock Exchange. Prior research has only focused on European countries to assess this association. Thus, by looking at another sample of data this thesis add evidence to the knowledge accumulation.

In short, there are two variables used for this thesis, namely goodwill impairment and stock price. The next paragraphs provide two models which can be used to measure the two above mentioned variables.

2.4 Association CAR and impairments

The first model which is explained is from MacKinlay (1997). MacKinlay developed a specialized model for the measurement of stock price when a goodwill impairment occurs. MacKinlay used his Cumulative Abnormal Return (CAR) to look into the specific period of time when there has been a goodwill impairment announcement. MacKinlay makes a distinguish between the normal return and the abnormal return. The normal return is defined as the expected return without conditioning on the event taking place. The abnormal return is the actual ex-post return of the event window minus the normal return. He stated that the abnormal return is defined as follows:

ARi,t = Ri,t - E(Ri,t|Xt)

This is a statistical model which focus on the abnormal return. In this model the ARi,t is

the abnormal return, the Ri,t is the actual return and E(Ri,t|Xt) is the normal returns. Xt is defined for the conditioning information.

17 In the following model the cumulative abnormal return can be measured. This model includes all the cumulative values compared to the abnormal return model. CAR is a measurement of a concept. The concept is a construct of the stock prices during the relevant year. The concept measures the fluctuations of the stock prices and relate these results to the goodwill impairment.

CARi,t = yt=x ARi,t

This model includes the daily abnormal return for firm i on date t. The x and y variables are represented for the first day of the period and the last day of the period. The usefulness of this model can be explained because effects of an event are immediately reflected in the stock price. The rationality of the marketplace resulted in this outcome. The use of this event study in this model have been applied to different economy specific firms in accounting and finance research. Examples of these includes earning announcement, new debt or equity issues and mergers or acquisitions. In the majority of researches there is a mainly focus on the effect on stock price. However, the difficulty for the implementation of this model is to determine the right procedures to select the right event window. MacKinlay concludes that there is no unique structure for selecting this. There is only a general flow of analysis which can be followed. For example, when there is a research about information content of a daily earnings database, the event is the earning announcement and the event window is about the specific dates which are included. Usually the window event is larger than the specific period of interest. This can help for the examination of the results of the event window. The period of interest is in practice expanded over multiple days. It includes at least the day before and after the impairment. Thus, on this way the price effects of the impairments are captured. When a broad event window is used, it is easier to get a better picture of what is going on. In this case it is possible to investigate pre-event returns which could influence the stock price reactions. The downside of such a long window could be that there arise a lot noise during that period. If a narrow scope is chosen, there is a less noise which influences the outcome.

Due to some information asymmetry the valuation of a company could be slightly different than the sum of the book value and the net income. Van Hulzen (2011) conducted a research where he found out that investors use the accounting information of the company in their valuation and also in their decision making.

In this thesis I want to add the variable goodwill impairment and CEO change to this model. The goodwill impairment is highly important for the determination of the stock price (Henning, 2000). Stock holders evaluate the goodwill impairments as highly valued objects. They rely on this information to decide if they want to invest in a company. The goodwill impairment could reveal information that there is sometime going on in the company. Researches are trying to explain why a stock price has a certain value. Therefore, a lot of variables are included in models and one of these determinations of the stock price is goodwill impairment. This accounting information highlights the events of an acquisitive company. Furthermore, the CEO change could influence the amount of goodwill impairment (Zang, 2008). Zang found evidence that there are more goodwill impairments taken when there is a recent change in CEO. This is an important development because the goodwill impairment might be a perfect opportunity for a company to write-off some of their subsidiaries. The CEO could argue that this needs to be done because the predecessor CEO did not do a good job. With this knowledge, stakeholders would probably add less value to the impairment losses during a transition period of a CEO. There is a likeliness that recently appointed CEO’s use this opportunity to manage earnings. So investor know that there is a recently changed CEO and they evaluate the goodwill impairment information as less valuable. They know that CEO are trying to take a big bath strategy and therefore they pay less attention to this information.

