1
Determinants of Abnormal Returns
An empirical research on what drives the abnormal returns resulting from a takeover announcement
Abstract
There are many different factors that influence the abnormal returns shareholders earn around takeover announcement dates. This thesis examines these abnormal returns for shareholders of the bidding firm. An empirical research is done on a sample of 150 takeovers in the United States for the period 2005 till 2009. In this sample, a cumulative abnormal return (CAR) of -1.13% was found. Because the sample period covers 2.5 years of the financial crisis, this factor was taken into account by dividing our original sample in two sub-samples: pre-crisis and during-crisis. In these samples of 89 and 61 takeovers
respectively, we found a CAR of -1.02% for the pre-crisis period and -1.28% for the during-crisis period. In all samples we found a positive relationship between the use of cash as means of payment and the CAR, which proves the positive effect the use of cash has. Also, in every sample we found a negative
relationship between the relative size of the target firm and the CAR, which means a better bargaining position of the target firm results in lower returns for the bidding firm. The period of the financial crisis showed us that in this period it were the financially oriented firms that performed on aggregate better than firms which were not. All results support findings of similar studies on this topic.
Nard Wiegerinck
10444475
Economics & Finance Finance
Number of credits: 12 ECTS Thesis supervisor: dr. Liang Zou
2
Statement of originality
This document was written by Nard Wiegerinck who declares to take full responsibility for the contents of this document. With this statement I declare that no other sources than the ones mentioned in the references section were used for this document. The University of Amsterdam is in no way responsible for the contents of this document.
3
Table of contents Page nr.
Abstract 1
Statement of originality 2
1: Introduction 4
2: Theoretical framework 5
2.1: Merger and acquisition waves 5
2.1.1: First and second wave 6
2.1.2: Third wave 7
2.1.3: Fourth wave 7
2.1.4: Fifth wave 8
2.2: Motives for takeovers 8
2.2.1: The synergy motive 9
2.2.2: Hubris 11
2.2.3: The agency motive 11
3: Results from the literature 12
3.1: Testing the CAR 13
3.2: Determinants of bidder returns 14
3.2.1: Firm-specific variables 15 3.2.2: Deal-specific variables 15 4: Empirical research 16 4.1: Data 16 4.2: Methodology 17 4.3: Hypotheses 20 5: Results 22 5.1: Summary 22 5.2: Regression analysis 26
5.3: Pre- and during-crisis 28
6: Conclusion 33
4 1: Introduction
The wealth effect from takeovers is a topic investigated a lot in the past decades. It should not come as a surprise that this has been done because takeovers can have a large economic and sociological impact. Employment rates, shareholder’s wealth, market competition and many more economic- and non-economic areas could be influenced by takeovers. A large driver for takeovers is the synergy potential between firms (Fuller, Netter & Stegemoller, 2002). Potential synergies influence shareholder wealth directly, both for target and acquiring firm’s shareholders. Whether these synergies are sought via a larger market share, better growth opportunities,
diversification, or some other value enhancing way, differs from one transaction to another.
Shareholders will only desire a takeover when they can gain from one. According to this reasoning the most important motivation for a takeover should be, at least from the shareholders point of view, to raise shareholder wealth. Recent studies showed that shareholders of target firms gain substantially from takeover announcements. For shareholders of the bidding firm, however, the wealth effects are ambiguous (Cybo-Ottone & Murgia, 2000).
Results from other studies, which will be examined later in this thesis, are in line with Malatesta’s (1983) finding that the net present value of a takeover is immediately captured in the share price when assuming an efficient capital market. When there is a belief that synergies between the firms exist, the share price will reflect this as soon as the market knows about the takeover, i.e., when the announcement is made. Firth (1980) came to the same conclusion. He beliefs that when we assume efficient capital markets, the effects of specific events, like takeover announcements, are captured in share price fluctuations around this date.
Although there is no consensus on whether the wealth effects are positive or negative for shareholders of the bidding firm, there are variables that influence the outcome of these returns. A well-known determinant influencing the bidder return is the method of payment (Fuller et al., 2002). The empirical reasoning behind this is that a
5
bidder will offer equity when its stock is overvalued, and will make a cash payment when it is undervalued. This is a well examined issue which will be taken into account in this thesis, besides other factors of influence. What those other determinants are and whether they have a significant influence seems worth investigating.
In this thesis, we will try to get an answer to the following dual research question: Do takeover announcements result in abnormal returns for bidding firm’s shareholders and what variables influence these returns?
This thesis will be organized in the following order: In chapter 2, a theoretical framework for takeovers will be given. This chapter explains the reasons why firms enter in an acquisition. Also, this chapter examines some explanatory variables others found that influence the abnormal returns. Chapter 3 will give an overview of the findings from similar studies in this field. The chosen methodology and data will be reviewed in chapter 4. Also, the hypotheses used to determine the answer to the research question will be stated. In the fifth chapter, the results from the event study and empirical analysis will be presented. Chapter 6 will summarize this thesis and will draw a conclusion regarding the hypotheses and the research question.
2: Theoretical framework
This chapter sheds some light on the theories behind takeovers. First, the largest merger and acquisition waves from the past will be addressed. Next, the reasons firms engage in takeovers are explained in more detail. Finally, possible explanatory variables for abnormal returns from the literature on this topic will be examined.
2.1: Merger and acquisitions waves
Takeovers tend to centre on specific periods. A cyclical pattern can be observed in takeover activity. This phenomenon is called a merger and acquisition, or takeover, wave (Martynova & Renneboog, 2005). In the literature, it has been brought forward that takeover waves are caused by shocks. These shocks can be of very different
6
origin: economic, regulatory, or technological (Harford, 2004). Whether these shocks result in a takeover wave depends on the overall capital liquidity. Martynova &
Renneboog (2005) argue that each wave has a unique set of motives for it to happen. However, there are also some common factors that influence these takeover waves. First of all, they found that every wave occurs in a period of economic recovery, which means the economy was hit by a depression or catastrophe of some sort before the wave. Next, these waves happen in periods with a fast growth of the credit- and stock market. All the large merger waves ended in the collapse of the stock market, but more on that later. Third, as already pointed out by Harford (2004), waves are preceded by shocks. The last common factor is the fact regulatory changes occur during these periods of takeover waves.
2.1.1: First and second wave
The first large takeover wave, which dates back to the 1890s, was called the Great Merger Wave. This takeover wave was mainly characterized by a desire to reduce competition and gain more market power (Martynova & Renneboog, 2008). The result was a couple of large firms in many different industries. The wave ended in the
beginning of the 20th century. The monopolies that emerged led to a lot of public protest. The centred market power came under attack, and anti-trust legislation was necessary to ensure a fair market. According to Sudarsanam (2003), the execution of these anti-trust laws was the dominant factor that helped starting the second takeover wave. This second wave, which started shortly after the First World War, ended with the start of the great worldwide depression, also known as Black Tuesday. The larger firms had to endure more competition during this second takeover wave, as this wave was characterized by the creation of oligopolistic markets. The focus during this second large takeover wave was more towards achieving economies of scale, instead of increasing market power (Martynova & Renneboog, 2005).
