• No results found

Mergers and Acquisitions in different industries – A shareholder perspective

N/A
N/A
Protected

Academic year: 2021

Share "Mergers and Acquisitions in different industries – A shareholder perspective"

Copied!
44
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Mergers and Acquisitions in different industries – A

shareholder perspective

Author: Frank van Rossum

Student number: S4665007

University: Radboud University

Date: 30-06-2020

(2)

i

Abstract

The decision to engage in mergers and acquisitions (M&A) is often made to fast. That is proven by the fact that nowadays still a lot of mergers and acquisitions fail. Even though companies worldwide perform about 50.000 mergers per year with a value of 4000 billion, most of these deals still end up as a negative investment. This might be due to investment appetite of CEO’s, but it can also be due to the absence of literature. This research shows that there is a gap in the literature regarding the relation between of industry factors and performance of mergers and acquisitions. The goal of this paper is to provide an overview of variables to consider when a company wants to involve in a merger or acquisition. This research was performed by looking at 1726 deals of listed companies that were completed worldwide across all kinds of industries. This paper tested whether the following industry variables affected M&A performance: (1) Labour intensity, (2) Board diversity, (3) Technology intensity, (4) Profitability, (5) Human capital intensity. We did so by making proxies for these variables, because that was the best way to measure these variables. Furthermore, a distinction was made between cross industry versus within industry deals and cross border versus domestic deals. The results show that there is a positive effect of labour intensity and human capital intensity on M&A performance when it comes to cross industry deals or domestic deals. Furthermore, there is a negative effect of board diversity on cross border deal performance. We did not find a significant effect of profitability and M&A performance. Similarly, we did not find significant results for technology intensity and M&A performance. Concluding, this paper opens up the field of research for the role industry factors when it comes to M&A performance. It shows that there is a gap in industry variables literature, which is partly filled by this research, and that can be filled even more by doing additional research.

(3)

ii

Acknowledgements

This paper is written as an assignment for a master’s thesis. Special thanks go out to dr. Jianying Qiu, drs. Maarten Gubbels and the Radboud University. I wish to thank dr. Jianying Qiu for supervising me during the process of writing my thesis. His feedback helped me improve my research and make the best out of it. Even though communication was more difficult due to Covid-19, he found a way to help me on a distance. I wish to thank Maarten Gubbels from the library team for helping me gathering my data. Gathering the data was more difficult than with firm level variables. Maarten Gubbels helped me gather my data, explore databases and he helped me to create proxies to replace the initial independent variables. Finally, I want to thank the Radboud university for providing knowledge, data and support.

(4)

iii

Content

1. Introduction 1

2. Literature review 2

2.1 Why this research is important 2

2.2 What has been done? 5

2.3 Why industry factors? 5

3. Methodology 6 3.1 Dependent variable 6 3.2 Independent variables 8 3.3 Hypotheses 12 3.4 Conceptual framework 13 3.5 Dataset 14 3.6 Analytical tools 17 4. Results 19 4.1 Overall results 19 4.2 Testing hypotheses 20

4.3 Cross industry versus Within industry M&A 22

4.4 Cross border versus domestic M&A 24

4.5 Additional results 25

4.6 Comparison with literature 27

5. Conclusion 28

6. Discussion and future research 30

7. Reference list 32

Appendix 36

1. Skewness and kurtosis 36

2. Histograms 36

3. Interaction variables 39

4. Control variables 40

(5)

1

1. Introduction

Mergers and acquisitions (M&A) are an important part of investment strategies for large, medium and small companies. Yearly, more than 50.000 mergers or acquisitions are completed. These mergers and acquisitions account for more than 4000 billion dollars annually (Imaa Institute, 2020). Improving the performance of mergers and acquisitions with only 1% will change global M&A earnings drastically. For that reason, it is very important to improve knowledge on how one can make M&A’s successful. Engaging in mergers and acquisitions is a matter of all times. Since companies started investing, mergers and acquisitions were part of the process. The number of mergers and acquisitions that have been completed worldwide, have grown rapidly over the years. In the last 10 years there was a total of 22.47% growth in number of mergers and acquisitions (Statista, 2020). When positive economic times arrive, companies start to engage in mergers and acquisitions even more. Very often, the main reason is that they have capital to invest, that they cannot invest in regular investment opportunities (Jensen M. C., 1986). However, the decision to engage in mergers or acquisitions is often made too fast. The same holds for periods after crises, Grave, Vardiabasis and Yavas (2012) show that when economies start growing again after a crisis, the number of mergers or acquisitions increases rapidly as well. The only periods that have decreased the number of mergers and acquisitions during history, were periods of crises. However, shortly after these periods, the number of mergers or acquisitions increases even more (Statista, 2020). Therefore, this topic is currently even more relevant than ever due to the current Covid-19 crisis and possible following economic effects. A wide variety of companies engage in M&A. CEO’s, shareholders and other stakeholders of these companies have different reasons for engaging in mergers and acquisitions. Possible reasons for engaging in M&A are synergy creation, stimulation of growth, increasing supply chain power and eliminating competition (Stanwick & Stanwick, 2001). This paper will elaborate more on those reasons later.

Historical results show that mergers or acquisitions on average do not pay off. Statistics have proven that the failure rate of M&A is between 70% and 90% (Christensen, Alton, Rising, & Waldeck, 2011). McGabe and Yook (1997) made a small distinction within this observation. They state that cash mergers do pay off when there is access cash flow available and that stock mergers do not pay off on average. Even though high rank managers should know that M&A’s on average often do not payoff, they still engage in mergers or acquisitions. The question that rises is, why? Reasons could be to boost corporate performance or creating fast short-term growth. This is also what was found by Tao, Liu, Gao and Xia (2017). The authors use the signalling theory to show that a merger announcement leads to a short-term performance boost. There might also be a more personal reason. Managers often want to

(6)

2 create or maintain a reputation. That is found in research performed by David Hirschleifer (1993). He found that managers behave in a certain way because of their personal goals. Often, they do this to improve or maintain their own reputation. For example, managers do this by engaging in large investment deals, when there is no positive profitability forecast. In addition, engaging in mergers or acquisitions often leads to higher bonusses for top managers, whereas the punishment for failing is often not that high (Spraggon & Bodolica, 2011). Altogether, all in all enough reasons for top managers to engage in risky M&A deals. But if top managers continue to perform mergers and acquisitions, it is at least important to improve the performance of mergers and acquisitions. One way to do so, is by looking at reasons that mergers or acquisitions went wrong in the past. It is already known that specific managers characteristics can influence M&A performance. For example, in the research by Yang, Ma and Li (2012). The authors show that a specific manager characteristic, namely overconfidence, has an influence on M&A performance. It is also known that integration mechanisms play an important role. For example, the research performed by Bauer and Matzler (2013). They show that the degree of integration play an important role when it comes to making M&A successful. They also show that cultural factors play a role in this process as well. However, there is little known about the forces at an industry level. This paper specifically focusses on the industry side of mergers and acquisitions. The main question that will be answered in this paper is which industry characteristics play a positive or negative role when examining M&A performance? For example, industry profitability might positively influence the performance of M&A. Top managers can use this insight when looking for future mergers or acquisitions to engage in. Therefore, this research is useful for top managers and shareholders in future mergers and acquisitions.