Sloan (1995) developed a model to estimate if stock prices correctly reflect all publicly available information. This model consists of three variables as leading indicators of goodwill impairments (GWDLIA). This paper is mainly focusing on the short term results. Goodwill impairments are mainly subjective estimated, because companies prepare the financial statements with their own interpretations and assumptions. It is harder for investors to assess the future performance of the company if there is a lack of adequately information. The first variable is the relative magnitude of goodwill balance (GTA). This variable scales the goodwill amount by total assets. Therefore, the goodwill impairment amounts are corrected by the size of the company which gives a better reflection of the goodwill impairment impact. This variable measures the goodwill impairment amount corrected by the scale of the total assets. On this way we can measure the relative impact of this variable. The formula which is used to calculate the GTA is: GTA = goodwill impairment / total assets. The second variable is the return on assets (ROA). ROA is measured as operating income after depreciation divided by total assets. The formula which is used to calculate the ROA is: ROA = operating income / total assets. If there is a deterioration in earnings this indicates that the fair value of firm’s assets is lower. Therefore, accounting earnings should be negatively associated with a future

19 goodwill impairment, because they let both see a decline in the future expected value. The third variable is cumulative abnormal return (CAR), because a lower stocker prices indicate a decline in future cash flows. Stock return and goodwill impairment should be negatively associated when a company makes a loss. These variables can be summarized in the following model of predicting goodwill impairments (Sloan, 1995):

CARt+1 = α + β1GTAt + B2ROAt + B3GDWLIAt + Εt+1

In summary, MacKinlay developed a model to measure the stock price. MacKinlay is mainly focusing on cumulative abnormal returns and Sloan developed a model for the measurement of goodwill impairment. The next paragraphs provide the research question and hypotheses development.

2.5 Research question and hypotheses development

The purpose of this thesis is to examine the association between goodwill impairments and the stock prices for companies in the United States. For this research I have developed the following research question:

Are goodwill impairments associated with CAR for companies in the United States during 2009 - 2016?

In this thesis I estimate the association of goodwill impairment and CAR. I use the CAR as my dependent variable and goodwill impairment as my independent variable. I make a further understanding about the effects of goodwill impairment. This research question is important because prior research has not focussed on the post-crisis period regarding this association with the different related variables, namely CEO change, risk rating and board independence. Prior research has only focussed on one of the mentioned variables and has not included all of these into one study. With this thesis I give a clear and complete overview of the most important literature with has discussed these topics. This thesis studies the effect of the different variables on the same dataset which leads to unbiased results. Each individual variable is measured under the same assumptions.

Stock prices are mostly heavily affected by the company’s profit. The profit of a company reflects the future cash flows that investors use to determine the value of the stock price. This research question is focusing on the period after the Great Recession. The financial crisis

caused a lot of bankruptcies and therefore also goodwill impairments. I exclude this period of my research to get robust findings which are not influenced by the market conditions of that specific period. The end of the Great Recession is chosen for this thesis as starting point so the data sample gives a fair view of reality. Lapointe-Antunes (2009) shows in his research that there was a negative relation in a IFRS setting between the goodwill impairment and stock price. In this thesis there are only companies from the United States included. Prior research has mainly focused on a Europe setting instead of looking to companies from the United States. This lead to the formulation of my first hypothesis:

H1: Goodwill impairment has a negative effect on CAR

The hypothesis above is stated in alternative form. The corresponding null hypothesis is that goodwill impairment has no effect on the CAR.

To make a better understanding of this association I added a corporate governance variable. The corporate governance variable which I use is the recent change of CEO. If a new CEO overstate the goodwill impairment in his transitional period of CEO change they can reduce the possibility that this impairment happens in the future. A goodwill impairment reduces the operating income of a company and therefore this is a bad signal to the stakeholders. A CEO tries to avert all events that lower the operating income. He is incentivized to prevent such decisions. This is because the CEO owns mostly a certain percentage of shares of the company. Abughazaleh (2011) did research about the relation of goodwill impairment and stock price when there is a recent change in CEO. There is a likeliness that recently appointed CEO use this for managing their earnings. Investors are aware of a recently changed CEO and they probably evaluate the goodwill impairment information as less valuable. The recent change in CEO is an important extension to include in my thesis because it can help in better understanding the association of goodwill impairment and CAR. My second hypothesis of this thesis is formulated as follows:

H2: There is a significant weaker association between goodwill impairments and CAR when there is a recent CEO change

Masters-Stout (2008) did further research about this topic. He extended the study on the subject of CEO change by looking at a broader scope compared to the transition period. With this broader scope they evaluate the influence of the regulations requirements. They try

21 to find an answer if law and regulations facilitates influences the goodwill impairment decisions. Moreover, the authors found evidence that CEO are more likely to use a big bath strategy when there is a negative operating income compared to a positive operating income.