7
2.1.2: Third wave
Due to the aftermath of the great depression, and the subsequent economic downturn during the Second World War, the third large takeover wave started in the early 50s (Martynova & Renneboog, 2005). It lasted for almost 2 decades, and had its peak in 1968, before collapsing in 1973. As pointed out by Sudarsanam (2003), there was a difference between the waves in the US and the UK. The takeover wave in the US was characterized by diversification and a further development of the largest conglomerates as a motive for takeovers, while the wave in the UK had its focus on a more horizontal integration. A possible explanation for the move towards diversification has been given by Matsusaka (1993). He argues that so called managerial synergies can be achieved when the target’s management complements the core business of the bidding firm. Although it’s an expensive form of diversification, it is a rather fast option and the easiest way to obtain expertise in other markets. However, the literature has no unique explanation for the patterns of this third takeover wave (Martynova & Renneboog, 2005).
2.1.3: Fourth wave
The fourth large takeover wave started in 1981 and ended in 1989. This 80s wave started when the market recovered from the economic recession due to the oil crises from the 1970s (Martynova & Renneboog, 2005). This proofs once more the theory brought forward by Martynova & Renneboog (2008) that takeover waves occur during, or right after, a period of economic recovery. The fourth wave coincided with a couple of developments: the period was characterized by new financial instruments and markets, for example, the junk bond market. Also, there was technological innovation in the electronics industry. Furthermore, changes in anti-trust legislation and the deregulation of the financial sector lead to the realization of this fourth wave (Martynova &
Renneboog, 2005). These changes are in line with Harford’s (2004) theory that takeover waves are caused by shocks. This wave had its focus on the reversal of the inefficient diversifications of the previous takeover wave (Martynova & Renneboog, 2005). Due to larger competition, firms started to refocus on their core business, and the takeover
8
market was characterized by divestitures, leveraged- and management buyouts, and hostile takeovers.
2.1.4: Fifth wave
For now, the last large takeover wave started in 1993. This wave was, compared to the previous waves, of a more global nature. The deals during this wave were for a larger part cross-border driven (Martynova & Renneboog, 2005). The fifth takeover wave was, just as the previous ones, established trough economic, technological, and regulatory changes and shocks (Harford, 2004). Because of ever increasing competition,
previously home-based firms entered in overseas takeovers to be able to withstand the rough international competition. This was also the main goal of the takeovers during this fifth wave (Martynova & Renneboog, 2008). Firms engaged in global deals as a means to grow in the more internationally oriented markets of the late 90s.
All takeover waves had different, unique characteristics. However, there have been predominant factors that have been an influence on all these waves. One factor that seemed to recur frequently as a factor of influence is the preceded economic recovery. With the global markets recovering from the recent financial crisis, is it fair to expect a new wave emerging in the foreseeable future?
2.2: Motives for takeovers
In the literature, three major motives for takeovers have been brought forward: possible synergies, managerial hubris, and the agency motive (Berkovitch & Narayanan, 1993). The first motive, synergies, suggests firms engage in takeovers because it is more profitable to merge the resources of two firms instead of keeping them as separate entities. A synergy is the extra cash flow two firms can obtain when their resources are combined (Firth, 1980). Besides Firth, who acknowledges this in his article on the theory of the firm, most literature argues there are two main theories explaining why firms engage in takeovers. These two theories are: the neoclassical profit maximization
9
theory and maximizing management utility.
2.2.1: The synergy motive
According to the neoclassical profit maximization theory, firms are motivated by competitive market forces to maximize their shareholders’ wealth (Firth, 1980). In essence, firms will only enter in a takeover if it enhances the wealth of its shareholders. The synergy motive captures this theory. Synergies are retrieved if the sharing of resources results in enhanced competitive power and in greater cash flows than those that could be attained when the firms worked as separate entities (Shleifer & Vishny, 2002).
Synergies can appear in a lot of ways. Goergen & Renneboog (2002) define two broad types of synergies: operating synergies and informational synergies. Operating synergies between firms exist if the combined firm has economies of scale and/or scope which the separate firms could not achieve (Matsusaka, 1993). Economies of scale and scope come in many different ways. When economies of scale exist, the combined firm has lower cost due to the higher output the combined firm achieves (Goergen &
Renneboog, 2002). An example is greater market power in the existing market. Greater market power could mean a larger market share or more power in the price setting stage. Other possible examples are: lower fixed cost per unit or lower personnel cost due to the cutback of otherwise double executed jobs within certain departments (Espen Eckbo, Giammarino & Heinkel, 1990). Economies of scope occur when the combination of two firms results in lower cost due to the broader scope of products it produces
(Goergen & Renneboog, 2002). For example, the combined firm can now produce more products in the same production plant. Because now one plant has to be run, this
should result in lower cost.
The other broad category is the informational synergies. According to Modigliani and Miller’s theory (1958), financial synergies shouldn’t arise when a perfect capital market exist. Of course, in reality this theory doesn’t hold, largely due to the fact asymmetries between firms exist. Because asymmetries like transaction costs,
10
regulations and informational asymmetries exist, there is room for informational
synergies. Informational synergies arise when the value of the combined firm is larger than the sum of its parts (Goergen & Renneboog, 2002). Such synergies could, for example, arise when a cash-rich firm with a low growth potential merged with a firm with a high growth potential but limited financial resources. Because the funding for new profitable projects can now be done within the new firm it is potentially less expensive compared to the opportunity cost when the funds had to come from other sources, like leverage (Slusky & Caves, 1991). This idea is captured in the internal capital markets theory (Goergen & Renneboog, 2002). A firm can also engage in a takeover to use its target to receive tax benefits. When merged, a firm allows itself to have a higher level of debt. Besides the financial stability this might bring, it can also provide the firm with tax savings when an optimal debt level is reached. Also, the net operating loss of the target firm could be used to relieve its own tax burden.
If managers indeed serve in the shareholders’ interest, they would only acquire firms if the transaction is profitable for its shareholders. When takeovers are driven by these synergies the total gain to the target- and bidding firm’s shareholders will be positive. The ultimate division of the synergies between the two parties depends on the relative bargaining position (Berkovitch & Narayanan, 1993). When the target firm has some bargaining power, a larger part of the potential synergies will flow to the target firm’s shareholders. This could be the case when the target is relatively large compared to the bidding firm. Ceteris paribus, a larger target firm will have a better bargaining position compared to a smaller one (Fuller et al., 2002). Another possibility for a better bargaining position is increased competition in the bidding process. Potential buyers would bid up the price until the firm who values the target highest remains. This bidding process would reveal which firm has, or believes it has, the highest synergy value with the target firm, otherwise it would be outbid by another firm. This shows that there is room for human error when valuing a target.