2. Literature review

2.1 Why this research is important

As mentioned before, CEO’s have several reasons for engaging in M&A. These reasons are: synergy creation, stimulation of growth, increasing supply chain power and eliminating competition (Stanwick & Stanwick, 2001). First of all, synergy creation is examined. Recent literature regarding M&A provides information about how M&A can lead to creating synergies (Röller, Stennek, & Verboven, 2001). They found that engaging in mergers and acquisitions offers possibilities to create efficiency gains. However, most companies fail to get the most out of this, because there are other mechanisms influencing this outcome. These mechanisms can be corporate culture, domestic culture, integration mechanisms, managerial characteristics or it could be due to the particular industry characteristics that apply on the target or acquiror. This is an important factor to examine. As mentioned before, M&A’s should lead to creating

(7)

3 synergies, stimulation of growth, increasing supply chain power and eliminating competition (Stanwick & Stanwick, 2001). But what specific characteristics of an industry could be beneficial for creating synergies? Laabs and Schiereck (2010) show that a short term view, often implies that M&A’s lead to synergies, whereas a long term view often shows negative results of M&A, resulting in no synergy creation. In this paper, the focus is only on the short term. In addition, it is interesting to ask ourselves what characteristics might influence the success of creating synergies? Previous research has failed to provide a detailed overview on what industry characteristics might play a role and the size of this effect. For that reason, this paper aims to shed a light on this topic.

A second reason for managers to engage in M&A is stimulating growth. Growth is often limited by agency problems, internal inefficiencies and capital market imperfections. Jensen and Ruback (1983) show that mergers and acquisitions can be a solution for these agency problems, internal inefficiencies and capital market imperfections (Jensen & Ruback, 1983). This claim is supported by Servaes and Tamayo (2014) who show that a takeover threat will cause a reduction of internal inefficiencies and market imperfections. For this reason, one would expect M&A to be profitable on average. These findings imply that M&A on average should be profitable. However, there are intrusive factors that limit the positive effect of M&A on agency problems, internal inefficiencies and capital market imperfections. This could be due to cultural factors, integration mechanisms or possibly by industrial factors. Colvin and Selden (2003) provide a different point of view on this topic. Colvin & Selden (2003) found that three out of four acquisitions fail, due to the fact that executives focus on sales and profits going up, whereas in reality they are destroying the companies true profitability, namely return on invested capital (Selden & Colvin, 2003). This implies that the far-reaching pressure to get rid of these imperfections might lead to an opposite effect, namely the destroying of the company’s true profitability.

Another reason for CEO’s to engage in M&A is increasing the company’s supply chain power. The discussion on whether vertical or horizontal integration is profitable is an often reoccurring one. Rozen-Bahker (2018) argues that vertical integration success is not something that stands on its own. It is a combination of synergy creation and integration mechanisms. She found that in general vertical M&A is not successful, because companies often fail to integrate the acquired company properly. This might have to do with corporate culture. Similarly, it might be due to industrial factors as well. A large difference between horizontal and vertical M&A is that vertical M&A is often a cross industry deal, while horizontal M&A are often within industry deals. This paper will make a distinction in this factor. Further elaboration on this topic can be found in the results section.

(8)

4 The last argument for engaging in CEO was eliminating competition. Peter Stanwick and Sarah Stanwick (2001) showed that a merger or acquisition could strengthen the competitive position of the company through expansion of their products to different markets and though different distribution channels. This allows the company to take a larger market share. In most cases even more market share than that of the acquiror and target had together. However, it is often very difficult to map a company’s competitive position. Especially when the company operates cross border. Then other factors play a role, such as culture and industrial factors. Kim (2020) argues that dependent on the type of industry, a competitor’s behaviour is different. In some high tech industries, acquiring high tech start-ups might lead to even more start-ups, hence lead to more competition (Kim D. J., 2020). Therefore, it is important to know what industry to perform a merger or acquisition in.

Previous literature has shown that there is some ambiguity on whether or not to engage in mergers and acquisitions. In this section, this paper will provide some more contrasting views on this matter. Laabs and Schiereck (2010) show that previous research on M&A has proven that short term returns for shareholders are positive after engaging in a merger or acquisition. However, long term results show significant losses for the acquiring firm. One would expect that this negative performance would lead to less M&A activity by acquiring companies. In reality, the opposite of this statement is true. Even though most M&A’s lead to long term losses for shareholders, most large firms keep participating in them. All of this can be explained by the research done by Steger and Kummer (2007). They found that there is a vicious circle within a company that goes from pressure to failure. When a merger wave has resulted in a negative outcome, the pressure from shareholders starts to increase again. In the long run, this leads to engaging in M&A again, and often in failing to make M&A successful. Therefore, it is unlikely that more research, proving that M&A has long term negative effect, will change the behaviour of these companies. For that reason, it is important to focus on improving M&A performance rather than changing companies’ investment appetite. Large investors often invest in companies they think are profitable. By looking at companies’ financial statements, they forecast which companies will be profitable in the long run. Even though most companies use the same overall ledger, the business environments that the companies operate in are very different. Therefore, it is very important to analyse the business environments of companies and to weigh the pros and cons of this environment. The business environment of a company consists partially of characteristics of the country that the company operates in (Kimelberg & Williams, 2013). The other parts can be found in the industry where the company operates in. Previous paragraphs have shown that industrial characteristics might be the missing link in M&A research. The industry that a company is participating in plays an important role when it comes to determining the business environment. By examining what

(9)

5 industry characteristics influence M&A performance negatively or positively, investors can make better decisions regarding what industry to invest in. It might help managers in making the most profitable investment decision.

2.2 What has been done?

In order to be able to reach CEO’s goals for mergers or acquisitions, more knowledge on factors that play a role is needed. Previous paragraphs have shown that most errors might be nested in national culture, corporate culture or industry differences. A lot of previous research has been done with the focus on national culture and corporate culture in relation with M&A. National culture can be described as the way that citizens understand their world and their place in it. Corporate culture is slightly different. It can be described as the patterns of actions, words, beliefs and behaviours of a company’s members. It is often referred to as a company’s DNA (Able, 2007). First of all, the factor national culture will be explored. Previous research, performed by Chakrabarti, Gupta-Mukherjee and Jayaraman (2009) has shown that cross border acquisitions performs better in the long run, when the target and the acquirors national culture differs. In addition, their research found that this effect is even larger when it is a cash merger or acquisition and a friendly takeover. When corporate culture is examined, current literature shows that there is a lot of previous research on the topic. For example, the research done by Bargeron, Lehn and Smith (2012). Lehn and Smith show that companies with stronger corporate culture, engage more in smaller mergers or acquisitions. When they do engage in larger mergers or acquisitions, the M&A performance of companies with strong corporate culture is proven to be lower than that of companies with weaker corporate culture.