H3: There is a significant weaker association between goodwill impairments and CAR when there is low risk rating

The third hypothesis looks into the relation with special focus to the Standard & Poor rating. The impact of a negative rating change is stronger compared to a positive rating change on the stock market (Gonzalez, 2004). Linciano (2004) is more nuanced about this association, because he found that it is strongly dependent on the sector of the company. This hypothesis helps to understand what the influence of the S&P rating does with the stock price when there is a goodwill impairment. This hypothesis mapping the risk extent of the company, because the S&P rating reflects the risk that investor face when they want to invest in this company. The risk explains what the likelihood is that a company cannot fulfill his obligations. This risk can be split into two components: probability of default (PD) and loss given default (LGD). PD explains the chance of failing, while LGD the expected loss during a bankruptcy measures (S&P, 2011). The risk rating is classified into two categories. The first category is the low risk rating which is divided into the A-rating. The second category is the high risk rating and for this thesis we use the B, C and D-rating.

The board of directors is a small group within the company who oversee and make decisions regarding the activities of the company. These decisions are mostly concerning complex tasks which relate to the strategy. The board of directors acts as the party that is between the stockholders and the managers of the company (Hillman & Dalziel, 2003). Outside board members are members who are not executives of the company and have no personal interest in the company. Outside directors are important for the firm’s ability to prevent failures and help to better estimate certain risks. The opportunistic behavior of manager will be eliminated by the outside public board members (Jensen & Meckling, 1976). The monitoring role of the outside board members is expected by stockholders as a crucial task (Dunn & Sainty, 2009). Another important issue with board independence is that the members can be split into three general categories regarding the characteristics of the members. According to Johnson, Schnatterly, and Hill (2013), the first category is demography which refers to the age, education, ethnicity, gender and race. The second category is human capital and looks into the

individual skills and experiences of the board member. The thirds category, social capital, is about the personal relationship of the board member to other firm managers.

H4: There is a significant weaker association between goodwill impairments and CAR when are more outside public board members

In short, the research question is focusing on the goodwill impairment and CAR for United States companies from 2009 – 2016. The first hypothesis takes a closer look into this association and the second hypothesis determines if this association is weaker when a recent CEO change has occurred. The third hypothesis examines the influence of the S&P rating. The next chapter is about the research design where all the information from this chapter is incorporated.

23

3 Research design

First of all, I will explain how I selected my sample for this thesis. Thereafter, I will explain the model that I used and describe the variables which I added to this model to test the hypotheses.

3.1 Sample

To conduct this study, I have selected data form the Wharton Research Data Services (WRDS). Within WRDS, the data from listed firms in the United States are collected. Some companies are established in the United States, but valued at Canadian reporting standards. I excluded these companies, because Canada needs to report on IFRS regulation and the United States needs to report their financial statements in line with applicable US GAAP regulations. WRDS is a database platform and the specific samples comes from the S&P dataset. I select the starting period at 01/01/2009 and ending period on 31/12/2016. This period is chosen because it excludes the financial crisis influences prior to 2009. First of all, this thesis is an event study which means that a statistical method is used to assess the impact of an event on the value of the firm. Secondly, the return on the stock market will be declared by using the designated variables. There are 2.090 observation classified for the final sample regression.

Table 2 - Observations determination

Observations of available goodwill impairment 5.124 (-) Observations with missing goodwill 1.235 (-) Observations with missing stock price 872 (-) Observations with other missing values 831 (-) Correction for outliers 86

(=) Final sample 2.090

Observations related to outside board members 913 Observations related to CEO change 1.235 Observations related to risk 1.297

Notes: This table explains the observations calculation of the final sample.