11
2.2.2: Hubris
As just explained, takeovers can also result from managerial mistakes. This idea is captured by the hubris hypothesis (Roll, 1986). An example of a managerial mistake is the expectation that there are synergies between two firms, while in reality this turns out to be lower, or isn’t even there at all. As explained by Berkovitch & Narayanan (1993), the hubris hypothesis assumes managers make mistakes when evaluating a potential target. When they eventually engage in a bid on the target firm they will tend to overpay due to their overestimation of the potential synergies. Firth (1980) also points out that the gains for the target firm are merely a transfer of wealth from the bidding firm. This leads to the conclusion that there is no relationship between the total gains from the takeover and the gains for the target firm. Since the synergies resulting from the
takeover are near zero under the hubris hypothesis, the premium payment is basically a transfer from the bidding to the target firm. When the hubris hypothesis holds, the gains to the target- and bidding firm from the transaction should be negatively correlated, and change one-for-one (Berkovitch & Narayanan 1993).
2.2.3: The agency motive
Proponents of the maximizing management utility theory believe that managers try to maximize their self-interest in a takeover (Baumol, 1959). This manager incentive theory is captured in the agency motive as an explanation for engaging in a takeover
(Berkovitch & Narayanan, 1993). As argued by Berkovitch & Narayanan, there are numerous reasons to explain the divergence between a management’s interest and the maximization of shareholder’s wealth. An example is when managers engage in a takeover which increases the dependency on that very same management (Shleifer & Vishny, 2002). Managers will acquire firms in the line of business where their expertise lies (Berkovitch & Narayanan, 1993). As Firth (1980) points out, the success of the takeover, and thus the new combined firm, depends on a firm’s management. These actions will assure managers they won’t lose their jobs easily after a takeover. Other examples are: the wish to raise market power (empire-building), raising managerial salary or other form of income, or using diversification as a way to increase the
12
managers’ portfolio (Berkovitch & Narayanan, 1993). All these possible factors of influence are derived from the managers’ desire to a quick and large growth of the firm. As managerial income is almost always positively influenced by the size of the firm managers tend to favour a bigger firm. Besides that, the status of leading a large firm is something advanced as a reason managers engage in empire-building activities. But, ultimately, the agency theory suggests managers act in a way they maximize their own benefits from the takeover, where shareholder’s wealth isn’t a priority.
In essence, all these actions results in large agency cost. This will in turn reduce the value of the new merged firm. The ultimate bearers of these negative actions are the bidding firm’s shareholders. The reasoning behind this is that the target firm is specifically chosen by the manager to fit its interest best. When the target firm realizes this agency problem, it will try to exploit this managerial incentive (Berkovitch &
Narayanan, 1993). The part of the value obtained by the target is positively related to its bargaining power. In the end, the bidding firm’s shareholders lose more value the more severe the agency problem and the larger the bargaining position of the target are (Malatesta, 1983). As Berkovitch & Narayanan (1993) conclude, because the bidding firm’s gains are negatively correlated to the magnitude of the agency problem, the gains between the bidding- and target are inversely related as well.
This chapter showed the historical background behind takeover waves. Also the three large theories for engaging in a takeover were explained. The next chapter will present findings of similar articles on this topic. First, an overview of the results of the cumulative abnormal returns for the bidding firm as well as the target firm from different articles will be presented. In the end we will review possible explanatory variables for the bidder’s abnormal returns from these articles.
3: Results from the literature
The cumulative abnormal returns resulting from a takeover announcement have been a topic of interest for many years now. Empirical research has shown a lot on the wealth effects for bidder- and target firm’s shareholders. However, till this day no unanimous
13
conclusion is present on what the wealth effects exactly are. Some articles provide results confirming that takeover announcements have positive wealth effects for both groups of shareholders (Goergen & Renneboog, 2004). Others, however, have findings showing a negative abnormal return for bidder shareholders. The difference in the results on the wealth effect for the bidding firm’s shareholders is exactly why this topic has no conclusion yet (Fuller et al., 2002).
In the following section, we will review some articles that investigated this phenomenon. We will look at the wealth effects for both bidder- and target
shareholders, as both have been analysed a lot. The second and final section will give an overview of the most important factors of influence discussed in the literature.
3.1: Testing the CAR
A lot has been said about the cumulative abnormal returns around takeover announcement dates. The motivation to engage in a research on this topic varies across the literature. Dodd (1980), for example, looked at the effect of the
announcement as well as the effect of the completion or cancellation of takeover
proposals on abnormal returns. In his analysis he found that target shareholders earned a significant average of 13% cumulative abnormal returns around the announcement date. For bidding firm’s shareholders, there was an average loss of 6%. This is quite a difference between both parties in terms of returns.
In their paper, Goergen & Renneboog (2004) found somewhat different results. In their article investigating the wealth effects in intra-European takeovers they too found a large positive wealth effect for target firm shareholders of approximately 9%. For the bidder’s shareholders, the wealth effects were a positive 0.7% around the announcement date. They chose a small time frame around the announcement date as it is expected that all abnormality is captured in those days. Although both articles had the same reaction for target shareholders, Dodd’s (1980) research had a bit more negative results on the bidder’s abnormal returns.
14
A difference that arises between many of the articles is the time frame used to analyse the abnormal returns. Firth (1980) did a similar research to the one Dodd (1980) did during the same year. Firth (1980), however, used larger time frames. He looked at the effects years before or after the announcement date. The results for these returns were not significant, revealing that the effects are not to be found that far before or after the announcement date. His results for the target firm one month before or at the announcement date were positive and significant. For bidder shareholders the only significant returns were negative ones in the month of the announcement date.
The banking industry is the centre of interest in the article by Cybo-Ottone and Murgia (2000) on wealth effects in the European banking market. Just as Goergen & Renneboog (2004), they chose a small time frame to capture the abnormal returns. The results of both articles also align reasonable well. Cybo-Ottone and Murgia (2000) found a gain of 12.55% for the target shareholders, where the bidding firm’s shareholders lost 0.17% on average.
The difference in results across the literature could be due to numerous reasons. The choice for the time period, time frame, geographical area, or some other
requirement leads to different results. However, the majority of the literature found at least for target shareholders more or less the same results. On average the
shareholders of the target firm gain significantly around takeover announcements. The wealth effect for shareholders of the bidding firm is not unambiguously positive or
negative. This is one of the reasons this thesis focuses solely on the wealth effect of the bidding firm’s shareholders.