2.3 Why industry factors?

Previous paragraph provided a small summary of research on culture in relation with M&A that has been done. However, research on industry level effects is lacking in providing theoretical statements about the impact of industry characteristics on the success of M&A. Previous paragraphs showed that the answer to why M&A’s are not always successful might be in the fact that industry level factors are often neglected. For example, previous research has shown that economic exposure is higher at industry level than at firm level (Butt, Rehman, Khan, & Safwan, 2010). Furthermore, Schroeder and Ahmad (2003) state that recent trends in M&A and globalization make it a necessity to look at M&A and other practices from both a country as an industry context. That means that there is an important difference, which might influence M&A performance. For sake of precision, it is good to measure factors at an industry level, rather than only looking at firm level domestic or cross border effects. That is because these companies that are operating within the same industry, compete with each other on global scale. It is not possible to compare them only by national factors or cultural factors, because that does not capture the external environment of the company as a whole. Therefore, looking

(10)

6 at the topic from an industry perspective, might be key in determining what influences M&A performance. This research aims to do so by finding an answer to the following research question:

RQ: Which industry characteristics influence the performance of a merger or acquisition?

3. Methodology

3.1 Dependent variable

In this paper the impact of industry characteristics on M&A performance will be examined. Therefore, M&A performance needs to be measured. Meglio and Risberg (2011), argue that there is not one universal way to measure M&A performance. Instead, the authors argue that there are several measurement approaches, that apply on different research setups. One way to measure M&A performance is by using the capital asset prising model (CAPM) as the valuation model. Another way is by valuating stocks using the Fama and French three factor model. This model uses the capital asset prising model and adds size risk and value risk to this valuation (Hayes, 2020). Griffin (2002) states that this model is only accurate when one uses domestic factors to calculate size risk and value risk. However, since this research relies on a wide set of deals from different industries, this is difficult to measure. Industries operate across the entire world and are not bound by borders. Assigning country specific factors to industries is therefore really difficult and often inaccurate. This in turn might lead to unreliable results. In Addition, research done by Bartholdy and Peare (2005) made a comparison between the Fama and French three factor model and CAPM. They found that CAPM is better for making estimations. Therefore, the CAPM is used as benchmark to calculate the cumulative abnormal returns in this research.

As mentioned earlier by Selden & Colvin (2003), one should focus on creating higher return on invested capital. For this reason, M&A success is measured by the increase of cumulative abnormal returns of combined stock. That means that the cumulative abnormal returns of the target and acquiror are calculated and in the end added up. In Addition, by adding both up, the combined effect of the merger or acquisition is measured, rather than just the effect of the target. The approach used is different than in most researches. Most researchers only examined the cumulative abnormal return of the acquiror. However, since both can be affected by for example synergy potential, economies of scale or economies of scope, these are added up in this research. This in in line with the approach used by Ikram and Nughroho (2014). This measurement approach reflects return on invested capital best, when looking at it from a shareholder perspective. That is in line with the measurement approach suggested by Das and Kapil (2012). They state that when performing an event study, looking at combined

(11)

7 cumulative abnormal returns in the event window compared to the estimation window is a good way at estimating M&A performance. When analysing this data, there are two important time periods to consider. These periods are based on the approach used in multiple researches regarding M&A performance like the research done by Diana Pop (2006). The first period is the estimation window. This is the phase where the companies did not have contact and the stock prices are not yet affected by M&A related events. This method assumes that investors respond in a short window after the merger, because they make predictions about possible synergy creation (Das & Kapil, 2012). The estimation window chosen for this research is one year prior to the announcement date. The second period to consider is the event window. In this period, combined cumulative abnormal returns, as a result of the merger or acquisition, are measured. In this paper the event window used is two months after the completion date of the merger or acquisition. This is a called the short run performance of the deal. Long run performance would be if two to five years after the deal are examined.

Figure 1: Estimation and event window

This indicates the actual returns on invested capital of the acquiring company. In this research the capital asset pricing model is used as a benchmark to calculate the cumulative abnormal returns of companies in the event window (Zaremba & Płotnicki, 2016). That has been done by looking at the systematic return of the different indices in combination with the correlation (beta) of the firm with the market. The beta of the market is calculated by using data on the index of each company. This data was used to calculate the beta in Stata. If the return of the stock exceeds the market return considering the beta, then this is called abnormal return. These abnormal returns are added up to come up with combined cumulative abnormal returns. Finally, this research will examine whether these cumulative abnormal returns are different in different industry contexts, due to particular industry characteristics. The formula used to calculate the cumulative abnormal returns is the following:

𝐶𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒 𝑎𝑏𝑛𝑜𝑟𝑚𝑎𝑙 𝑟𝑒𝑡𝑢𝑟𝑛 = 𝑅(𝑡) + 𝑅(𝑎) − 𝑅𝑚(𝑡) ∗ 𝑏 − 𝑅𝑚(𝑎) ∗ 𝑏

Where R(t) is the return of the target, R(a) is the return of the acquiror. Rm(t) is the market return of the index that applies to the particular target, Rm(a) is the same but from the index of the acquiror. Finally, the b is the beta calculated by Stata.

(12)

8

3.2 Independent variables

This research aims to provide a broad view on what industry characteristics might actually play a role in M&A performance. In order to do so, the effect of seven industry characteristics on M&A performance will be examined. The following variables will be examined: (1) Labour intensity, (2) Board diversity, (3) Technology intensity, (4) Profitability, (5) Human capital intensity. Data on industry characteristics is very hard to gather. Most databases do not have data on industries as a whole, which explains the limited amount of research done on this topic. To still be able to perform this research, the decision was made to gather data on the top 50 peer firms in the industry. The data on the top 50 firms in each industry provided us with a good proxy for the initial industry variables. In total, data of 72000 unique peer firms was used to come up with good proxies for our initial independent variables. Next paragraph will elaborate further on why these variables are expected to have an effect on M&A performance and how these variables will be measured. In addition, two kinds of dummy variables are used in order to measure different effects. The dummies that will be used are (a) cross industry versus within industry dummy and (b) cross country versus domestic deals dummy. More details on these dummy variables will be provided later in this paper.