For this thesis I have merged various datasets to be able to work with the data. The total dataset consists of 5.124 observations, but there were some observations which I cannot use and therefore I exclude some of them. Although the sample is significantly narrowed down, it still remains large enough to be useful for statistical research. The goodwill impairment is the relative magnitude of goodwill. This variable scales the goodwill amount by total assets, which is in line with the approach of Sloan. Therefore, the goodwill impairment amounts are corrected by the size of the company which gives a better reflection of the goodwill impairment impact. Firms with a combination of a high degree of assets and a low return on assets are more likely associated with an overvalued goodwill. Furthermore, the stock price also reflects a percentage change. There are 86 outliers detected, these comes from the top and bottom 1% of the observation. The outside board members’ variable is for the 2.090 goodwill impairments in 881 observations available. There are some observations where there is no information about the outside board members. For the CEO change there are 1.235 observations. The risk variable has 1.297 observations. The risk variable is defined from the Standard and Poor database; the variable has rating variated from A to D. An A-rating is relatively good and associated with a lower risk compared to a D-rating. With an A-rating the company is likely capable of meeting his financial commitments.

To give a better understanding of the goodwill impairment process I will explain how the regressions are made. In WRDS the goodwill impairment amount is given at the end of the book year. For this thesis I am looking at that particular year and take the stock price of the first date of the year. I will compare the change of that stock price with the stock price of the end of the year. For this thesis I will only look at big goodwill impairments. I have defined this when the goodwill impairment is minimal 5% of total assets. The publication of the goodwill impairment in the annual report is taken as leading indicator. To get a good understanding of the stock price change I have calculated a percentage and compare this with the goodwill impairment amounts. For example, when there is a goodwill impairment given at 13/3/2010. I will take the stock price of 1/1/2010 and of 31/12/2010. Because there are only big goodwill impairments I can take this time frame. The investors will reflect their opinion of the goodwill impairment in the stock price and therefore this method can be used to research this relation.

In summary, there are over two thousand observations available for the fundamental testing. For all of the hypothesis there are sufficient observations because of the usable data from WRDS.

25

3.2 Model and variables

As described in my literature review there are mainly two models. MacKinlay used a model to measure the cumulative abnormal return. This model includes all the cumulative values compared to the abnormal return model. Sloan also developed a model to estimate if CAR correctly reflect all publicly available information related to the goodwill impairment. This model includes three variables as leading indicators of goodwill impairments. Goodwill impairments are estimated subjective because companies prepare the financial statements with their own interpretations and assumptions. Investors found it difficult to measure companies’ future performance when there is little information available. Therefore, investors focus more on the goodwill impairment amounts in these situations.

3.2.1 Effect of CEO change

I added the effect of CEO change to the relation of the goodwill impairment on the CAR. This dummy helps to understand if a recent change CEO affect this relation. Prior research found that a CEO will cut his R&D costs and increase his own reward in his last year (Sloan and Dechow, 1991). The corporate governance variable CEO change helps to better understand the association. The recent change of a CEO could result in an overstatement of the goodwill impairment in a transitional period. The goodwill impairment will decrease the operating income and negatively signals to the investors. A CEO is incentives to show goodwill impairment when he just became CEO. In this situation the CEO would not have to make these adjustments in the future. The CEO is stock holder for a percentage of the companies’ shares. However, investors are aware of these developments and incentives of the CEO. Investors will therefore evaluate a goodwill impairment as less important during this period.

When the CEO change coefficient is significant and positive this means that the stock price reacts when there is a recent CEO change. The stock price will increase when a goodwill impairment has occurred during a recent CEO change. We do not expect this result because it would contradict the negative effects of a goodwill impairment. When the CEO change coefficient is significant and negative this means that the stock price negatively reacts. We expect this relation because a goodwill impairment will decrease the goodwill amount. The goodwill amount on the balance sheet informs the stakeholders about the future economic benefits of the subsidiary and this will likely influence stockholders’ investment decisions.

3.2.2 Effect of board of director’s independence

The outside board members reflect the independence of the board of directors of the company. Therefore, I added this dummy to the model of Sloan, to examine if the relation between goodwill impairments and CAR significantly changes. The board of directors oversees the company and future related decisions for the activities. These are complicated tasks and are related to the strategy of the company. The board of directors helps the stockholders and the managers of the company to better align both of their interest. The members do not have a personal interest in the company and therefore they can act as the perfect mediator. They can prevent important failures and will help to better focus on the future performance. Managers of the company are trying to manage earnings for a better presentation of their performance. The outside board members have a monitoring role and also includes the interest of the stockholders into their decisions.