3.2: Determinants of bidder returns
In this thesis, we will try to answer the dual research question via en empirical research. Examining similar studies gives us possible explanatory variables which can be used in our regression analysis. The literature has shown us that bidder announcement returns depend on a variety of firm-specific- and deal-specific variables (Moeller, Schlingemann & Stulz, 2004). First of all we will examine the firm-specific variables that were used. In
15
the final section we will take a look at the deal-specific variables used in the literature.
3.2.1: Firm-specific variables
An important firm-specific feature is whether the target firm is publicly or privately held, or even if it is a subsidiary (Fuller et al., 2002). In their article it has been shown to be of significant influence on abnormal returns. It has been shown that the method of
payment has a different influence on returns for public and private targets. The method of payment will be covered later on.
The pre-acquisition debt level of the bidding firm has been addressed in different articles. An important feature of debt is the fact that the firm is obliged to repay it
eventually. This means a firm’s management cannot take on risky investment
opportunities as the providers of this debt have probably set some restrictions on the behaviour of the management. Besides this, debt levels are used for other reasons: tax shielding, signalling effect (Schlingemann, 2004). In his article, Schlingemann assumed a negative relation between abnormal returns and existing debt in a firm’s capital
structure.
3.2.2: Deal-specific variables
The relative size of the target firm has been an interesting point. As the relative size resembles the relative importance of the target compared to the bidder, it is often used as a measure of the target firm’s bargaining position (Moeller et al., 2004). However, as pointed out by Asquith et al. (1983), the relative size is also a measurement of the importance of the transaction for the bidding firm. In aggregate, however, a higher relative size means more of the gains from an acquisition will flow to the target firm’s shareholders. This means relative size has an expected negative relationship with the cumulative abnormal returns of the bidding firm’s shareholders (Moeller et al., 2004). This important feature will be taken into account in this thesis due to its proven explanatory power.
16
The method of payment has been proven to be one of the most important factors of influence on the abnormal returns from takeover announcements. Many articles, like for instance the one by Yook (2003), stress there is a signalling effect involved in the choice of the method of payment. Issuing equity for the funding of a takeover has been assumed to signal an overvaluation of the share price of the bidding firm (Bruner, 1988). This signalling effect has a negative influence on the abnormal returns for bidder’s shareholders. The use of cash as means of payment has proven to be beneficial for bidding firm’s shareholders by the majority of the literature (Peterson & Peterson, 1991). However, some articles found opposite results due to numerous reasons. Because of the proven importance of this feature, it will be included in the regression analysis.
The next chapter will explain the methodology and data collection used in answering the dual research question. Afterwards, there is a chapter presenting and reviewing the results of the regression analyses. In the final chapter a conclusion will be drawn on this topic.
4: Empirical research
The dual research question will be answered by examining a sample of takeovers. To do empirical tests, data on this topic had to be gathered using several databases. The next section describes how the data was collected. Afterwards the methodology used in this thesis will be described. Finally, the expectations will be presented in the
hypotheses-section.
4.1: Data
To test the dual research question, data had to be collected. First the Zephyr database was used to receive the desired sample. Some requirements on the takeovers were submitted into the database. All takeovers had to meet the following requirements: the transaction had to be completed within the time period 2005 till 2009; this covers five years of material. The decision for this time frame was made because it covers a fairly recent period, as the majority of similar articles are from several decades ago. Besides
17
that, the time period can be split into two sub-periods to look at effects of the financial crisis. In this thesis, the period 2005 till June 2007 will be addressed by the pre-crisis period. The during-crisis period is covered by the takeovers completed within the time frame of July 2007 till 2009. Next, the geographical area investigated will be the United States. The reason behind the choice for the United States as the area of interest was because of the fact that the financial crisis more or less started there. Furthermore, the United States are in general an area worth examining due to its importance on the well-being of the rest of the world. Next, we only include listed bidding firms and listed or delisted target firms in this sample. Because the share price of private firms does not necessarily reflect the market price, and because less precise information on those firms is available, private firms are omitted (Goergen & Renneboog, 2004). This all means that both firms engaged in the transaction had to be listed in the United States at the time of the takeover. In the takeover the bidder has to acquire all outstanding shares of the target firm, and cannot have had any interest in the target before the takeover. With this restriction we receive only the ‘purest’ cases, which is useful when examining the effects of the method of payment. The final requirement is that the deal size has to be $500 million or more. After controlling for these requirements, a sample of 150 takeovers remains.
After having selected the desired sample, the financial statistics for the sample had to be retrieved from the databases. For these numbers and other measures, the databases Compustat and Datastream were used. These databases were accessed to determine the abnormal returns and to find the values for the different variables of interest for each of the 150 takeovers. In the next section, the research method and all variables used will be explained.
4.2: Methodology
With an empirical research the wealth effects for bidding firms’ shareholders due to takeover announcements will be examined. A takeover should bring above normal returns to shareholders, reflecting an abnormal wealth effect in comparison to ‘normal’ times. The abnormal return measures this difference between the actual return and the
18
expected ‘normal’ return. When calculating the expected return of a security, its beta with respect to the market has to be known. The beta can be estimated by running a market model over a 6 month period which ends 1 month, or 20 trading days, before the announcement date (Brown & Warner, 1980). The month prior to the announcement date is omitted from this research period because it would influence the beta of the security, while the beta should measure volatility under ‘normal’ times. The equation for the market model is:
With an ordinary least square regression the parameters of this model, α and β, can be estimated
.
The market return (Rmt) is measured by the daily return of the S&P500 over the same 6 month period. The abnormal returns can then be calculated using the parameter-estimates derived before. The abnormal return-equation is:
Here, Rit is the actual return of a security on a certain date. We will follow Fuller
et al. (2002) in their choice for the duration of the time interval. The time interval studied will be [-2 , 2]: this means 2 trading days prior to the announcement date till 2 trading days afterwards. This period of 5 days should capture nearly all abnormal returns
resulting from the takeover announcement. When the abnormal returns are determined, the cumulative abnormal returns over the time interval can be calculated to test the abnormality of the wealth effect. The CAR is calculated with the following equation:
In the regression, the CAR is the dependent variable. For target shareholders the CAR has been shown to be unambiguously positive. This conclusion does not always hold for bidder shareholders. Via an OLS regression analysis we will look at what
19
determines the CAR around the announcement date of the takeover, and what the direction of those determinants is. The regression function will have the following form:
Yi = β0 + β1Crisis + β2Cash + β3Stock + β4Financial + β5Target + β6Bidder + β7Size +
β8Debt + β9Size*Cash + β10Size*Stock + εi
In table 1 the descriptions of all variables can be found.
Table 1
Description of the various variables
Variable Description Dependent Variable CAR Dummy Variables Crisis Cash Stock Financial Control Variables Target Bidder Size Debt Interaction Variables Size*Cash Size*Stock
Cumulative Abnormal Return for the bidding firm.
This variable has value 1 when the takeover happened during the financial crisis period, and value 0 if not.