The first independent variable that is examined is labour intensity. Research on this topic has been done by Nunnenkamp and Spatz (2003). They argue that there are large differences on how much companies spend per USD value added. Therefore, there are differences in practices of companies, which might be difficult to integrate. That in turn, might influence the possibility to create synergies. This insight might indicate that certain industries are better in adding value, and therefore better in creating synergies. Previous research has shown lower labour intensity ratio’s provide certain economic benefits (Spitsin, et al., 2015). This might be beneficial for acquiring companies. However, research has also shown that it leads to negative social results. This might go at the expense of synergy creation. Previous research show that there are contrasting results in literature. Therefore, this research aims to provide clarity on the topic, when it comes to mergers and acquisitions. According to Tregenna (2014) there are two ways to measure the labour intensity of an industry. This can be done by looking at the ratio between employment & capital stock or by looking at the ratio between employment and value added. In this research the first approach is used, because this provides clear indications about how labour absorbing activities in particular industries are. In addition, this measure makes a clear distinction between labour intensive and capital-intensive industries, which might help us explain a possible effect in the results. Capital can be defined as amount of regular and preferred stocks issued times the par value per share (Akhilesh, 2019). Labour intensity can be explained as the number of employees per million US Dollar (USD) in capital stock. This is the inverse of labour productivity, which calculates the ratio

(13)

9 between value added and number of employees. As mentioned before, this paper does not use actual industry variables, but rather a proxy of the top fifty companies in the industry. The formula that is used to measure labour intensity in this paper is the following:

𝐿𝐼 =𝑁𝑟. 𝑜𝑓 𝐸𝑚𝑝𝑙𝑜𝑦𝑒𝑒𝑠 𝑆𝑡𝑜𝑐𝑘 𝑐𝑎𝑝𝑖𝑡𝑎𝑙

The second variable that is examined in this research is board diversity. In this paper board diversity can be defined as the degree of presence of females on the board of directors. This is operationalized by looking at the percentage of females on boards of directors within an industry. Research by Chen and Crossland (2014) has shown that female representation on the board of directors has a negative influence on number of acquisitions and the size of the acquisitions. However, this does not necessarily mean that high representation of females on the board of directors is bad for M&A performance, because most large mergers and acquisitions fail. One might even expect that female representation on the board leads to higher success rate of M&A, because they appear to be pickier in choosing M&A deals. This is in line with the research by Bellinger and Hillman (2000). The authors performed a research on how board diversity can influence merger performance positively. They find that board diversity has positive influence of merger performance. This can be explained by the level of tolerance that woman have, compared to men. However, their research is performed with firm level data. In contrast to other research, Hagendorff and Keasey (2019) found that board diversity, when it comes to gender diversification, does not show significantly different results. However, this research was only focussed on the banking industry. Moreover, Bazel-Shoham, Mook Lee, Rivera and Shoham (2020) performed a research on board diversity with cross border deals. They found that when there are more females on the board of directors, than there is less cross border M&A activity. In this research, the influence of board diversity on an industry level will be examined. This will be measured by measuring the percentage of woman on the board of directors in the industry as a whole. Similarly to other independent variables, this variable based on a proxy, based on the top fifty companies in the industry.

𝐵𝐷 =𝐹𝑒𝑚𝑎𝑙𝑒𝑠 𝑜𝑛 𝐵𝑜𝑎𝑟𝑑

𝐵𝑜𝑎𝑟𝑑𝑚𝑒𝑚𝑏𝑒𝑟𝑠 ∗ 100%

The next variable that is examined in this paper is technology intensity of industries. Technology intensity can be defined as research and development costs divided by total sales (Jones, 2007). A previous study performed by Cassiman, Colombo, Garrone and Veugelers (2005), showed that on average, technology expenses will go down after completion of a merger or acquisition. However, sales of these companies do not change. That indicates that companies cut their costs while sales stay the same. That would result a higher profit, hence a higher M&A performance. This effect is stronger in industries that are very technology

(14)

10 intensive. This finding is supported by Lin, Huang and Liu (2010), who found that more M&A leads to higher monopoly power, which in turn leads to dampening the innovation of companies in terms R&D and patents. In contrary Fernandez, Triguero, & Alfaro-Cortes (2019) argue that technology intensity will go up in the short and long term after the merger or acquisition. This might indicate that there is an effect of technology intensity on M&A performance, but the direction of this effect is not clear yet. Moreover, Miyazaki (2009) investigated the effects of technology intensity on the likelyhood of M&A deals. He found that high level technology intensity leads to higher profitability expectations, because of synergy creating possibilities. In addition, Blonigen & Taylor (2000) show that companies with relatively low technology intensity are more likely to participate in the acquisition market. One would expect that they only do this because history has proven it to be beneficial. The literature review shows that there is a lot of ambiguity on this topic. Therefore, this research strives to examine the real relationship by looking at it on an industry level. As mentioned before, technology intensity can be defined as research and development (R&D) costs divided by total sales (Jones, 2007). However, most companies do not directly publish their R&D figures. Therefore, the decision is made to use a proxy for this. The number of patents of the top fifty peer firms will be used as a proxy for R&D costs. The formula that is used in this paper for determining the technology intensity of industries is the following:

𝑇𝑒𝑐ℎ𝑛𝑜𝑙𝑜𝑔𝑦 𝑖𝑛𝑡𝑒𝑛𝑠𝑖𝑡𝑦 =𝑁𝑟. 𝑜𝑓 𝑃𝑎𝑡𝑒𝑛𝑡𝑠 𝑇𝑜𝑡𝑎𝑙 𝑠𝑎𝑙𝑒𝑠

In this paper, industry profitability is also examined as variable that might influence M&A performance. Research performed by Li and Islam (2019) has shown that firms that are profitable, become less leveraged. That means they reduce the amount of debt in their firm. However, the authors add that the relation between profitability and leverage is different in different industries. Different industries might react differently to profitability. Therefore, the influence of profitability might vary across industries as well. Profitable industries are often the target for a large merger or acquisition because it seems profitable. However, acquiring a firm in such a profitable industry, might be extensively more expensive. The question that rises is: is offering more money, profitable on the long run. Fernandez, Triguero and Alfaro-Cortes (2019) found that on a firm level, profitability will have no short term effect but only a long term effect. That means that the performance of a merger or acquisition will grow, but not in a short period after the merger. However, this is on a firm level, this research will examine whether difference in industry profitability has influence on M&A performance as well. Industry profitability can be defined as earnings before interest, depreciation and amortization of the industry divided by total assets of the industry. There is no data available of industry assets or

(15)

11 industry EBITDA, therefore a proxy will be used to measure this. Fifty peer companies of all targets in the dataset are used. These peer companies form a proxy for the entire industry.

𝑃 = 𝐸𝐵𝐼𝑇𝐷𝐴 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠

Finally, the variable human capital intensity is measured. Russell Coff (2002) performed research on the effect of human capital on acquisitions. He argues that in most mergers and acquisitions happen in capital intensive industries, because human capital is difficult to acquire in efficient labour markets. This in turn, might cause acquiring firms to overbid on firms that are human capital intensive. Therefore, human capital intensity might have impact on M&A performance as well. In addition, Lin, Hung and Li (2006) performed research in the banking industry. In contrast to the findings of Coff, they found that M&A could be very effective when the company has high HR capability. This could be the case for high human capital industries as well. They also argue that this effect is stronger when the deal is performed domestically. The effect is weaker when the deal is performed cross border. This variable can be measured by the employee compensation in USD divided by number of persons employed (BEA, 2003). Similarly to the other independent variables, there is no industry data available. A proxy based on the top fifty companies in the industry is used to come as close to the actual variable as possible. The degree of human capital intensity within industries is given by the following formula.