When the board of director’s independence coefficient is significant and positive this means that the stock price reacts when there are outside board members. The stock price will increase when a goodwill impairment has occurred when there are less outside public board members. We do not expect this result because it would contradict the expectations that less outside members improve the trust of stock holders. When the board of director’s independence coefficient is significant and negative this means that there is a negative stock price reaction. We expect this relation because more outside board members can better align the interest of the stockholders with the interest of the managers.

CARt+1 = α + β1GTAt + B2ROAt + B3GDWLIAt + B5BODINDt

3.2.3 Effect of risk rating

Standard & Poor are comparing big U.S. companies with each other and evaluate which companies have a higher risk profile compared to others. This dummy can explain if investors react stronger to companies with a bad rating on a goodwill impairment. The assumption is that investors are more distrustful to bad ranked companies and therefore they will faster get rid of these stocks. The risk rating is an important indicator for investors to evaluate the companies’ performance. The risk that an investor faces when he invests in the company is categorised into different levels. The likelihood that a company cannot fulfil the obligations to the stakeholders is therefore split into certain levels.

When the risk rating coefficient is significant and positive this means that the stock price increases when there is a negative risk rating. In this case the stock price will increase

27 when there is a bad risk rating. We do not expect this result because a bad risk rating would negatively affect the stock price. When the risk rating coefficient is significant and negative this means that the stock price negatively reacts. We expect this relation because a good risk rating is associated with a less strong decrease in stock price when there is a goodwill impairment. Investors are assumed that they pay less attention to a goodwill impairment when they have a good risk rating.

CARt+1 = α + β1GTAt + B2ROAt + B3GDWLIAt + B6RISKt

3.2.4 Adjusted model

In this thesis I will include the dummies CEO change, board independence and risk rating into the Sloan model. This will help to better understand the relation of the goodwill impairment and the stock price because there are more variables taken into account. The first three variables come from the model of Sloan and helps to understand the degree of importance for the goodwill impairment. To find an answer to the research question I will use the model of Sloan. This will lead to the following model:

CARt+1 = α + β1GTAt + B2ROAt + B3GDWLIAt + B4CEOt + B5BODINDt + B6RISKt + Εt+1

Concluding, the model of Sloan includes the relative magnitude of goodwill balance, return on assets and stock price variable. For this thesis the CEO change, outside board members and risk rating variables are added to this model.

4 Results

In this chapter I will present the research results of each hypothesis. I will also give an explanation of the given outcomes.

4.1 Correlation

Table 3 shows the descriptive statistics for the all of the variables. The mean value for GDWLIA was -0.037 and implies that there is a 3,7% decrease for this variable. The CAR, which reflects the annual stock price changes, is 0.141 and means that the 40.308 observations increase 14,1%. We can see that there is an annual increase which is also in line with the assumptions that companies are trying to make money and therefore create more firm value. This outcome is reflected in the increased stock price by stock holders. The CEO change variables seems to have quite much observations, but I have given all the observations a “0” when there was no CEO change in that particular year. This explains the relative high number of observations.

TABLE 3 - Descriptive statistics

Stats GDWLIA CAR ROA GTA CEO OPB RISK

N 2,143 40,308 21,268 29,609 1,272,730 11,334 19,184 Mean -.037 .141 .288 .079 .010 .759 .794 St. dev .059 .738 2.929 .130 .099 1.013 .404 Min -.220 -.902 -6.283 0 0 0 0 P25 -.044 -.231 .051 0 0 0 1 P50 -.009 .029 .203 .011 0 0 1 P75 -.001 .303 .375 .106 0 1 1 Max -.000 2.407 2.274 .807 1 1 1

Notes: This table presents a clear overview of all different variables. The observation count (N), mean, standard

deviation, minimum value, 25% value, 50% value, 75% value and maximum value is given in this overview. This table helps to understand the subdivision of the variables.

29 The Pearson correlation matrix measures the strength and direction of the association that exist between the variables. This summarizes that the relation between goodwill impairment and stock price is with incorporation of the different variables still significant. This implies that the relation between these two is very strong and in all cases more informative when a variable is included. Sloan (1995) included in his model the return on assets and the goodwill amount divided by total asset. The results of these variables are also included in the statistics.