This variable has value 1 when the takeover is financed with cash, and value 0 if not.
This variable has value 1 when the takeover is financed with equity, and value 0 if not (the alternative to Cash and Stock is a mixture of the two).
This variable has value 1 if the bidding firm is a financially oriented firm, and value 0 if not
This is a variable for the size of the target firm. It is measured as the natural logarithm of the transaction value.
This is a variable for the size of the bidding firm. It is measured as the natural logarithm of the bidding firm’s market
capitalization 1 fiscal year prior to the takeover.
This is a variable for the relative size of the firms. It is measured as the transaction value divided by the bidding firm’s market capitalization 1 fiscal year prior to the takeover.
This is a variable for the debt position of the bidding firm. It is measured as the bidding firm’s total level of debt divided by the market capitalization both 1 fiscal year prior to the takeover.
This is an interaction variable between the relative size of the target firm and whether the bidding firm financed the takeover with cash.
This is an interaction variable between the relative size of the target firm and whether the bidding firm financed the takeover with equity.
20
Following the majority of literature on this topic, we include the method of payment into the regression. Firms make choices regarding the medium of exchange in takeovers for numerous reasons: tax incentives, signalling effect, lack of liquidity, and more (Peterson & Peterson, 1991). The Cash and Stock variable account for the financing choices for the different takeovers. Just as articles like the one by Moeller et al. (2004) we choose to include the relative Size of both firms in the regression analysis. It has been shown to have a significant influence on the cumulative abnormal returns. As both the relative size and the medium of exchange have resulted in a significant impact in researches over the years, we will include interaction terms for them. It is possible that the combined effect is a significant factor of influence on the abnormal return for bidder shareholders. As this thesis also focuses on the effect of the financial crisis on abnormal returns, we will pay special attention to the dummy variable Crisis, representing whether a takeover took place before or during the crisis. The other variables have been proven to have an influence on abnormal returns and they will be included to gain more
statistical power.
In the final section we will look at some hypotheses regarding this empirical research. Via the regression analyses we will try to determine the outcome of these hypotheses, and use them in answering the dual research question set at the beginning of this thesis.
4.3: Hypotheses
By examining the literature on this topic some hypotheses will be set. First and foremost, the CAR will be tested. Next, we will set hypotheses for some of the independent variables:
Shareholders of the bidding firm have a negative wealth effect resulting from a takeover announcement. Following the majority of researches on this topic, we expect the CAR for the bidding firms to be significantly negative. We will have a look at the sample as a whole, and will also divide it into sub-samples to see the effects of certain
21
factors on this matter. As we expect this to be negative, a one sided T-test will be sufficient to check for significance.
The use of cash as means of payment in takeovers has a positive effect on the CAR of bidder shareholders. This is a common hypothesis followed by for instance Fuller et al. (2002), among others. Because some articles found no significant results (Peterson & Peterson, 1991), or possible negative effects, we will use a two sided T-test to check for a significant influence on the CAR.
Takeovers during the crisis in which the bidding firm is a financially oriented firm have a higher CAR than firms which are not. The reasoning behind this hypothesis is that during the early stages of the crisis it were especially the financial institutions that got into trouble. Because a lot of financially oriented firms had to be sold for them to be able to maintain operations, the ones buying these firms could get a better deal than they would ever get before the crisis. The few financial organisations that were still sufficiently stable to engage in takeovers could swoop in and buy out the less fortunate firms. Because their relative bargaining position was better than it would be pre crisis, their CAR is expected to be higher. Because this is no common hypothesis, and we have no preliminary certainty about the direction, we test for significance with a two sided T-test.
The relative size of the target compared to the bidder has a negative impact on the CAR of bidder shareholders. This is a hypothesis followed by Fuller et al. (2002). This variable looks at the relative importance of the target to the acquirer. The higher this relative value gets, the larger the bargaining position of the target firm will be. When the target firm has a larger bargaining position, it is able to obtain a larger premium from the bidding firm. This should thus in turn mean that the CAR of the bidding firm is
negatively affected by the relative size of the firms. We test this hypothesis by checking for significance using a two sided T-test.
In the next chapter the results of the regression analyses will be presented and examined. The hypotheses will be put to the test with these results. In the final chapter
22
a conclusion on the research question will be drawn with respect to these results.
5: Results
In this chapter, we will review the results of the regression analyses on our sample. First an overview of the most important statistics for the variables will be given. In the final two sections the results of the ordinary least squares regressions will be presented and reviewed.
5.1: Summary
In the previous chapter in table 1 the description of the variables could be found. Table 2 represent a summary of the statistics for these variables.
Table 2
Summary statistics for the total sample
Variable N Mean St. Dev. Min. Max.
CAR Crisis Cash Stock Financial Target Bidder Size Debt Size*Cash Size*Stock 150 150 150 150 150 150 150 150 150 150 150 -0.0113 0.4067 0.4733 0.2067 0.1533 21.665 23.285 0.5191 0.3476 0.1087 0.1459 0.0639 0.4929 0.501 0.4063 0.3615 1.0490 1.6093 0.6660 0.2329 0.2561 0.4085 -0.2088 0 0 0 0 20.0300 20.145 0.0027 0 0 0 0.2100 1 1 1 1 24.944 26.429 4.1282 0.9324 1.4679 2.6146
The average cumulative abnormal return for the whole sample was -1.13%. This negative return is what was expected at the beginning of this thesis. We do, however, still have to check this for significance. In the table we see that the relative size is almost 52% on average. The highest value for the relative size is over 400% of the
23
target value compared to the bidder. The debt-to-market value ratio is on average 34.76%, with the highest ratio of 93.24% and the lowest having no debt at all.
In table 3 and table 4 we present the same statistics after dividing the original sample in two sub-samples: one for the pre-crisis period and one for the during-crisis period. This way we can look at any differences in the statistics between these two periods.
Table 3
Summary statistics for the pre-crisis sample
Variable N Mean St. Dev. Min. Max.
CAR Cash Stock Financial Target Bidder Size Debt Size*Cash Size*Stock 89 89 89 89 89 89 89 89 89 89 -0.0102 0.5056 0.2135 0.1011 21.698 23.338 0.5201 0.3240 0.1244 0.1386 0.0644 0.5028 0.4121 0.3032 1.0659 1.6114 0.6018 0.2207 0.2919 0.3824 -0.2088 0 0 0 20.030 20.145 0.0027 0 0 0 0.1942 1 1 1 24.928 25.971 2.6146 0.9324 1.4679 2.6146 Table 4
Summary statistics for the during-crisis sample
Variable N Mean St. Dev. Min. Max.