𝐻𝐶𝐼 =𝐸𝑚𝑝𝑙𝑜𝑦𝑒𝑒 𝑐𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛 𝑁𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑒𝑠

As mentioned before, there might be differences within our results. Therefore, this paper will use two kinds of dummies in order to measure separate effects of the independent variables on the dependent variable. The first dummy variable is cross industry versus within industry. It can also be seen as vertical versus horizontal integration. Datta, Kodwani and Viney (2013) show that there is a difference in effect from within industry deals and cross industry deals. They concluded that within industry deals benefit, because of economies of scale and economies of scope. In addition, Chon, Choi and Barnett (2003) argue that cross industry deals might have benefits as well. By sharing knowledge, industries become more integrated. This causes companies to change their value chain. Furthermore, cross industry M&A can be a way to generate required capital to compete in a competitive industry. The basis for this dummy is the 4-digit US SIC code. This dummy provides a 1 if the 4-digit US SIC codes are not equal and it provides a 0 if industry codes are the same. By using this dummy it can be examined whether the results hold for all deals or if they hold only for, for example cross industry deals.

(16)

12 Secondly, a dummy variable for cross border versus domestic deals was added. Previous research by Anand, Capron and Mitchell (2005) provides evidence that cross border mergers and acquisitions enhances the acquirors capabilities and it improves their financial performance. Investors value global operating companies higher than domestic operating companies. In addition, Reynolds and Teerikangas (2015) state that there is no absolute difference between domestic and cross border, because domestic deals are still embedded in international contexts. Employee experience in domestic operating companies still bears international characteristics. In this research it will be tested whether this is indeed the case, or whether other results are found. This dummy is based on the country code of both the acquiror as the target. If both country codes are the same the dummy variable provides a 1 if it is a domestic deal and a 0 when it is a cross border deal. This way it is possible to measure whether there is a statistically significant difference in cross border deals and domestic deals, when it comes to measuring the effect of industry variables.

3.3 Hypotheses

As mentioned in previous paragraph, this research expects that several industry characteristics will have positive or negative influence on the performance of mergers and acquisitions. Based on the literature discussed, the following hypotheses are formulated:

In the literature review it was argued that there are contrasting theories on the effect of labour intensity on M&A performance. However, we believe that the argumentation for a negative relation is stronger. Research has shown that higher labour intensity leads to negative social results. This can affect one of the key goals of M&A, namely synergy creation. Moreover, previous research has shown that lower labour intensity ratio’s provide certain economic benefits as well (Spitsin, et al., 2015). Therefore, it is expected that labour intensity will negatively influence the performance of mergers and acquisitions, due to the fact that is makes integration, and therefore synergy creation, more difficult.

H1: Labour intensity will negatively influence M&A performance.

Secondly, this paper examined board diversity. Previous literature showed contrasting theories. The most compelling theory argues that board diversity has a positive effect on M&A performance if one looks at it from a firm level perspective. This is supported by Bellinger and Hillman (2000). They found that, on a firm level, board diversity has positive effect on merger performance. This can be explained by the level of tolerance that woman have, compared to men. Therefore, the same results are expected when it is measured at an industry level.

(17)

13 Zooming in on technology intensity shows that most theories expect that M&A performance will be positively influenced by technology intensity. The reasons for that are that merged companies often decrease their technology spending, while sales stay at the same level (Cassiman, Colombo, Garrone, & Veugelers, 2005). Hence, they cut costs where possible. Another explanation is given by Miyazaki (2009). He investigated the effects of technology intensity on the likelyhood of M&A deals. He found that high level technology intensity leads to higher profitability expectations, because of synergy creating possibilities. Therefore, it is expected that technology intensity will influence M&A performance positively.

H3: Technology intensity will positively influence M&A performance.

The next independent variable that is used in this paper is the profitability of industries. Fernandez, Triguero and Alfaro-Cortes (2019) show that M&A performance, when looking at firm level, will increase in the long run. When industries are very profitable, this usually leads to higher M&A performance, whereas low profitability would most likely lead to worse M&A performance.

H4: The higher the profitability of the industry, the higher M&A performance.

The last hypothesis focusses on human capital intensity. As mentioned before, high human capital intensity leads to relatively more mergers and acquisitions in an industry. The reason for that, is that the acquiring firm wants to buy human capital. Contrasting theories were found. However, the most compelling theory was that this often leads to overbidding on firms (Coff, 2002). Therefore, it is expected that the M&A performance will be influenced negatively by human capital intensity.

H5: The higher human capital intensity, the lower M&A performance.

3.4 Conceptual framework

In order to find answers to the research question, it is important to define how the research will be performed. To do so, a conceptual framework was made. The conceptual framework below provides a small overview of the expected effects of this research. As mentioned before, this paper looks at the difference in combined cumulative abnormal returns of deals to measure the success of the merger or acquisition. Therefore, this is going to be the dependent variable of this research. As mentioned in the hypotheses section, this paper expects that M&A performance is affected by a group of independent variables, that reflect particular industry characteristics. The independent variables are the following: (1) Labour intensity, (2) Board diversity, (3) Technology intensity, (4) Profitability, and (5) Human capital intensity. In this research we also consider interaction variables. They are not added in this conceptual

(18)

14 framework. Further elaboration on these interaction variables can be found in chapter 4. Combining these variables results in the following conceptual framework:

Figure 2: Conceptual framework

3.5 Dataset

3.5.1 Dataset dependent variable

In order to be able to examine the hypothesis several databases are used. The Zephyr database is used to gather data on 1726 deals that were made by listed companies worldwide. Non listed firms were excluded, because there is not sufficient stock data available on these firms. Moreover, the focus of this paper is only on firms that actually performed and completed a merger or acquisition. Other takeover mechanisms were excluded from the dataset. Deals that were performed, where the target and acquiror have the same ultimate owner were excluded from this research as well, because these companies might have been strongly interconnected before the merger or acquisition. Hence, the effect of this deal is not reliable. Finally, all deals in the dataset that were completed before 01/01/2009 and after 01/01/2019 were removed, because the returns of the deals before 2009 might have been affected by the 2008 credit crisis and it is not possible to measure the effect of deals after 2019. After excluding some deals, there are 1194 unique deals in the dataset.

Secondly, the Eikon database is used to gather data on stock returns of the acquiror and target within one year before the announcement date and two months after the completion date. Over this period in time, the percentage change in stock value is measured. For the same period in time, the corresponding indices are matched and used to gather data on the percentage change in value of these indices. Later on, this data will be used to calculate the abnormal returns. Moreover, this data can be used to calculate the beta as well. This can be done using Stata.