4.2 H1 test: relation between goodwill impairment and CAR

The following observations were made as a result of the empirical research with respect to goodwill impairment and CAR. The graph illustrates that there are 2.090 observations and that the goodwill impairment resulted in a 1,26% average decrease in stock price. The probability value is significant and this means that we can reject the null hypothesis. Goodwill impairment has a significant negative effect on the cumulative abnormal return. This statistical results were expected because a goodwill impairment is associated by investors as bad news. A decline in stock price is therefore logically expected. The mean of goodwill impairment is -0,037 and means that there is an average decline of 3 percent scaled by the total asset of the company. The absolute number of the mean for goodwill impairment is -160.3981. The mean of the stock price change is 0,1407677 and describes the average of all stock prices. The actual stock price has a mean of 29,27961 and summarizes that the average stock price is around the 29.

We see that the CAR decreases with approximately 1,2% when there is goodwill impairment. This was also expected because the selected companies have big goodwill impairments in that particular year. In the next sections there will be investigated if this coefficient changes due to an added variable.

TABLE 4 – Regression results H1 Dependent variable: CAR

Regression results Coefficient Prob. GTA (0.075) 0.417 ROA (0.039) 0.510 GDWLIA (0.012) 0.001*** N Adjusted/Pseudo R2 2.090 0.340

Notes: This table gives the most important values regarding the association of goodwill impairment and CAR. A

P>|t|-value lower than 0.05 indicates that the association is significant. The coefficient indicates the relative change of the association. A negative coefficient means that the stock price will decrease when there is a goodwill impairment. The ***, **, * indicate a significance of 1%, 5% and 10%.

31

4.3 H2 test: relation between goodwill impairment, CAR and CEO change

The second hypothesis examines if the relation between goodwill impairment and CAR differs significantly due to the CEO change. I have given all the observations a “1” when there was a CEO change. When there was no change I give the observations a “0”. On this way it is possible to measure the association for the variable. In table 5 we can see that the probability value of the goodwill impairment and stock price relation is still below the critical p-value of 0.01. This means we can accept the second hypothesis. We see that the coefficient of the relation with the added CEO change variable is now -1,22%. There is a relatively small difference compared to the first hypothesis. For the first hypothesis the coefficient was -1,26%. This means that the stock price reacts less strong when there is a recent CEO change. The relevant p-value is still significant and means that investors value a goodwill impairment less strong when there is a recent CEO change. This could be because the investors know that a new CEO takes a big bath strategy when he is appointed. Therefore, investors know that a new CEO brings expectations and investors anticipate on this. Furthermore, the R-squared is for this relation 3,3% and describes that 3.3% of the population can be explained by the model. We see a small decrease compared to the first hypothesis, because there was the R-squared around the 3.5%. The results are still significant despite the fact that there are now 1.235 observations.

TABLE 5 – Regression results H2 Dependent variable: CAR

Regression results Coefficient Prob. GTA (0.159) 0.582 ROA (0.113) 0.623 GDWLIA (0.013) 0.000*** CEO (0.088) 0.009*** N Adjusted/Pseudo R2 1.235 0.320

Notes: This table gives the most important values regarding the association of goodwill impairment and CAR

with respect to CEO change. A P>|t|-value lower than 0.05 indicates that the association is significant. The coefficient indicates the relative change of the association. A negative coefficient means that the stock price will decrease when there is a goodwill impairment. The ***, **, * indicate a significance of 1%, 5% and 10%.

4.4 H3 test: relation between goodwill impairment, CAR and risk rating

The third hypothesis examines if there a significant weaker association between goodwill impairments and CAR when there is low risk rating. As described in the literature review there are two categories for risk rating. For the determination of the risk I gave all the A-rating a “1” and all the B, C and D-rating a “0”. I have chosen for this subdivision because there are 4.000 observations classified with an A-rating, 13.000 observations with a B-rating, 2.000 observations with a C-rating and 200 observations with a D-rating. This means that there are two main categories, namely the group 1 with all the A-ratings and group 0 with B, C and D-rating. In this way it is possible to see if there is a significant change regarding the goodwill impairments and CAR. In table 6 we see that the relation between goodwill impairment and stock price is still significant. The relevant probability value is now 0.001. This is below the concerned 0.01 p-value, so we accept the third hypothesis. In other words, this extra variable has affected the p-value because the p-value for this hypothesis is a little bit higher. So, this means the relation is less strong compared to the first hypothesis. But the relation has still an acceptable p-value and this means that the risk rating is a good variable to add to this model.

We see that the coefficient is -0,007 for the goodwill impairment and CAR relation with the variable risk rating. This means that the stock price will approximately stay the same when an A-rating is given to a company. Because the stock price will decrease with an average of 0,7% when a big goodwill impairment has occurred during a particular year. Based on this results we can see that the A-rating is very important for investors and they trust in the company’s future. We can conclude that an A-rating is outweighs the relatively big goodwill impairment.

Furthermore, the R-squared is for this hypothesis 7,62%. This percentage explains that because of these three variables 7,62% is explained. This is almost twice the size as it was with the first hypothesis. However, there are less observations classified because not all risk ratings are available. For this hypothesis we have still 1.393 observation, which is quite much. The significant relation could possibly exist because there are only two categories used for the risk determination.

33 TABLE 6 – Regression results H3

Dependent variable: CAR

Regression results Coefficient Prob. GTA (0.159) 0.582 ROA (0.113) 0.623 GDWLIA (0.007) 0.000*** RISK (0.032) 0.017** N Adjusted/Pseudo R2 1.393 0.029

Notes: This table gives the most important values regarding the association of goodwill impairment and CAR

with respect to risk rating. A P>|t|-value lower than 0.05 indicates that the association is significant. The coefficient indicates the relative change of the association. A negative coefficient means that the stock price will decrease when there is a goodwill impairment. The ***, **, * indicate a significance of 1%, 5% and 10%.

To get a better understanding of the risk rating influence I have split the risk rating into the relevant two categories. The regression describes that there is no significant other outcome than the previous regression. However, it is valuable that this distinguish is made because now we can see that the A risk rating is very important and outweighs the negative goodwill impairment. I would like to add to this results that the biggest group of observations was the B-rating. For 1,393 observations there was a risk rating available when there was also a goodwill impairment.

4.5 H4 test: relation between goodwill impairment, CAR and outside public board

The outside board members within the board of directors are mostly overseeing complex tasks which relate to the strategy. As described in the literature review the outside board members are important for a company. These people act between the interest of the stockholders and the interest of the managers of the company. In table 7 we see that the goodwill impairment and stock price have a relevant probability value of 0.018. This is above the significant 0.01 level but below the 0.05 level. This means that we can accept this hypothesis. There are 913 observations which have available data to examines this relation. The R-squared less than 1% and this is critically low. The results show that the outside board members can be used to better understand the relation between goodwill impairment and stock price.

TABLE 7 – Regression results H4 Dependent variable: CAR

Regression results

Coefficient Prob.

GTA (0.120) 0.431

ROA (0.086) 0.712

GDWLIA (0.006) 0.018**

OUTSIDE PUBLIC BOARD (0.009) 0.089*

N

Adjusted/Pseudo R2

913

0.006

Notes: This table gives the most important values regarding the association of goodwill impairment and CAR

with respect to outside public board members. A P>|t|-value lower than 0.05 indicates that the association is significant. The coefficient indicates the relative change of the association. A negative coefficient means that the stock price will decrease when there is a goodwill impairment. The ***, **, * indicate a significance of 1%, 5% and 10%.

35

4.6 Summary main findings

To provide a clear overview of the main findings of this study a summary table is included. This table provides an overview of the most important findings. All of the three hypothesis shows significant results.

Table 4 Summary of main findings Hypothesis Variables Findings

H1 GDWLIA

& CAR

The probability value is significant and this means that we can reject the null hypothesis. Goodwill impairment has a significant negative effect on CAR. This statistical results were expected because a goodwill impairment is associated by investors as bad news.

H2 CEO We can see that the probability value of the goodwill impairment and CAR relation is still below the critical p-value of 0.01. This means we can accept the second hypothesis.

H3 RISK The relation is less strong compared to the first hypothesis. But the relation has still an acceptable p-value and this means that the risk rating is a good variable to add to this model.

H4 OPB We see that the goodwill impairment and CAR have a relevant probability value of 0.018. This is above the significant 0.01 level but below the 0.05 level. This means that we can accept this hypothesis.