CAR Cash Stock Financial Target Bidder Size Debt Size*Cash Size*Stock 61 61 61 61 61 61 61 61 61 61 -0.0128 0.4262 0.1967 0.2295 21.616 23.209 0.5176 0.3822 0.0857 0.1566 0.0635 0.4986 0.4008 0.4240 1.0306 1.6166 0.7552 0.2473 0.1922 0.4469 -0.1689 0 0 0 20.134 20.499 0.0054 0 0 0 0.2100 1 1 1 24.944 26.429 4.1282 0.8931 0.9143 2.5789
24
When looking at table 3 and table 4, we see an average CAR of -1.02% for the pre-crisis period compared to a -1.28% for the during-pre-crisis takeovers. This means that on average the bidding firms pre-crisis lost less from takeover announcements than bidding firms engaged in a takeover during the crisis period. However, these returns have to be checked for significance to be able to draw any conclusions from them. We see that the relative sizes of the two sub-periods are quite close: 52.01% and 51.76% for the pre-crisis- and the during-crisis period respectively. The bidding firms engaged in takeovers pre-crisis had a slightly higher debt ratio than takeover during-crisis with an average of 38.22% compared to an average of 32.40% respectively. Also, a higher percentage of firms engaged in takeovers during the crisis period were financially oriented firms. The pre-crisis percentage was 10.11% while the during-crisis sample showed 22.95% for the percentage of financial firms. The number of takeovers that met all the requirements is higher in the pre-crisis period than in the during-crisis period: 89 observations compared to 61 observations respectively. Whether this depends on the requirements set or
because fewer takeovers were completed in the second period is unfortunately unclear from these statistics.
As the first hypothesis stated that the abnormal returns from takeover
announcements were negative for bidder shareholders, we check for significance in a couple of sub-samples. The results are shown in table 5. As our hypothesis expected a negative relationship between abnormal returns and takeover announcements, a one sided T-test was done to check for the significance of the returns. We test the CAR for the sample as a whole, but will also divide it into the pre-crisis and during-crisis period. We will also check the CAR for significance with relation to the method of payment and on whether the bidder is a financially oriented firm.
We see that the -1.13% abnormal return for the total sample is statistically significant at the 5% level. For the precrisis and duringcrisis samples the 1.02% and -1.28% respectively are both statistically significant at the 10% level. The cumulative abnormal returns for takeovers financed with cash show a slight positive abnormal return of 0.72%. Although it is not statistically significant, it shows that cash offers have a probability of receiving higher returns than offers which are financed another way.
25 Table 5
CAR for the different samples. This table contains the different values for the cumulative abnormal returns. A test has been done to see whether the CAR was significantly negative under certain scenarios. First, the total CAR tests the CAR for the sample as a whole. Next, the CAR will be examined under restriction of the different dummy variables. A division will be made between the pre-crisis CAR and the during-crisis CAR to see the effect of the financial crisis on this issue. Furthermore, the CAR will be checked for takeovers financed with cash (Cash CAR) and for takeovers financed with equity (Stock CAR). Finally the CAR for takeovers in which the bidder is a financially oriented firm is tested (Financial CAR). As the effect was expected to be negative, a one sided T-test is done to check for significance. The number in brackets is the T-statistic. Superscripts *, ** and *** denote significance at the 10%, 5% and 1% level. CAR N Total CAR Pre-crisis CAR During-crisis CAR Cash CAR Stock CAR Financial CAR -0.0113** (-2.16)** -0.0102* (-1.50)* -0.0128* (-1.58)* 0.0072 (1.13) -0.0319*** (-2.51)*** -0.0045 (-0.34) 150 89 61 71 31 23
This can, for instance, be seen from the -3.19% abnormal return for takeovers financed with equity as it is statistically negative at the 1% level. Takeovers by financially oriented firms show a cumulative abnormal return of -0.45%, which is not statistically significantly below zero. The numbers in table 5 show that for various (sub) samples the return is on average significantly negative. This confirms our hypothesis that bidding firm’s
shareholders lose around takeover announcements.
In the next two sections we will review the regression analyses that were done in order to shed some light on the other hypothesis. We will look at the whole sample at first, but the sub-samples regarding the crisis period will be examined too.
26
5.2: Regression analysis
The hypotheses will at first be tested for the selected sample of 150 takeovers. Via three different regressions we will look at the effect the variables have on the CAR of bidder shareholders. In table 6 these results are presented. The first regression looks at the effect of the four dummy variables on the CAR. Special attention here is paid to the method of payment dummies. The descriptions for these dummy variables can be found in table 1. The second column presents the regression in which the CAR is regressed on the same four dummy variables and now the four control variables too. The last regression looks at the effect of the interaction variables between the relative size of the firms and the method of payment. The method of payment dummy variables are omitted from the third regression analysis because this enables us to see the effect of the
interaction variables without the effect of the dummy’s. This has been done to make sure the effects do not overlap in a way that makes us unable to draw any conclusions from the results.
By looking at the results in table 6 we see an interesting result for the Cash variable in the first regression. Takeovers financed with cash result in an average 3.94% higher CAR than takeovers financed with a mix of equity and cash. The mixture
financing is the alternative to Cash or Stock. This 3.94% is significantly different from zero at the 1% significance level. We also see that Financial bidding firms have an average 2.94% higher return than bidding firms which are not. This coefficient is significant at the 10% level. However, because the intercept is significant at the 1% level, we probably have to include more variables to be able to draw sensible
conclusions. The adjusted R-squared measures the explanatory power of a regression. For the first regression the adjusted R-squared is 0.0732.
In the second regression, the intercept is not statistically significant anymore. By adding the control variables, the Financial coefficient lost its significance too. We see from the adjusted R-squared, which is now 0.0973, that we obtained some more
explanatory power by adding the control variables. The coefficient on the variable Cash, which is 3.54%, is significant at the 5% level now. This means that after the addition of
27 Table 6
Regression output for the total sample. This table contains the regression output for three different regression analyses. In regression number 1, the CAR is regressed on the dummy variables. In the second regression control variables are added. In the final regression we look at the influence of the interaction variables on the CAR. The descriptions for the variables can be found in Table 1. The number in brackets is the T-statistic. Superscripts *, ** and *** denote significance at the 10%, 5% and 1% level.
(1) (2) (3) Crisis Cash Stock Financial Target Bidder Size Debt Size*Cash Size*Stock Intercept N Adj. R2 -0.0031 (-0.30) 0.0394*** (3.25)*** -0.0050 (-0.35) 0.0274* (1.83)* -0.0318*** (-2.96)*** 150 0.0732 -0.0043 (-0.42) 0.0354** (2.41)** -0.0098 (-0.69) 0.0107 (0.57) 0.0076 (1.15) -0.0100* (-1.88)* -0.0310*** (-2.63)*** 0.0376 (1.44) 0.0458 (0.40) 150 0.0973 -0.0016 (0.15) -0.0030 (-0.18) -0.0002 (-0.03) -0.0002 (-0.04) -0.0259** (-2.11)** 0.0325 (1.26) 0.0632*** (2.95)*** -0.0120 (-0.87) -0.0052 (-0.04) 150 0.1066
28
the control variables bidding firms financing their takeover with cash perform on average still better. When looking at the variable Bidder we see that controlling for the natural logarithm of the bidder’s market capitalization gives us a coefficient of -1.00% with a 10% significance level. The Size variable was added to see the effect of the relative size of the firms. This variable too is of special interest as it is used to determine the outcome of one of the hypotheses. In this regression, the coefficient on this variable is -3.10% with a 1% significance level. This means that a higher relative size of the target firm results on average in a lower CAR for bidder shareholders. For now, this confirms our hypothesis, but we will continue to examine this.