The data gathered in these databases can be used to calculate the abnormal returns of target and acquirors after performing the merger or acquisition. First of all, some calculations

(19)

15 have to be made in order to be able to use the data that is gathered earlier. The dependent variable is measured by looking at combined cumulative abnormal returns. Cumulative abnormal returns can be defined as the returns of a stock, minus the benchmark returns. In general, that means that returns of the target and acquiror are compared with the average return in the index. Afterwards, these returns are added up to come up with combined cumulative abnormal returns as a result of the merger or acquisition. This will be done in order to accurately measure the effect of the takeover. By looking at abnormal returns, this paper accounts for systematic returns. Moreover, a company’s beta is used, because this provides information about the correlation of stocks with the market. The beta of all companies in the dataset is used to calculate their particular benchmark return. The abnormal return can be calculated using the following formula:

𝐶𝑜𝑚𝑏𝑖𝑛𝑒𝑑 𝐶𝑢𝑚𝑢𝑙𝑎𝑡𝑖𝑣𝑒 𝐴𝑏𝑛𝑜𝑟𝑚𝑎𝑙 𝑅𝑒𝑡𝑢𝑟𝑛 = 𝑅(𝑡) + 𝑅(𝑎) − 𝑅𝑚(𝑡) ∗ 𝑏 − 𝑅𝑚(𝑎) ∗ 𝑏

Where R(t) is the return of the target, R(a) is the return of the acquiror. Rm(t) is the market return of the index that applies to the particular target, Rm(a) is the same but from the index of the acquiror. Finally, the b is the beta calculated by Stata. This formula is used to calculate the abnormal returns of companies in the dataset in a short window after the merger or acquisition. Later on, in this paper, this data will be used to see what industry characteristics influence M&A performance positively or negatively. The formula shows that the returns of the target (t) and the acquiror (a) are added up and the market returns times the beta of both are subtracted. This way this paper makes sure that the effect that is measured is not due to a specific event in the market. Moreover, by adding both up, the combined effect of the merger or acquisition is measured, rather than just the effect of the target. The approach used is different than in most research. Most researchers only examined the cumulative abnormal return of the acquiror. However, since both can be affected by for example synergy potential, economies of scale or economies of scope, these are added up. This in in line with the approach used by Ikram and Nughroho (2014).

Table 1: Combined Cumulative Abnormal Returns

3.5.2 Dataset independent variable

As discussed before, the effect of the following variables will be examined: (1) Labour intensity, (2) Board diversity, (3) Technology intensity, (4) Profitability, and (5) Human capital intensity. In order to measure the effect of these variables, data is gathered from several datasets. Most of these variables are not available in existing databases, because these databases do not have data on industries as a whole. Therefore, is was decided to make proxy’s in order to

Variable Obs Mean Std. Dev. Min Max

(20)

16 come as close to the actual variable as possible. Using these proxies can provide useful insights on possible effects of these industry variables. The proxies are based on the US SIC code of the target firms. For all US SIC codes data was gathered on the top 50 companies in that industry. The data of these companies represent the industry as a whole. These proxies are used to replace the industry variable, without losing the ability to form conclusions on industry variables. Moreover, the data that was gathered from these companies is always data of the year of the announcement date. That way it is assured that the relations that are found, are due to actual data from that particular date. The table below provides a small overview of how the independent variables are calculated and with what data.

Table 2: Independent variables

Variable Formula Databa

se(s) Data (1) Labour intensity =𝑁𝑟. 𝑜𝑓 𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑒𝑠 𝑆𝑡𝑜𝑐𝑘 𝑐𝑎𝑝𝑖𝑡𝑎𝑙

Orbis Proxy based on number of employees and stock capital of 72850 peer companies. (2) Board diversity = 𝑁𝑟. 𝑜𝑓 𝑓𝑒𝑚𝑎𝑙𝑒𝑠 𝑜𝑛 𝑏𝑜𝑎𝑟𝑑 𝑁𝑟. 𝑜𝑓 𝑚𝑎𝑙𝑒𝑠 + 𝑓𝑒𝑚𝑎𝑙𝑒𝑠 𝑜𝑛 𝑏𝑜𝑎𝑟𝑑 ∗ 100%

Orbis Proxy based on number of females and males on board of directors of 72850 peer companies. (3) Technology intensity =𝑁𝑟. 𝑜𝑓 𝑔𝑟𝑎𝑛𝑡𝑒𝑑 𝑝𝑎𝑡𝑒𝑛𝑡𝑠 𝑇𝑜𝑡𝑎𝑙 𝑠𝑎𝑙𝑒𝑠

Orbis Proxy of the technology intensity by looking at number of patents of 72850 peer companies. (4) Profitability = 𝐸𝐵𝐼𝑇𝐷𝐴 𝑇𝑜𝑡𝑎𝑙 𝑎𝑠𝑠𝑒𝑡𝑠

Orbis Proxy based on the EBITDA and total assets of 72850 peer companies. (5) Human capital intensity =𝐸𝑚𝑝𝑙𝑜𝑦𝑒𝑒 𝑐𝑜𝑚𝑝𝑒𝑛𝑠𝑎𝑡𝑖𝑜𝑛 𝑁𝑟. 𝑜𝑓 𝑒𝑚𝑝𝑙𝑜𝑦𝑒𝑒𝑠

Orbis Proxy based on the employee

compensation and number of employees of 72850 peer companies.

(21)

17 All independent variables had a small number of missing values. These missing values were replaced with the mean of the variable. This is in line with the approach suggested by Alan Acock (2005). The picture below shows some summary statistics of these new variables:

Table 3: Summary statistics: independent variables

This table shows that there are large differences in scores of companies on labour intensity and human capital intensity. These differences are a lot smaller when it comes to profitability, technology intensity and female board representation.

3.6 Analytical tools

Previous paragraph elaborated on the data needed for this research and the data gathering method. This section will discuss how the data is used in relation to the hypotheses. In order to do so, several methodological tools will be used. An event study will be performed, because this research wants to measure the different effects of an event, in this case a merger or acquisition, on the abnormal stock returns of companies. The difference between the events come from difference in industry that the merger or acquisition takes place in. For example, M&A in high tech industries might have a significantly different effect on M&A performance than M&A in high human capital-intensive industries. In order to measure these potential effects, this research looks at two periods in time. The estimation window and the event window. The estimation window that is used in this paper is the period of one year before the merger announcement. That way, the stock returns at x-1 are not affected by the event itself. The event window that is used, is a window of two months after the merger is completed. This is also called the short run performance window. It is expected that this captures the effect best, because investors respond quick to completed mergers and they anticipate on synergy potential. Therefore, the stock value of the company two months after the merger or acquisition should represent the M&A performance most accurately.

This paper is based on cross sectional data. That means that there is no specific time ordering in the data. Because of that, the normal checks for autocorrelation and trends are not applicable on this data. Hence, this paper will not check for these kinds of assumptions in the dataset. The only check that is possible is the test for normal distribution of the data. Therefore, the first step in examining the data is to check whether the data is normally distributed. To do so, the histograms of the different variables are examined. If the data is normally distributed, it

Variable Obs Mean Std. Dev. Min Max

Labourintensity 1194 144944.79 144073.37 15135.798 990366.79

Humancapital 1194 90.64 105.456 9.989 537.07

Profitability 1194 -.937 3.137 -17.645 3.226

TechIntensity 1194 .612 1.588 0 6.125

(22)

18 is possible to run a t-test to test for testing the hypotheses. If the data is not normally distributed, the data will be transformed by taking at the logarithm of the variable (LaLonde, 2005). The normality test showed us that the following variables were not normally distributed: (1) technology intensity, (2) Human Capital intensity and (3) Labour intensity. As mentioned before, the logarithm of these variables is used in order to make the distribution normal. By using the logarithm, the normality of the variables is improved. The rule of thumb regarding skewness and kurtosis is that the skewness should be between 0,5 and -0,5 and kurtosis should be as close to 3 as possible. As shown in appendix 1, using the logarithm strongly improved the variables normality. Appendix 1 provides more information on how the transformation improved the variables’ skewness and kurtosis. Unfortunately, the variable profitability was not normally distributed and is not transformable to a normal distribution. A second control of normality is looking at the histograms of the variables. The histograms of the different variables after the logarithm transformation are attached in appendix 2. This shows us that the variables are much better normally distributed after the transformation, hence the logarithmic transformation was useful.

This paper does not only want to measure whether the merger or acquisition causes a significant effect. Rather, the aim is to measure what industry characteristics influence the effect size most. In order to measure that, an ordinary least squares (OLS) regression will be performed. This OLS regression will point out what the effects of the different industry characteristics are on the degree of combined cumulative abnormal return effect. The independent variables in this regression are (1) Labour intensity, (2) Board diversity, (3) Technology intensity, (4) Profitability, and (5) Human capital intensity. The dependent variable used in this OLS regression is M&A performance. This paper controls for interaction effects by adding interaction variables to the equation. An interaction variable will be created if the correlation in the correlation table is higher than 0.25 or lower than -0.25. Moreover, all variable/dummy combinations are added in order to examine the pure effect of the variables. See the table below for the correlation figures.

Table 4: Interaction table

The following interaction variables were added to the regression (see appendix 5 for further explanation on abbreviations): (a) LogHC*LogLI, (b) Profit.*LogHC, (c) Log TI*LogHC, (d)

Variables (1) (2) (3) (4) (5) (6) (1) CombinedCAR 1.000 (2) LogLI 0.022 1.000 (3) LogHC 0.048 0.500 1.000 (4) Profitability 0.002 -0.147 -0.277 1.000 (5) LogTI 0.084 0.031 0.375 -0.026 1.000 (6) FemBoardRep -0.030 0.011 0.147 0.275 -0.083 1.000

(23)

19 Profit.*TBR (e) and an interaction variable that captures all the interactions between the industry and country dummy variable and the independent variables, this adds up to another 10 interaction variables. See appendix 3 for an overview of interaction variables. Moreover, this research controls for several previously proven factors that influence M&A performance. Das and Kapil (2012) provide an overview of explanatory variables that has been used in previous research to explain M&A performance. To control for the effect of these variables several control variables were included in this paper. The control variables used in this paper are the following: (A) Acquiror to target relatedness, (B) Debt to equity (D/E) ratio of acquiror, (C) dummy for industry differences and (D) dummy for country differences (E) country risk. The acquiror to target relatedness and the D/E ratio of the acquiror are based on the paper of Das and Kapil (2012). They state that those are variables might affect the combined cumulative abnormal returns after deals. The dummy for industry differences makes a difference in within industry mergers and cross industry mergers. This variable is based on the US SIC 4-digit industry code. The dummy for country differences makes a difference between cross border deals and domestic deals. This is based on the approach used by Doyten and Cakan (2011). They separated the effects of domestic and cross border to get a more detailed perspective on the topic. The distinction of cross border and domestic is based on the country code of both the acquiror as the target firm. If this is the same, it is a domestic deal, if it differs than the deal is cross border. Finally, this research controls for country risk as described by Das & Kapil (2012) as well. This dummy accounts for all other country differences, like culture, regulation etcetera. An extensive list of the control variables and their measurement approach is provided in appendix 4. Now that the data is properly prepared, this paper continues with the analysis of the data. As mentioned before, an OLS regression will be performed to do so. In next chapter this paper will elaborate on the results of this OLS regression.

4. Results

4.1 Overall results

In this section the results of the OLS regression are shown. For sake of simplicity, the country difference indicators were removed from the regression output table. However, they were still part of the initial regression. They were removed from the regression output that is shown in this article, because the output of these country indicators themselves do not contain any information that is relevant for this research. In the table below you can find the OLS regression output. You can find an overview of the abbreviation explanation of the variables in appendix 5.

(24)

20 In order to analyze this regression table, overall statistics are examined first. The R-squared figure of this regression equation is 0.177. That means that 17,7% of the variation in the combined cumulative abnormal returns is due to variables included in this regression. That means that more than 82% of the variation can be explained by other factors, which leaves room for future research. Moreover, we find significant results for the logarithm of labour intensity and the logarithm of human capital intensity. Both factors show a positive coefficient, which means that there is a significant positive effect of these variables on M&A performance. In next paragraph this paper will elaborate more on what the results mean for the hypotheses.

4.2 Testing hypotheses

Prior to this research, several hypotheses based on literature of prior research were provided. Particular results were expected based on these hypotheses. In this section this paper will discuss whether these predictions were accurate and if not, why they were not accurate. The hypotheses that were introduced in the beginning of this research were the following:

Combined

CAR Coef. St.Err. t-value p-value [95% Conf Interval] Sig

LogHC LogLI LogHC*LogLI LogTI LogTI*LogHC Profitability Profit.*LogHC FemBoardRep Profit.*FBR LogLI*Profit. DumInd DumInd*LogLI DumInd*Profit. DumInd*LogLI DumIndLogTI DumInd*FBR DumInd*LogHC DumCountry DumCountry*LogLI DumCountry*Profit. DumCountry*LogTI DumCountry*FBR DumCountry*LogHC AcquirorTotalAssets EquityOnLiabilities 180.73 69.998 2.58 .01 43.132 318.327 ** 63.434 24.248 2.62 .009 15.768 111.099 *** -14.733 5.658 -2.60 .01 -25.855 -3.611 *** 3.72 6.775 0.55 .583 -9.597 17.037 -.658 1.477 -0.45 .656 -3.561 2.245 -58.789 116.76 -0.50 .615 -288.31 170.731 4.241 6.6 0.64 .521 -8.732 17.214 -.078 3.524 -0.02 .982 -7.006 6.849 4.28 3.375 1.27 .206 -2.356 10.915 -.578 8.687 -0.07 .947 -17.655 16.499 122.975 120.463 1.02 .308 -113.824 359.773 -18.766 11.887 -1.58 .115 -42.133 4.601 52.104 103.036 0.51 .613 -150.437 254.645 -4 8.304 -0.48 .63 -20.324 12.323 -3.715 2.32 -1.60 .11 -8.276 .845 .394 3.765 0.10 .917 -7.008 7.796 18.621 10.716 1.74 .083 -2.444 39.686 * 115.977 159.939 0.73 .469 -198.42 430.375 3.345 15.219 0.22 .826 -26.573 33.262 -2.289 3.86 -0.59 .553 -9.876 5.298 3.517 2.998 1.17 .241 -2.376 9.41 -5.16 4.99 -1.03 .302 -14.97 4.65 -16.819 15.404 -1.09 .276 -47.099 13.461 0 0 -1.34 .179 0 0 .011 .033 0.33 .738 -.053 .075 Mean dependent

var 6.797 SD dependent var 58.242

R-squared 0.177 Number of obs 494.000

F-test 1.091 Prob > F 0.291

Akaike crit. (AIC) 5484.738 Bayesian crit. (BIC) 5829.346

(25)

21

H1: Labour intensity will negatively influence M&A performance.