In the final regression the interaction variables were added. The only two
statistically significant variables are Size and Size*Cash. The coefficient on Size has the same direction and is now significant at the 5% level. What’s worth mentioning is that the coefficient on Size*Cash is 6.32% with a 1% significance level. This means that takeovers financed with cash result in a higher CAR for the bidding firm as the relative size of the target gets higher. The addition of the interaction variables resulted in even more explanatory power, as our adjusted R-squared in the final regression is 0.1066.
By looking at the results so far the hypotheses regarding the use of cash in the financing decision and the relative size hypothesis have proven to hold in our sample. In the next section we will divide the sample into two sub-samples: a pre-crisis sample and a during-crisis sample. This enables us to search for any differences due to the crisis.
5.3: Pre- and during-crisis
To see whether the financial crisis has had an impact on abnormal returns resulting from takeover announcements, separate samples were created and were tested for the same variables as the ones used in the previous regressions. The only difference arises due to the fact we omit the Crisis variable because of multicollinearity. The results of the three regressions for the pre-crisis period are presented in table 7. Again, in the first regression the CAR is regressed on dummy variables only. In the second regression the control variables are added, and in the third regression we add the interaction variables
29 Table 7
Regression output for the pre-crisis sample. This table contains the regression output for three different regression analyses. In this sample only the takeovers with value 0 for the dummy variable Crisis are taken into account. This is done to examine the difference between the pre- and during-crisis sample. In regression number 1, the CAR is regressed on the dummy variables. In the second regression control variables are added. In the final regression we look at the influence of the interaction variables on the CAR. The descriptions for the variables can be found in Table 1. The number in brackets is the T-statistic. Superscripts *, ** and *** denote significance at the 10%, 5% and 1% level.
(1) (2) (3) Cash Stock Financial Target Bidder Size Debt Size*Cash Size*Stock Intercept N Adj. R2 0.0344** (2.15)** -0.0130 (-0.69) 0.0003 (0.01) -0.0248* (-1.88)* 89 0.0705 0.0281 (1.37) -0.0172 (-0.88) -0.0287 (-1.00) 0.0011 (0.13) -0.0052 (-0.11) -0.0204 (-1.05) 0.0622* (1.67)* 0.0688 (0.41) 89 0.0695 -0.0380 (-1.37) -0.0055 (-0.68) 0.0024 (0.32) -0.0165 (-0.83) 0.0538 (1.43) 0.0443* (1.73)* -0.0146 (-0.73) 0.0446 (0.27) 89 0.0577
and lose the method of payment dummies. If we take a look at the coefficients of the first regression, we see that the effect of the use of cash in financing a takeover has a
30
significant effect on the CAR for bidder shareholders. Cash has an impact of 3.44% with a 5% significance level. The other dummy’s, and especially the Financial dummy, seem to have no significant influence on the dependent variable.
In the second regression we lose the statistical power from the Cash variable. After adding the control variables, the only significant coefficient is the one for the Debt variable. With a coefficient of 6.22%, this variable has a significant influence on CAR at the 10% level. From the adjusted R-squared we see that adding the control variables led to a weaker explanatory power from these variables. This is contrary to what we found in table 6 for the regressions on the whole sample, where adding variables resulted in more explanatory power on what determines the CAR for bidder shareholders.
In the final regression we see from the adjusted R-squared that adding the interaction variables resulted again in a lower explanation power from the regression. This is again the opposite of what we saw in table 6 on the effect of interaction variables on the whole sample. In the third regression in table 7, we see that we again have only one significant variable. This time it is the coefficient on Size*Cash. The coefficient is significant at the 10% level with an influence on CAR of 4.43%.
The pre-crisis period regression analyses tell us that controlling for the same variables results in lower explanatory power compared to the regressions under the whole sample. The fact we receive only one significant coefficient in both the second and third regression reveals that our analysis does not have an explicit conclusion on the pre-crisis period. This could mean that the explanatory power from the regression analysis for the whole sample presented in table 6 is due to the effect of takeovers completed during the crisis period. Next, we will have a look at the results for the during-crisis sample to see whether this is indeed the case.
The results of the same three regressions for the during-crisis period are presented in table 8. We again control in the three regressions for dummies, control variables, and finally interaction variables. If we take a look at the first regression, where we just control for the dummy’s, we already see a higher adjusted R-squared than
31 Table 8
Regression output for the during-crisis sample. This table contains the regression output for three
different regression analyses. In this sample only the takeovers with value 1 for the dummy variable Crisis are taken into account. This is done to examine the difference between the pre- and during-crisis sample. In regression number 1, the CAR is regressed on the dummy variables. In the second regression control variables are added. In the final regression we look at the influence of the interaction variables on the CAR. The descriptions for the variables can be found in Table 1. The number in brackets is the T-statistic. Superscripts *, ** and *** denote significance at the 10%, 5% and 1% level.
(1) (2) (3) Cash Stock Financial Target Bidder Size Debt Size*Cash Size*Stock Intercept N Adj. R2 0.0463** (2.48)** 0.0063 (0.29) 0.0487** (2.43)** -0.0450*** (-3.03)*** 61 0.0923 0.0454** (2.04)** -0.0005 (-0.02) 0.0421* (1.69)* 0.0179* (1.68)* -0.0162* (-1.89)* -0.0398*** (-2.69)*** 0.0222 (0.58) -0.0417 (-0.25) 61 0.1305 0.0257 (1.20) 0.0093 (0.93) -0.0030 (-0.38) -0.0311** (-2.01)** 0.0259 (0.72) 0.1173*** (2.76)*** -0.0108 (-0.57) -0.1531 (-0.91) 61 0.1653
under the same regression for the pre-crisis period. Here we find a R-squared of 9.23% compared to 7.05% we found in table 7. If we take a look at the coefficients, the
32
influence of Cash and Financial are both statistically significant at the 5% level. Especially the impact of the Financial variable is noticeable, as the effect of this
variable for the pre-crisis period was nearly zero with almost no explanatory power. For now, this confirms our hypothesis regarding the impact financially oriented bidding firms had on abnormal returns during the crisis. However, as our intercept is significant at the 1% level, we will control for more variables to be able to draw a sensible conclusion from the results.
In the second regression, after adding the same control variables as before, we find a much higher R-squared. By adding the control variables we improved our
explanatory power from 9.23% to 13.05%. We also see that we lost the significance of the intercept by adding these variables. As for the significance of the independent variables, we now control for five different significant variables. The Cash coefficient of 4.54% is again significant at the 5% level. We still see a significant influence of the Financial variable. The 4.21% coefficient lost a bit of explanatory power, but is still significant at the 10% level. We find a 10% significance for the Target and Bidder coefficients. What is important to note is the sign of the coefficients. The coefficient on the Target variable is 1.79%, whereas the coefficient for Bidder is -1.62%. This means that a higher market capitalization of the target results in a higher CAR for bidder shareholders, while a higher market capitalization of the bidding firm itself results in a lower CAR. With a coefficient of -3.98% and 1% significance we find the Size variable. We can conclude with a lot of certainty that the relative size of the target has a negative influence on bidder shareholders’ CAR. What is interesting to see is the influence of the method of payment on this issue in the third and final regression, where we add the interaction variables between the relative size and the method of payment.
First of all, we see an adjusted R-squared of 16.53% in the final regression. This means that adding the interaction variables resulted again in a higher explanatory power. From the results in table 8 it can be seen that the coefficient of the variables Financial, Target and Bidder lost their significance. Under this regression the hypothesis for the financially oriented firms performing better during the crisis period cannot be confirmed anymore. We see from the coefficient of the Size variable that it still has a
33
significantly negative influence on the dependent variable. With a coefficient of -3.11% and a 5% significance level the negative relation remains between Size and CAR. The fact we lost a bit of power here comes from the fact we added the interaction variables. The combination of the size and the use of equity as means of payment has no
significant influence on CAR in this sample. The Size*Cash variable, on the other hand, does have an impact on the dependent variable. With a positive 11.73% impact, and a 1% significance level, we see that Size*Cash has something to say about the meaning of the CAR. What might be surprising is the direction of this coefficient. We already found a positive coefficient in the various regressions for the Cash variable, but found the opposite results for the Size coefficient. The fact that we find such a high positive value for the interaction variable could be due to a couple of reasons. It is possible that during the crisis the use of cash in a relatively large takeover signalled a strong belief in the capabilities of the collaboration of the firms, and on the well being of the bidding firm pre-acquisition.
After reviewing the results from table 6 through 8, conclusions regarding the hypotheses will be drawn in the next chapter. With the help of these conclusions we will try to receive an answer to the dual research question that was set at the beginning of this thesis.
6: Conclusion
In this final chapter a conclusion will be drawn from the analyses of this thesis. First the hypotheses that were set are covered. With the help of the results and the conclusions from these hypotheses we will answer the dual research question: Do takeover
announcements result in abnormal returns for bidding firm’s shareholders and what variables influence these returns?
The wealth effect resulting from takeover announcements has been studied extensively. While it has been shown to be positive for shareholders of the target firm, this conclusion could not be drawn for shareholders of the bidding firm. As was our first hypothesis, it was expected that bidding firm’s shareholders lose around the date of a
34
takeover announcement. With the help of the test statistics, we are able to confirm this hypothesis. In table 5 the results for various one sided T-test were presented examining the direction of the cumulative abnormal returns. For the whole sample, the two sub-samples created around the crisis period, and the sample of equity financed takeovers only, we saw a significantly negative CAR at the 10%, 5% or 1% level. This is in line with results of similar studies, and confirms our first hypothesis.
In the second hypothesis we stated that the use of cash as means of payment in takeovers has a positive effect on the CAR of bidder shareholders. In the six different regressions were the Cash variable was included, we found five cases of a significant positive influence. The three remaining regressions were we controlled for the
interaction variables between the relative size and the method of payment all resulted in a significant influence of the use of cash on the CAR of bidder shareholders. From these results it is safe to say we can confirm the second hypothesis that the use of cash is of positive influence on the CAR.
The third hypothesis was a bit more specific. There we stated that takeovers during the crisis in which the bidding firm is a financially oriented firm have a higher CAR than firms which are not. When we look at the three regressions from the during-crisis sample, we see a significantly positive coefficient two out of three times. Only when the interaction variables are included we do not find any significance. From these results we can partially conclude that financially oriented firms performed better during the crisis regarding their shareholders’ CAR around the takeover announcement date. However, these results could be due to some other reason. More research on this issue would be wise as this hypothesis was not backed by similar studies and because the crisis is a fairly recent event.
Under the final hypothesis we stated that the relative size of the target compared to the bidder has a negative impact on the CAR of bidder shareholders. This was a more common hypothesis, and when reviewing the results from table 6 through 8 we can confirm this hypothesis. In the three different regressions were only the Size
variable, and not the interaction variables with the relative size, was included, we found two cases of a 1% significance. These cases of significance were from the regressions
35
of the whole sample and the during-crisis sub-sample. When adding the interaction variables, the results are somewhat different. For the Size*Stock variable we find only negative coefficients, showing a negative relationship between the relative size,
controlled for only equity funded takeovers, and the CAR of bidder shareholders. But, as there were zero cases of significance, we cannot conclude anything from this. On the Size*Cash variable, however, we found three significantly positive coefficients. As already explained, this is largely due to the signalling effect both methods of payment have on this issue, and no difference in the effect of the relative size arises because of these results. As our findings regarding this hypothesis are in line with the literature, we can confirm that the sole effect of the relative size of the target firm is of negative
influence on the CAR of the bidding firm’s shareholders.
We have set these hypotheses as means to answer our dual research question. This question was as follows: Do takeover announcements result in abnormal returns for bidding firm’s shareholders and what variables influence these returns? After reviewing the first hypothesis, we indeed find abnormal returns for bidding firm’s shareholders. The results showed significant negative returns around the
announcement date for those shareholders. As we find similar results in the literature, this thesis confirms the findings of earlier literature on this topic. Under the other three hypotheses we tested the influence of various variables on the CAR. We found a significant influence on CAR by the use of cash as means of payment and through the relative size hypothesis. Also, during the crisis period we found a significant positive influence of the Financial variable on this matter. These variables influence the abnormal returns of the bidder shareholders around the announcement date.
As we could not cover every possible aspect of influence on the CAR, this topic remains an issue worth investigating. The interaction variables between the relative size and the method of payment showed some interesting results. The effect of the firm-specific measures like Tobin’s Q or a firm’s credit rating, which were not covered in this thesis, could have some importance on this matter too. Also, the effect of subjective measures like the number of days between the announcement and completion date are variables that could be investigated in a further research. The many possible variables
36
not dealt with in this thesis are a motivation to pursue testing this topic in more complex regression and quantitative analyses, as the takeover market is of large importance to every nation’s economy.