The first hypothesis was that expected labour intensity was expected to have a negative influence on M&A performance. In this research M&A performance was measured by looking at cumulative abnormal returns of both the target as the acquiror. Labour intensity was measured by creating a proxy based on around 50 companies in that particular industry. As shown in the regression output, the measure for labour intensity was LogLI. This variable is proven to have a significant influence on the combined cumulative abnormal return when it comes to the overall regression output. That means that there is an overall effect. The direction of this effect can be shown by looking at the coefficient. However, this should happen with some caution, because the logarithm of this variable was used, for sake of normality. The coefficient of LogLI is 63.43. that means that an increase of 1 in logarithm of labour intensity, causes an increase of 63.43 in combined cumulative abnormal return. In contrast to the expectations, there is a positive effect of labour intensity on M&A performance. The question that rises is: why is the effect positive?

H2: Board diversity will positively influence M&A performance.

The second hypothesis was that board diversity was expected to have a positive influence on M&A performance. As mentioned before, M&A performance is measured by looking at the combined cumulative abnormal returns. The board diversity of industries is measured by using a proxy. From all industries in the dataset, the average score of the top 50 companies in that industry was taken. Within these industries the percentage of females represented in the board of directors was examined. On beforehand it was expected that more females on the board of directors would have a positive influence on the M&A performance. When the regression output is examined, then the results show that Female Board Representation (FBR) does not have a significant influence on the combined cumulative abnormal returns. That means that the H2 hypothesis is rejected. There is no proven effect of more female representation within an industry on M&A performance.

H3: Technology intensity will positively influence M&A performance.

In order to test this hypothesis, the prediction that was made for the third variable was based on the assumption that higher technology intensity would result in better synergy creation, and therefore higher M&A performance. Hence, a positive relation was expected. The technology intensity is measured by using a proxy. This proxy is calculated by looking at the amount of granted patents of the top 50 companies in the industry. A higher number of patents should indicate higher technology intensity. The logarithm of technology intensity was used in order to get a normal distribution of the data. That means that one should be careful with the interpretation of the coefficient. As shown in the regression output, the variable LogTI is not

(26)

22 significant. That means that higher technology intensity within an industry does not have a significant positive influence on M&A performance for mergers or acquisitions within that industry. That is different than the expected relation based on the hypothesis. Therefore, the H3 hypothesis is rejected as well.

H4: The higher the profitability of the industry, the higher M&A performance.

The fourth hypothesis that was made is based on the profitability of an industry. This paper argued based on literature that when an industry as a whole in more profitable, then M&A performance should be higher too. The profitability of industries is measured by using a proxy. This proxy consists of a ratio of the earnings and assets of the top 50 firms within that particular industry. Due to the fact that profitability had negative figures as well, it was not possible to get a normal distribution of the variable. However, when examining the regression output, results show that profitability does not have a significant influence on combined cumulative abnormal returns. Similarly to technology intensity, we did not find results that support the hypothesis. Therefore, the H4 hypothesis is rejected.

H5: The higher human capital intensity, the lower M&A performance.

The last hypothesis was regarding human capital intensity. Human capital intensity was expected to have a negative effect on M&A performance because high human capital intensity was expected to lead to overbidding firms, which should result in negative outcomes for the combined cumulative abnormal returns. In this paper human capital intensity was measured by using a proxy. This proxy was created by looking at the top 50 firms within the industry. A human capital intensity ratio was calculated based on the employee compensation and the number of employees. As shown in the regression output, there is a significant effect of human capital intensity on M&A performance. The coefficient shows that a rise in the logarithm of human capital intensity of 1, will cause combined cumulative returns to rise with 180.73. This coefficient indicates that there is a large positive effect, which is the contrary of what was expected based on literature. That in turn, means that rejecting the H5 hypothesis is correct. There is no negative relation between human capital intensity and M&A performance. Rather, this effect is positive. The question that rises is why?

4.3 Cross industry versus Within industry M&A

When looking at the overall output, the results show some significant results that are important for testing the hypotheses. However, it is also important to test whether these results hold when focus is on specific deal differences. In this chapter this paper will test what the results would be if a distinction is made between cross industry deals and deals that happen within an industry. The distinction is made by using the dummy variable called dumInd. This dummy

(27)

23 provides a 1 if the industry codes are not equal and it provides a 0 if industry codes are the same. Table 6 provides an overview of these dummy variables.

Table 6: Overview industry dummy distribution

This table shows that both groups have enough observations to be able to draw conclusions from it. The next step is to regress both groups separately, so they can be compared. The following tables show (1) cross industry deals and (2) within industry deals.

Table 7: Cross industry regression

Table 7 shows similar results to the overall regression output. This table shows that both the logarithm of human capital as the logarithm of labour intensity are statistically significant. A difference from the overall output is the effect size. The coefficient of human capital is higher. That means that an increase in logarithm human capital of 1, will cause the combined cumulative abnormal return to increase by 256. The coefficient of labour intensity is nearly equal to the overall regression output. The next table shows the regression of within industry deals.

DumIndustry Freq. Percent Cum.

0 814 68.17 68.17

1 380 31.83 100.00

Total 1194 100.00

CombinedCAR Coef. St.Err. t-value p-value [95% Conf Interval] Sig

LogHC 256.817 72.597 3.54 0 113.91 399.724 ***

LogLI 56.908 23.086 2.47 .014 11.463 102.353 **

LogTI -7.578 7.356 -1.03 .304 -22.058 6.903

Profitability -55.856 58.882 -0.95 .344 -171.766 60.053

FemBoardRep -.203 1.79 -0.11 .91 -3.726 3.32

Mean dependent var 7.090 SD dependent var 55.281

R-squared 0.244 Number of obs 349.000

F-test 1.303 Prob > F 0.072

Akaike crit. (AIC) 3832.618 Bayesian crit. (BIC) 4102.473

Referenties

GERELATEERDE DOCUMENTEN

archivalisch en een materieeltech- nisch luik. De vraagstelling, die door Peter te Poel van het Bonnefantenmuseum werd toegelicht, gaat als volgt: hoe meer te weten komen over

These results coincide with the results of Goergen and Renneboog (2004). Furthermore, 26% of the bidders make a deal that is diversifying and 30% of the targets is

It can be assumed that when organizations communicate their CSR activities with the information strategy or response strategy, the CSR videos will have fewer likes, views

De eerst bekende uitgever was Cornelis Banheyning (actief 1647-1657), daarmee moet de prent in of na 1647 zijn gemaakt.. Portret van Lodewijk de Dieu naar

This special issue is based on selected papers and extended papers of the Fifth International Conference on eHealth, Telemedicine, and Social Medicine, held on February 2013, in

Based on this thesis it seems that airlines in general are able to generate a positive abnormal shareholder return during the 5 years after a merger, as a significant positive

The enhanced deposition in the pneumatophore fringe followed from a sharp decline in the suspended sediment concentrations across this zone, as the reduction of near-bed velocities

This study has employed spatial hedonic speci fications to assess two spatial aspects in the marginal effects of distance to forests, parks, and fields on apartment prices: