The Acquisition Premium
An empirical analysis of public-to-private transactions in the period 2005-2010.Christine S. van den Bos
December 2011
UNIVERSITY OF GRONINGEN Faculty of Economics and Business
MSc Business Administration Finance
C.S. van den Bos University of Groningen 2
The Acquisition Premium
An empirical analysis of public-to-private transactions in the period 2005-2010.
ABSTRACT
This thesis is based on a sample of 4,986 worldwide public-to-private transactions that took place from 2005 to 2009. The objective of this analysis is to study: (1) the impact of a transaction announcement on the target’s share price and (2) the hypothesis that well-informed acquirers pay lower premiums than less well-well-informed buyers. We find evidence that the acquisition announcement causes the share price of the target company to increase significantly with 8.37% on the event date and a cumulative average abnormal return of circa 12.87% is measured in the event window [-2; +2]. In addition the median premium we found is 17.50%. Furthermore, we find evidence in favour of the hypothesis that well-informed bidders pay lower premiums: private equity pays less than strategic bidders, consortiums pay less than individual buyers, and industry insiders pay less than industry outsiders. Finally, we find evidence that acquirers based in the European Union pay significant lower premiums than American based bidders.
C.S. van den Bos Saxen-Weimarlaan 43 1075BZ Amsterdam
C.S. van den Bos University of Groningen 3 PREFACE
I would like to thank my supervisor, Andrea Schertler, for her guidance the past months. She helped to operationalize my ideas and provided me with useful feedback. I am also very grateful to my colleagues of Ernst & Young who helped to find my way in the Bloomberg database and who always made time for a brainstorm session. Moreover, I am very thankful for the assistance of Robert-Jan Maaskant and Jelle Makkinga with respect to the data set. Finally, I would like to express my gratitude to my parents, my brother and sister for their continuous support and their comments on previous versions.
C.S. van den Bos University of Groningen 4 TABLE OF CONTENTS ABSTRACT 2 PREFACE 3 1. INTRODUCTION 6 2. LITERATURE REVIEW 8 2.1 Value creation 8
2.2 Information asymmetry and resource based theory 10
3. DATA AND METHODOLOGY 14
3.1 Data selection procedure 14
3.2 Abnormal return 15
3.3 Acquisition premium 18
3.4 Normality of the distribution 19
4. RESULTS 20
4.1 Descriptive statistics 20
4.2 Event study 21
4.3 Sensitivity analyses of the premium 24
5. REGRESSION ANALYSIS 27
5.1 Regression methodology 27
5.2 Regression results 28
6. CONCLUSION 33
6.1 General conclusions 33
6.2 Limitations of this study 34
C.S. van den Bos University of Groningen 5
REFERENCES 36
DATABASES 40
LIST OF TABLES 41
APPENDIX A - Regression variables 42
C.S. van den Bos University of Groningen 6 1. INTRODUCTION
Due to the financial crisis that started in 2008, the financial markets are subject to radical changes and will not be the same as in the past decennia. For example, bidders have more difficulty finding financing for their transactions (Ang and Mauck, 2011; Zhou, Li and Svejnar, 2011), and the implementation of Basel II has far-reaching consequences for banks (Lenz, 2011). We believe that the financial markets are currently at a tipping point and this makes acquisitions and their corresponding premiums an interesting topic for a research. Once a company decides it is interested in an acquisition, it needs to decide which price it is willing to pay. The price per share is usually higher than the current stock market-price and therefore the term acquisition premium is used. Hayward and Hambrick (1997) argue that acquirers in general pay substantial premiums, but sometimes it occurs due to extraordinary circumstances that there is actually an acquisition discount. The shareholders of the targeted firms are likely to be aware of these premiums and therefore the share price will increase when an acquisition is announced. Although we expect the share price to have a positive reaction following the acquisition announcement, we do not know how much the share price will increase. Therefore, the first research question is formulated as follows: To what extent does the announcement of an acquisition have a significant positive effect on the target’s share price?
Although recognised potential acquirers of listed firms have access to the same data room in order to perform a due diligence review, the final offer varies among bidders. Multiple researchers have analysed the determining factors of the acquisition premiums and they attributed the size of the premium to different causes. The main objective of this thesis is to study the effect of several factors, which according to literature are related to asymmetric information between potential buyers, on the acquisition premium. This results in the second research question: Do well-informed buyers pay a lower premium than poorly informed buyers and if yes, to what extent?
C.S. van den Bos University of Groningen 7 premiums of which the calculation was based on different post-event windows, which is incorrect in our opinion. We believe that the weekly premium, which includes a correction for the different post-event windows, is more suitable to compare the different transactions.
C.S. van den Bos University of Groningen 8 2. LITERATURE REVIEW
In this chapter we will review several previous papers, and formulate our hypotheses. In the first part we focus on the reason why acquirers are interested in buying another firm and the influence of the transaction announcement on the target’s share price. In the second part we present the asymmetric information and the resource based theory and we will formulate our hypotheses that are related to these theories.
2.1 Value creation
In order to better understand the premiums buyers are willing to pay, it is essential to know the motives a buyer might have for the acquisition. Different reasons why a company would be interested in acquiring another firm are discussed by previous researchers. Walsh and Seward (1990) and Berkovitch and Narayanan (1993) distinguish three main motives: synergies, poor target management, and hubris. Strategic buyers have the opportunity to create synergies by generating economies of scale (Bradley, Desai and Kim, 1988; Slusky and Caves, 1991) and by profiting from complementary resources (Wansley, Lane and Yang, 1983). Private equity firms, on the other hand, could merge firms that are in their portfolio in order to create synergies. Poor target management suggests that a company would perform better if its management is replaced. According to Fama (1980) the stock of the target company under current management is likely to be underpriced and will therefore become an attractive target to acquirers. Potential buyers believe that they can eliminate the target’s inefficiencies and consequently maximize stockholder value. Closely related to poor target management is the hubris of the bidder’s management. March and Shapira (1987) and Hayward and Hambrick (1997) describe hubris in a corporate context as the extreme self-confidence of some managers. These managers are convinced that they can make a success of every acquisition and that they will certainly outperform the current target management, underestimating the risk of failure. Moreover, Roll (1986) argues that hubris causes managers to believe that the market capitalization does not reflect the combined economic value of both firms. Another type of hubris that is a motive for acquisitions is the wish to control a firm as large as possible, because the larger the firm the more status the management will obtain, megalomania (Peng and Heath, 1996).
C.S. van den Bos University of Groningen 9 different markets, the risks faced by the acquirer will spread and perceived risk reduction might therefore be a reason to expand through acquisition.
Although the studies mentioned above state different reasons for transactions to occur, they all agree on the main reason of the transaction: value creation. Therefore, we believe that the reasons described above should be considered as means to create value and that value creation per se is the ultimate reason why a company is interested in buying a firm. This justifies the observation that the bidder is willing to pay a premium on top of the share price as long as the premium does not exceed the expected benefits. The shareholders of the target expect the buyer, for reasons discussed above, to pay a premium. Consequently, the share price of the target is likely to increase as soon as the transaction is announced. Both the influence of the takeover announcement on the target’s share price (table 1) and the size of the acquisition premium (table 2) are analysed by multiple researchers.
Table 1
Overview previous event studies
The table contains a summary of the main characteristics and findings of previous research that analyse the AAR (Average Abnormal Returns) and the CAAR (Cumulative Average Abnormal Returns). “Period” reflects the years that are analysed. The number of announcements is the actual number of events that are included in the analysis. “AAR” presents the AAR at the event date and “CAAR” presents the CAAR in the event window.
Author(s) Period # of events Country Est. Window Event window Result (AAR) t=0 Result (CAAR) Dodd (1980) 1970-1977 151 US N.A. [-40; 0] Significant, 8.74% α=0.01 Significant, 24.0% α=0.01 Keown and Pinkerton (1981) 1975-1978 194 US N.A. [-60; 0] Significant, 12.02%, α=0.01 Significant, 25.3%, α=0.01 Wansley et al. (1983) 1970-1978 203 US [-200; -50] [-40; +40] Significant, 7.22%, α= 0.01 Significant, 25.19%, α= 0.01 Boone and Mulherin (2007) 1990-1999 400 US N.A. [-1; +1] N.A. Significant, 21.6%, α= 0.05 Alexandridis, Petmezas and Travlos (2010) 1990-2007 4,577 49 countries N.A. [-2; +2] N.A. Significant, 17.60%, α=0.01
C.S. van den Bos University of Groningen 10 to different premium calculations (e.g. different time frames) and to the different share values used (e.g. dividend adjusted or not). Although the extent of the reaction of the target’s share price differs among the previous researches, we expect, based on the arguments of value creation mentioned above, the share price to increase when a public-to-private transaction is announced. Therefore, our first hypothesis is that the announcement of an acquisition has a significant positive influence on the target’s share price.
Table 2:
Overview previous research on acquisition premiums
The table contains a summary of the main characteristics and findings of previous research that analyse the acquisition premium. “Period” reflects the years in which the analysed deals took place. The number of transactions is the actual number of deals that are included in the analysis. The column with the presented premium contains the mean [median] premiums. Some researchers have distinguished different bidders and calculated the corresponding premiums, which is indicated in these cases.
Author(s) Period # of
transactions Country Premium Jarrell and Poulsen
(1989) 1960 - 1985 663 US
1960 - 1969: 19% 1970 - 1979: 35% 1980 - 1985: 30%
Walkling and Edmister
(1985) 1972-1977 158 US 52%, [47%] 1973-1979: [47.2%] 1980-1989: [37.7%] 1990-1998: [34.5%] 1973-1998: [37.9%] 1973-1979: 74%, [65%] 1990-1999: 47%, [42%] Varaiya (1987) 1975-1980 77 US 69,7% 46.5% (public bidder) 40.9% (private bidder) 28.5% (PE)
Slusky and Caves (1991) 1986-1988 100 US 50.50%
Hayward and Hambrick
(1997) 1989 and 1992 106 US 49%
205 (SB) 54.4%, [41.6%]
205 (PE) 42.5%, [39.5%]
Mnejja and Sahut (2010) 2004-2007 200 France UK
PE paid 16-24% less than SB. Andrade et al. (2002) 1972-1998 4256 US Gondhalekar et al. (2002) 1973-1999 703 US 1667 Bargeron et al. (2008) 1980-2005 US Roosenboom, Fidrmuc and Teunissen (2009) 1997-2006 US
2.2 Information asymmetry and resource based theory
C.S. van den Bos University of Groningen 11 Another reason why some bidding firms might be better informed than their competitors can be explained by the resource based theory (Pringle and Kroll, 1997): the acquirer possesses a unique resource that enables the firm to obtain extra or more precise information. Therefore well-informed parties do not overvalue the company or the opportunities of an acquisition, which means that they pay lower acquisition premiums than poorly informed bidders (Wilson, 1967; Hendricks and Porter, 1988; Moeller, Schlingemann and Stulz, 2007; Dionne, La Haye and Bergères, 2010).
In order to test the previous statement, we formulate the following hypothesis: well-informed buyers pay smaller acquisition premiums than less well-informed parties. The target’s management may decide for different reasons to support and accept a lower bid, for example when the management itself can stay or when it concerns a management buy-out (Lee et al, 1992). So this hypothesis does not mean that for a given deal, the probability that a well-informed bidder closes the deal is 0.
The hypothesis is subdivided into different sub-hypotheses that are presented below and that are related to the information asymmetry theory and/or the resource based arguments.
Existing target shareholders
According to Heflin and Shaw (2000), Chen, Harford and Li (2007) and Edmans (2009) acquirers who are block holders, those who own at least 5% of the shares of the target company, have preferred access to management due to monitoring activities. This results in an information advantage over non-block holders. We assume that the more shares one owns, the accessibility of the management becomes better and therefore the shareholder has better resources. Hence, we hypothesize that the more target shares a potential buyer owns before the transaction announcement, the lower the acquisition premium will be. This means that the hypothesis is not limited to the 5% ownership and takes into account all existing shareholders and distinguishes all different sizes of shareholders.
Private equity bidders versus strategic bidders
C.S. van den Bos University of Groningen 12 who can profit from synergies. We will therefore hypothesize that private equity partners pay a significant lower premium than strategic buyers.
Consortiums
Consortiums have an information advantage compared to individual buyers, because the participants exchange the information that they possess (Boone and Mulherin, 2008; Officer, Ozbas and Sensoy, 2009). This will result in more precise information. In addition, consortiums share their post-acquisition risks and profits; hence the risk profiles are different from the ones faced by individual buyers (Fleck et al., 2005). Therefore we formulate the following hypothesis: Consortiums pay significant lower premiums than single buyers. Moreover, we hypothesize that consortiums negatively influence the acquisition premium.
Industrial activity
Industry insiders have an information advantage over industry outsiders, because they are familiar with the target’s markets and assets. This enables them to correctly value the assets (Shleifer and Vishny, 1992). An exception is made for cross-industry private equity deals, because these often concern management buy-outs as discussed above. For this reason we only take deals of strategic buyers into account for this sub-hypothesis: strategic buyers offer larger premiums for targets in other industries than for targets active in the same industry. Therefore, we also hypothesize that industry insiders have a negative influence on the acquisition premium.
Geographical location
C.S. van den Bos University of Groningen 13 Friendly bidders
C.S. van den Bos University of Groningen 14 3. DATA AND METHODOLOGY
This chapter starts by discussing the sample selection in section 3.1. Subsequently, the models used to perform the event-study will be explained in section 3.2. Section 3.3 defines the equation for the acquisition premium. The last section, 3.4, of this chapter presents the statistics used to test for normality.
3.1 Data selection procedure
The sample dataset that fulfils all requirements of the proposed research consists of 4,986 deals, which is a combination of the deals available in the databases published by Bloomberg and Zephyr. Both databases are well-known and have a large set of information with respect to acquisitions and the related companies. Although both databases are subdued to the same limitations there is only an overlap of 1,175 deals after applying the selection criteria.
First of all it is necessary to define the concept of an acquisition, since various definitions are formulated in literature. In this thesis we follow the definition formulated by Zephyr: an acquisition is any deal where the buyer ends up by owning more than 50% of the target’s shares. An additional advantage of this definition is that as soon as a company or institution owns more than 50% of the shares of another company, the acquirer in principle, depending on local legislation, has voting control (Prôa Bessane and Verdini Maia, 2010). The combined databases make notice of 42,324 acquisitions that involve listed target firms. We only analyse listed targets in order to ascertain the availability of information, such as the target’s share price and the offer price. In addition, we limit the research by analysing acquisitions that were announced in the period starting January 2005 up to and including December 2009. This reduces the number of deals in the dataset to 17,373.
Furthermore, we only take into consideration those deals that were completed by the end of 2010(Keown and Pinkerton, 1981; Bargeron et al., 2008). This results in 13,765 remaining deals. Prematurely terminated deals and those still pending are deleted from the dataset. This means that only successful deals are analysed and this ensures that the results will not be influenced by failed offers. Although withdrawn transactions are likely to have the same announcement effects as realised transactions, the calculation of the acquisition premium requires the target’s share price at the completion date which does not exist for withdrawn transactions.
C.S. van den Bos University of Groningen 15 premiums are larger for cash payments than for payment in securities, due to tax effects. According to Gondhalekar, Sant and Ferris, (2004) non 100% cash payments might introduce contaminating financial structure considerations. Finally, Moeller, Schlingemann and Stulz (2007) argue that stock offers reduce overvaluation, because target shareholders share in any subsequent decrease in the bidder’s stock and will therefore reduce the information asymmetry between bidder and buyer. Consequently, including both 100% cash bids and equity offers in the analysis increases the probability of biased results. For these reasons, just like Bargeron et al. (2008), we only take 100% cash deals into consideration. This brings the number of deals back to 7,642.
Among the remaining deals are several deals consisting of multiple targets, which are all deleted because the premium per target cannot be distinguished because the data source only gives the total transaction amount. In case the acquiring party consists of multiple firms however, the deal is preserved and the composition of the consortium is recorded for further analysis. Finally, if deals occur in the sample more than once due to an overlap between Zephyr and Bloomberg, only one observation is preserved. In case of disparity between the overlapping deals (387 cases) the companies’ websites and the databases of MergerMarket and OneSource are used to decide which data is most accurate. This results in a total sample of 4,986 deals.
The characteristics of the target, the acquirer, the deal and the market, which are necessary for the analyses, are collected from Bloomberg and Zephyr. The industry sectors are classified based on the US SIC codes, the United States Standard Industrial Classification codes. To ensure that private equity parties and strategic buyers are distinguished properly, the company description in the Bloomberg database is read and in case of any doubt the firm’s website is consulted.
3.2 Abnormal return
C.S. van den Bos University of Groningen 16 In order to calculate the abnormal return, we will use three models discussed by Brown and Warner (1980, 1985): (i) the Mean Adjusted Returns model, (ii) the Market Adjusted Returns model and (iii) the Market and Risk Adjusted model. These models calculate the abnormal return after determining the expected return, which is based on the return of stock i or the market. The MSCI world is used as index, because the database contains stocks of different countries.
The observed return is the actual return on a stock. Formula [1] shows the equation of the observed return, where is the observed return of stock i at time t and Pt is the price of
the stock at time t.
= ln ( ) − ln ( )
[1]
The Mean Adjusted Returns model [2] calculates the abnormal return by taking the difference between the observed return and the expected return. Equation [2] is the formula of the abnormal returns of stock i at time t (ARit) and is the expected return of
stock i. The expected return is the average of the estimation window. In accordance with the Capital Asset Pricing Model we assume that both the systematic risk of the stock as well as the return is constant.
= −
[2]
The Market Adjusted Returns model [3] calculates the abnormal return by adjusting the observed return for fluctuations in the market. is the return of the market index at time t. This means that we assume the expected market return at time t to be equal for every stock.
= −
[3]
The Market and Risk Adjusted Returns model [4] is different from the previous discussed models, because it takes into account the systematic risk of stock i () and the non-market related risk of stock i ().
= − −
C.S. van den Bos University of Groningen 17 Subsequently, the average abnormal returns (AAR) per day in the event window are calculated [5], where N is the number of events.
=∑ ,
[5]
In addition the cumulative average abnormal return (CAAR) [6] is calculated by adding up the AAR’s in the event window. Calculating CAAR is necessary in case the news of the upcoming transaction is out in the open a few days before the official announcement, due to rumours or a leakage.
, = , + ,
[6]
Furthermore, both a t-test [7] and a Wilcoxon signed rank test [8] are performed in order to test whether the announcement of an acquisition causes the AARs and CAARs to be significantly larger than the expected returns. A one-sided test is required, because previous research has proved a positive relationship between the announcement and the target’s share price. The t-test is a parametric test, which means that the data is assumed to be normally distributed. The t-statistic compares the estimated value of a parameter with its hypothesized value, scaled by the standard error of the estimate.
| | = − " ! #
[7]
Where t is the test statistic, is the estimated value of the parameter, ! is the hypothesized value of the parameter, and "# is the standard error of the estimate.
The Wilcoxon signed rank test is a nonparametric test, which means that the data is not assumed to be normally distributed. The Wilcoxon signed rank test is used to test the median difference in paired data.
$%= & '
C.S. van den Bos University of Groningen 18 Where φi is 1 if Xi > Yi and φi is 0 if Xi < Yi. The cases where Xi = Yi are excluded. Ri is the
rank, where the smallest positive difference between Xi and Yi gets rank 1, and to tied scores a median rank is assigned.
3.3 Acquisition premium
The acquisition premium is calculated as the percentage change in share price caused by the take-over [9].
()*+,*= ln - ,. ,(/ 0)1
[9]
Where Premiumi is the percentage lognormal premium paid by the acquirer for target i. Pi,C
is the target’s share price at the completion date and Pi,(A-n) is the target’s share price n days
prior to the announcement date of the takeover.
By estimating the premium as a percentage, scaling errors due to the equity value of the target firm do not occur. Some researchers base the premium on the target’s share price 20 trading days before the official announcement date (Varaiya, 1986; Varaiya and Ferris, 1987; Barclay and Warner, 1993). However, Schwert (1996), Gondhalekar, Sant and Ferris (2004) and Betton, Eckbo and Thorburn (2008) state that the run-up associated with the acquisition occurs prior to these 20 days and therefore suggest to look at the share price 40 trading days before the announcement. Hence, in this study we will examine the share price 40 trading days prior to the event (n=40).
The number of days between the announcement date and the completion date differs per transaction. In order to enable a comparison of the premiums and to exclude differences due to the length of the period, the holding period return (HPR) is calculated [10].
2 = - ,. ,(/ 0)1
⁄ − 1
[10]
C.S. van den Bos University of Groningen 19 3.4 Normality of the distribution
In order to test the normality of the distribution, a Jarque-Bera (JB) test is performed. JB tests whether the skewness [11] and the excess kurtosis [12] are 0. A skewness of 0 means that the distribution is symmetric about its mean, whereas excess kurtosis measures to what extent the fatness of the tails deviates from a normal distribution.
"5)6 =(:7,;)8 ;8⁄9
[11] <,( =7,(:;)=;9
[12]
Where E[u] is the expected error and :; its corresponding variance. As a consequence, the JB-test [13] is formulated as follows:
>? = @"5)66 +; (<,( − 3)24 ;E
C.S. van den Bos University of Groningen 20 4. RESULTS
In this chapter the descriptive statistics and the results of the analyses are presented and discussed. First, the descriptive statistics are presented in section 4.1. Secondly, the results of the event study are presented and discussed in section 4.2. Subsequently, the findings of the sensitivity analyses will be presented and discussed in section 4.3.
4.1 Descriptive statistics
The descriptive statistics of the returns in the estimation window [-250;-41] are presented in table 3. Both the average and the median return are negative values for the two market adjusted models, indicating that for a given deal the stock price in the estimation window is outperformed by the market. Jarque-Bera is not significant, which indicates that the data is normally distributed.
Table 3
Descriptive statistics of the return in the estimation window
The statistics are based on the three models discussed by Brown and Warner (1980, 1985), using [-250;-41] as an estimation window. The number of observations is N=4,986.
Mean Adjusted Returns model
Market Adjusted Returns model
Market and Risk Adjusted Returns model Average 0.00% -0.06% 0.00% Median 0.00% -0.07% -0.01% Minimum -0.27% -0.34% -0.26% Maximum 0.20% 0.14% 0.22% Standard deviation 0.00 0.00 0.00 Skewness -0.06 -0.04 0.06 Kurtosis 3.54 3.48 3.47 Jarque Bera 2.07 1.56 1.57 Probability JB 0.35 0.46 0.46
C.S. van den Bos University of Groningen 21 Table 4
Acquisition premium
The premium, the dependent variable, is based on the lognormal returns. The first column presents the total premium, whereas the premium in the second column is adjusted for the number of weeks between the announcement date and the completion date. The number of observations is N=4,986. Whole period premium Weekly premium Average 18.28% 1.17% Median 17.50% 1.00% Minimum -698.19% -10.17% Maximum 574.53% 12.42% Standard deviation 0.49 0.03 Skewness 38.41 0.07 Kurtosis -1.70 3.15 Jarque-Bera 1408.15 8.73 Probability JB 0.00 0.00
Most of the deals concern US targets (22%). Together with Japan, Canada and Great Britain, which account for 8% of the targets each, these four countries host almost half of the targets. However, the division of domicile countries of the acquirer is somewhat different, which is due to the number of cross-border deals (47.7%). A transaction is defined to be cross-border when the headquarters of the target and the bidder are located in different countries.
The four main sectors of industry in which almost two-thirds of the targets operate are consumer cyclical, consumer non-cyclical, financial and industrial. Of the acquirers, more than 40% are active in the financial sector, which can be explained by the large number of institutional buyers including private equity and venture capital that actively buy and sell companies. A more in-depth examination of the financial buyers reveals that private equity and investment companies represent 50% of the financial buyers. Private equity consists of both private equity and venture capital firms. The investment companies are divided into three subcategories based on the definition of the US Securities and Exchange Commission, the SEC: closed-end funds, mutual funds, and unit investment trusts.
4.2 Event study
C.S. van den Bos University of Groningen 22 Table 5 and graph 1 show that the three models result in AARs that are close to one another. Moreover, every model yields an AAR that is especially large at t=0, the event date, with an abnormal return between 8.37% and 8.42%. This AAR is in line with the AARs found by Dodd (1980), Keown and Pinkerton (1981) and Wansley et al. (1983), who found AArs of respectively 8.74%, 12.02% and 7.22% at the event date. The abnormal return at t=0 is highly significant according to all three models and both test statistics. This is in line with the hypothesis that the acquisition announcement has a positive impact on the target’s share price. In addition, the AARs of all three models are also significant at the 5%-level at t=1 according to both the parametric and the nonparametric test. The significance at t=1 could, for example, be due to a late reaction of some shareholders.
T=0 and t=1 are the only two days that yield significant AARs according to the models and both tests. This means that, although the positive values before t=0 indicate a run-up to the event, the real impact of the transaction announcement takes place on the event date and the day after the announcement, because these days show a significant abnormal return. We can conclude that the announcement of a public-to-private transaction has a positive impact on the target’s share price. Therefore we reject the null-hypothesis that states that a public-to-private transaction announcement does not influence the target’s share price.
Table 5
Significance of the AARs
The table presents the AARs of the Mean Adjusted Returns model, the Market Adjusted Returns model and the Market and Risk Adjusted Returns model and their corresponding p-values in the event window [-2; +2].
AAR P-value (t-test) P-value (Wilcoxon) AAR P-value (t-test) P-value (Wilcoxon) AAR P-value (t-test) P-value (Wilcoxon) -2 0.37% 0.211 0.546 0.31% 0.116 0.131 0.39% 0.232 0.621 -1 0.96% 0.126 0.070 0.93% 0.162 0.122 0.96% 0.136 0.111 0 8.47% 0.000 * 0.000 * 8.42% 0.000 * 0.000 * 8.47% 0.000 * 0.000 * 1 2.76% 0.024* 0.022 * 2.70% 0.030 * 0.025 * 2.67% 0.020 * 0.020 * 2 0.31% 0.459 0.856 0.22% 0.509 0.461 0.28% 0.497 0.147
Day Mean Adjusted Returns model
Market Adjusted Returns model
Market and Risk Adjusted Returns model
C.S. van den Bos University of Groningen 23 Graph 1: AARs in the event window. The three models show a comparable reaction to the transaction announcement, with a peak at t=0.
Table 6 and graph 2 present the CAARs for the event window and the corresponding p-values for both the parametric t-test and the nonparametric Wilcoxon signed rank test. Every period presented in table 4 yields CAARs that are highly significant according to all three models and to both tests. We would like to emphasize the size of the CAAR in period [0; +1]. Although this period exists of only 2 days the CAAR is almost as large as the CAAR in the other, larger periods. This is in harmony with our hypothesis stating that the announcement of a public-to-private transaction causes the target’s share price to increase.
Table 6
Significance of the CAARs
The table presents per model the CAARs in the event window and the parametric and nonparametric test results.
CAAR P-value (t-test) P-value (wilcoxon) CAAR P-value (t-test) P-value (Wilcoxon) CAAR P-value (t-test) P-value (Wilcoxon) [-2; +2] 12.87% 0.000** 0.000** 12.57% 0.000** 0.000** 12.77% 0.000** 0.000** [-1; +1] 12.18% 0.000** 0.000** 12.05% 0.000** 0.000** 12.10% 0.000** 0.000** [0; +1] 11.23% 0.000** 0.000** 11.12% 0.000** 0.000** 11.14% 0.000** 0.000** Period
Mean Adjusted Returns model
Market Adjusted Returns model
Market and Risk Adjusted Returns model
Note: |**| significant at the 1%-level. 0.00% 1.00% 2.00% 3.00% 4.00% 5.00% 6.00% 7.00% 8.00% 9.00% -2 -1 0 1 2 A A R Event window
Average Abnormal Return in event window
C.S. van den Bos University of Groningen 24
Graph 2: CAARs in the event window. The three models show a comparable reaction to the transaction announcement.
The CAARs of the entire event window (12.57%, 12.77%, and 12.87%) account for circa 70% of the average premium that we found in this study (18.28%). This means that as soon as the upcoming acquisition, together with its offer price, is out in the open, the target’s share price will approach the bid price. The reason why the share price only rises to 70% of the bid price during the event window is not analysed in this study, but could be due to uncertainty about the bidder’s ability to complete the acquisition. The data selection procedure in chapter 2 already stated that some offers are withdrawn or failed, which indicates the risk target shareholders face when an acquisition is announced.
These findings enable us to state that the announcement of an upcoming acquisition has a significant influence on the target’s share price. After the transaction is announced, the abnormal returns approach the final premiums that are paid for 70%.
4.3 Sensitivity analyses of the premium
Several sensitivity analyses are performed in order to examine subgroups of the data in detail. The differences between groups are tested with a Wilcoxon signed rank test, because the data is likely to be skewed due to the smaller amount of events per group.
We hypothesized that private equity buyers pay a lower premium than strategic buyers, because they have an information advantage and no synergy opportunities they are willing to pay for. Private equity buyers pay a median premium over the acquisition period of 15.47% and strategic bidders pay a median premium of 17.74%. This difference is
0.00% 2.00% 4.00% 6.00% 8.00% 10.00% 12.00% 14.00% -2 -1 0 1 2 C A A R Event window
Cumulative Average Abnormal Return in event window
Mean Adjusted Returns model
Market Adjusted Returns model
C.S. van den Bos University of Groningen 25 significant at a 5%-level, since it has a p-value of 0.036. Therefore it supports the hypothesis stating that private equity firms pay lower premiums than strategic bidders. Furthermore, this result corresponds to the findings of for example Bargeron et al. (2008), Roosenboom, Fidrmuc and Teunissen (2009) and Mnejja and Sahut (2010). Moreover, after correcting for the number of weeks between the announcement and the completion date, the analysis yields a similar result. Private equity firms pay a median weekly premium of 0.99%, which is significantly smaller than the 1.19% weekly premium paid for by the non-private equity companies (p-value is 0.011). Since both sensitivity tests result in a significant lower premium paid by private equity than the premium paid by strategic buyers, we can reject the null-hypothesis that private equity buyers pay the same premium as strategic buyers. This finding is in favour of the information asymmetry theory.
Consortiums of acquirers pay a medium weekly premium of 0.72% instead of the 1.01% paid by individual buyers. The corresponding p-value is 0.038, which implies that the difference is significant. Therefore, there is enough evidence to reject the null hypothesis, that consortiums pay lower premiums than individual buyers. This is in harmony with the information asymmetry theory.
Furthermore, we divide the strategic buyers in two groups; the inside industry buyers and the outside industry buyers. Inside industry buyers pay a median weekly premium of 1.17%, which is lower than the 1.23% premium for outside industry acquisitions. The corresponding p-value is 0.048, which means that the difference is significant at a 5%-level. Therefore we can reject the null hypothesis, which means that industry insiders pay lower premiums than industry outsiders. This is in accordance with the asymmetric information theory.
Foreign bidders pay a lower weekly premium than bidders that are located in the same country as the target; 0.97% compared to 1.01%. However, this difference has a p-value of 0.450 and is therefore not significant. As a consequence, we cannot reject the null-hypothesis that domestic bidders pay a lower premium than foreign acquirers. This result does not support the expectations based on the information asymmetry theory, where industry insiders are expected to have more information about the target and therefore offer a lower premium.
C.S. van den Bos University of Groningen 26 2009 (18.28%) is much lower than the 37.9% premium they found in the period 1990 to 1998. However, when we make two subsamples in our database and compare the weekly premium found in 2005 to 2006 (0.97%) with the premium of 2008 to 2009 (1.09%) we observe a higher premium for the second period. This difference is not significant (p=0.363). Therefore we do not find support for the statement of Jarrell and Poulsen (1989), Andrade, Mitchell and Stafford (2002), and Gondhalekar, Sant and Ferris (2004), that the premium decreases over the years. The reason could be due to the recent financial crisis: share prices worldwide collapsed after the start of the crisis. The drop in share prices is likely to increase the percentage premiums, although the premiums expressed in currencies might have dropped significantly.
Furthermore, a sensitivity analysis focused on the country where the bidder is headquartered is performed. The analysis shows that European based acquirers pay a weekly premium of 0.85% which is significantly less than their American counterparts that pay a weekly premium of 1.30% (p=0.000). The difference could be due to legal environment, since this might influence the ability of the acquirer to implement reorganisations; labour law in the European Union generally provides more employee protection than in the United States. Therefore it could be more costly to perform reorganisations in Europe than in the United States, which results in smaller opportunities, hence a smaller premium.
An overview of the above discussed sensitivity analyses and the corresponding p-values are presented in table 7.
Table 7 Sensitivity analysis
The second column contains the median weekly premiums paid by the first group of acquirers, whereas the fourth column contains the median weekly premiums of the second group of bidders.
Group 1 Median Group 2 Median p-value
(Wilcoxon)
Private equity 0.99% Strategic bidders 1.19% 0.011 *
Consortium 0.72% Individual bidders 1.01% 0.038 *
Industry insiders 1.17% Industry outsiders 1.23% 0.048 *
Cross-country deals 0.97% Domestic deals 1.01% 0.450
EU based bidders 0.85% US based bidders 1.30% 0.000 **
2005-2006 0.97% 2008-2009 1.09% 0.363
C.S. van den Bos University of Groningen 27 5. REGRESSION ANALYSIS
So far we studied the influence of a takeover announcement on the target’s share price and we performed several sensitivity analyses of the acquisition premium. In this section we will study the determinants of the acquisition premium by performing a regression analysis. We will start by presenting the methodology and the variables that are included in the regression. Thereafter, we will present the results of the regression analysis and the robustness checks.
5.1. Regression methodology
In addition to the above discussed variables that are related to the information asymmetry theory, we also put some control variables in the regression: the period length, the announced total value, a crisis dummy1, a United States dummy, a Europe dummy, a financial industry dummy, a consumer industry dummy and the country’s central bank lending rate2. An overview of all regression variables is presented in table 12, appendix A. In order to examine the determinants of the acquisition premium, we use an Ordinary Least Squares (OLS) regression analysis. Formula [14] represents the general OLS regression, where the dependent variable yi is the acquisition premium paid for target i, F is a constant
and u is the unobserved error term. The independent variables are represented by xj and their corresponding coefficients are formulated as betaj (βj). The independent variables are the different deal, target and acquirer characteristics on which the dependent variable is regressed with the OLS regression to obtain the betas. The error term reflects the portion of yi that cannot be explained by the independent variables and has a mean of 0.
G = F + & βIxKI+ uK M
N
[14]
The F-test [15] examines the extent to which the regression explains the variation in the dependent variable. The equation represents the ratio of the average regression sum of squares and the average sum of squared errors.
1 We assume the crisis to have started the day that Lehman Brothers Holding Inc. filed for bankruptcy, which is September 15th, 2008. This is in accordance with Cecchetti (2009).
C.S. van den Bos University of Groningen 28 O ="" 7 − (5 + 1)9⁄"" 5⁄
[15]
In this equation, RSS is the residual sum of squares, k the number of estimated betas, SSE the sum of squared errors and N the number of observations.
Furthermore, we will check the standard errors of the regression on heteroscedasticity, which means that the error terms across the observation points have different variances. A non-constant error term results in a non-minimum variance among the class of unbiased estimators. This could lead to incorrectly accepting the null-hypothesis, a so called type 1 error. In order to determine whether the error terms are heteroscedastic or not, a White’s general test for heteroscedasticity is performed. The White’s test for heteroscedasticity results in an F-statistic of 0.741 and a probability of 0.985. This means that the variance of the errors is homoscedastic or constant. Therefore the variances of the coefficients are not biased and the chance of a type 1 error is reduced.
We will also test the independent variables for multicollinearity, which means that two or more independent variables are highly correlated. The correlation coefficients are presented in appendix B. The correlation analysis indicates a high correlation between the EU dummy and the US dummy. Therefore we will delete the EU dummy from the regression. This dummy was only entered as a control variable and is most likely to cause a multicollinearity problem within the regression. In addition, we observe a high correlation between the financial dummy and the private equity dummy, which is caused by the fact that the group of financial buyers exists for one third of private equity. Since the financial dummy is just a control variable and the probability of multicollinearity problems increases by inlcuding both variables in the regression, we also delete the financial dummy from the regression. Although some other dummies are highly correlated we do not exclude other variables, because we believe their correlation is not causing multicollinearity problems. 5.2 Regression results
C.S. van den Bos University of Groningen 29 variables, except for the consortium dummy. According to the standard OLS regression a consortium has a significant negative effect on the premium, whereas the OLS regression with White correction states that this relation is not significant. Therefore we cannot reject the null-hypothesis that consortia pay the same premium as individual buyers.
As hypothesized, the percentage of target’s shares the acquirer already owns prior to the announcement of the deal is negatively related to the premium. The more shares the buyer already owns, the better access the buyer has to the management and the lower the acquisition premium. This finding is for both tests significant at a 1%-level. However, the coefficient is very small which means that the economic effect is not very large either. Subsequently, the private equity dummy has according to both test statistics a significant negative impact on the acquisition premium, which was also hypothesized. According to the regression analysis, the weekly premium decreases by 0.3% if the buyer is a private equity firm.
Furthermore, the cross-border dummy has, as hypothesized, a significant positive effect on the premium. The premium paid by a buyer that is headquartered in a different country than its target increases with 0.2%.
C.S. van den Bos University of Groningen 30 Table 8
Regression analysis, weekly premium
The table presents the results of the regression analysis with the weekly premium as the dependent variable and the factors presented below as the independent variables.
Variable Coefficient
Constant 0.020 9.478 ** 7.690 **
Percentage owned before 0.000 -2.822 ** -3.167 **
Private equity dummy -0.003 -2.302 * -2.142 *
Consortium dummy -0.006 -2.161 * -1.704
IndustryInsider dummy -0.001 -0.674 -0.721
Cross-border dummy 0.002 2.018 * 2.104 *
Hostile bid dummy 0.002 0.417 1.096
US based bidder dummy 0.001 0.766 0.675
Consumer industry dummy 0.002 1.319 1.527
Announced total value 0.000 -0.639 -1.377
100% ownership dummy 0.000 0.184 0.184
Crisis dummy -0.005 -3.304 ** -2.859 **
Interest rate 0.000 0.796 0.881
Period length 0.000 -4.006 ** -4.470 **
Number of observations after adjustmentsa 3771
R-squared 0.017 Adjusted R-squared 0.014 F-statistic 5.045 F-statistic probability 0.000 T-statistic (White correction) T-statistic
Note: |*| significant at 5%-level, |**| significant at 1%-level. a: E-views adjusts the sample number for lagged values.
C.S. van den Bos University of Groningen 31 Table 9
Country sensitivity
The table presents the regression results for the subsample of US and EU based buyers. The White variance-covariance matrix of the coefficients is used for both subsamples.
Variable Coefficient T-statistic Coefficient T-statistic
Constant 0.026 2.973 ** 0.023 5.129 **
Percentage owned before 0.000 -0.944 0.000 -3.877 **
Private equity dummy -0.013 -0.761 0.001 0.152
Consortium dummy -0.003 -1.258 -0.003 -1.948
IndustryInsider dummy -0.001 -0.459 0.001 0.740
Cross-border dummy 0.001 0.329 0.000 -0.230
Hostile bid dummy -0.003 -0.756 -0.003 -1.768
Consumer industry dummy 0.009 3.112 ** 0.001 0.627
Announced total value 0.000 -0.770 0.000 -2.832 **
100% ownership dummy 0.003 0.473 0.002 1.477
Crisis dummy -0.009 -2.071 * -0.005 -1.937
Interest rate 0.000 -0.160 0.000 0.562
Period length 0.000 -1.161 0.000 -5.433 **
Number of observations after adjustmentsa 960 1072
R-squared 0.035 0.065
Adjusted R-squared 0.023 0.055
F-statistic 2.898 6.146
F-statistic probability 0.001 0.000
US EU
Note: |*| significant at 5%-level, |**| significant at 1%-level. a: E-views adjusts the sample number for lagged values.
Table 9 shows the regression results for those transactions where the bidder is based in the US or in the EU. A comparison between the two regions shows some interesting differences. The premium paid by European Union based bidders is significantly influenced by (i) the percentage of target’s shares owned by the bidder prior to the transaction announcement, (ii) the announced total value of the transaction and (iii) the length of the period between the announcement date and the completion date. However, the economic effects are not very large, since the coefficient of all three variables is 0.000. The premium paid by US based acquirers, on the contrary, is significantly influenced by two variables: (i) the consumer industry dummy and (ii) the crisis dummy. If the bidder is active in the consumer industry the premium increases with 0.9%, whereas the premium decreases with 0.9% if the transaction occurred before the bankruptcy filing of Lehman Brothers on September 15th, 2008.
C.S. van den Bos University of Groningen 32 Table 10 shows the sensitivity checks for the bidder type; strategic bidders versus private equity. The premium paid by the strategic buyers is significantly influenced at the 1%-level by (i) the percentage of shares the buyer already owned, (ii) the crisis dummy, and (iii) the length of the period between the announcement date and the completion date. The regression variables explain 2.5% of the premium paid by the strategic buyers. On the contrary, the premium paid by private equity is hardly influenced by the variables included in this regression analysis. The adjusted R2 states that only 0.2% of the premium is explained by these variables. Moreover, none of the variables has a significant influence on the acquisition premium paid by private equity. This means that the premium paid by private equity buyers is hardly influenced by the analysed variables.
Table 10
Bidder type sensitivity
The table presents the regression results for the subsample of strategic buyers and private equity. The White variance-covariance matrix of the coefficients is used for both subsamples. |-| indicates that the variable is not included in the regression analysis of the subsample.
Variable Coefficient T-statistic Coefficient T-statistic
Constant 0.023 0.000 ** 0.017 1.424
Percentage owned before 0.000 0.000 ** 0.000 0.518
Consortium dummy -0.017 0.219 -0.018 -1.760
Industry insider dummy -0.003 0.062 -
-Cross-border dummy 0.002 0.112 0.002 0.440
Hostile bid dummy 0.002 0.306 0.003 0.581
US dummy 0.002 0.300 0.001 0.196
Consumer dummy 0.001 0.329 -
-Announced total value 0.000 0.163 0.000 0.341
100% ownership dummy 0.001 0.746 0.009 1.621
Crisis dummy -0.006 0.003 ** -0.007 -0.850
Interest rate 0.000 0.422 0.001 1.922
Period length 0.000 0.000 ** 0.000 -1.602
Number of observations after adjustmentsa 2992 363
R-squared 0.030 0.035
Adjusted R-squared 0.025 0.002
F-statistic 6.048 1.065
F-statistic probability 0.000 0.389
Strategic buyer Private equity
C.S. van den Bos University of Groningen 33 6. CONCLUSION
6.1 General Conclusions
This study analyses 4,986 acquisition premiums that have been paid worldwide in the period January 2005 up to December 2009. The median premium the bidders paid is 17.50%, which is lower than the premiums presented in previous papers. The weekly premium, which is the premium adjusted to the number of weeks over which the premium is calculated, is 1.00%.
An event study that uses the Mean Adjusted Returns model, the Market Adjusted Returns model and the Market and Risk Adjusted Returns model, shows that the average abnormal return on the announcement date of the transaction ranges from 8.37% to 8.42%. During the five days surrounding the event date [-2; +2], the three models yield a cumulative average abnormal return that lies between 12.57% and 12.87%. Both the abnormal returns and the cumulative abnormal returns are significant and therefore we can conclude that the target’s share price is significantly influenced by the announcement of an upcoming transaction. However, when comparing the (cumulative) abnormal returns to previous papers that used the same calculations, we must conclude that the influence on the share price is less in the period 2005-2009 than before.
C.S. van den Bos University of Groningen 34 In addition to the sensitivity analyses related to the information hypothesis, we also find a significant lower premium for bidders based in the European Union (0.85%) than those based in the United States (1.30%).
In addition, the regression analysis confirms almost all expected relations. The percentage of target shares that the acquirer already owns before the transaction is negatively correlated to the acquisition premium. Moreover, buyers that are classified as private equity or assembled in a consortium have a significant negative effect on the acquisition premium. Non-domestic buyers positively influence the premium significantly. However, although the industry insiders and the hostile bidders seem to be related to the acquisition premium in the hypothesized way, these results are not significant.
The regression analysis, although several significant independent variables are found, only explains 1.4% of the premium. The adjusted R-squared increases however, if the sample is divided into European Union en American based bidders.
6.2 Limitations of this study
This study, just like previous papers, assumes the final price paid per share to equal the share price at the completion date due to a perfect market. However, in reality the price per share actually paid might differ from the ultimate share price and would therefore result in a different premium. This could, for example, be caused by post-closing events triggered by warranties or unwarranted risks.
6.3 Recommendations for further research
The R-squared resulting from the regression analysis in this study is quite small. Therefore it would be interesting to analyse other factors that could influence the premium. Previous research (Roosenboom, Fidrmuc and Teunissen, 2009; Gondhalekar, Sant and Ferris, 2004; Dittmar, Li and Nain, 2008) states that the premium is not significantly influenced by the target’s characteristics. Therefore, a study of the bidder characteristics other than asymmetric information, the deal features or the environment characteristics could yield interesting findings and are recommended for further research.
C.S. van den Bos University of Groningen 35 parties recharge their losses into the premiums they are willing to pay or could they find other ways to ensure their profits?
C.S. van den Bos University of Groningen 36 REFERENCES
Alexandridis, G., D. Petmezas and N.G. Travlos, 2010, Gains from Mergers and Acquisitions Around the World: New Evidence, Financial Management 39, 1671-1695.
Andrade, G., M. Mitchell and E. Stafford, 2002, New Evidence and Perspectives on Mergers, Journal of Economic Perspectives 15, 103-120.
Ang, J. and N. Mauck, 2011, Fire Sale Acquisitions: Myth vs. Reality, Journal of Banking & Finance 35, 532-543.
Ayers, B.C., C.E. Lefanowicz and J.R. Robinson, 2003, Do Firms Purchase the Pooling Method?, Review of Accounting Studies 7, 5-32.
Barclay M.J. and J.B. Warner, 1993, Stealth Trading and Volatility: Which Trades Move Prices?, Journal of Financial Economics 34, 281-305.
Bargeron, L., F. Schlingemann, R. Stulz and C. Zutter, 2008, Why Do Private Equity Acquirers Pay so Little Compared to Public Acquirers?, Journal of Financial Economics, 89, 375-390. Berkovitch, E. and M., Narayanan, 1993, Motives for Takeovers: An Empirical Investigation, Journal of Financial and Quantitative Analysis 28, 347-362.
Betton, S., B.E. Eckbo and K. Thorburn, 2008, Corporate Takeovers, in B.E. Eckbo, ed.: Handbook of Corporate Finance: Empirical Corporate Finance, Vol. II, (Elsevier, Amsterdam). Boone, A.L. and J.H. Mulherin, 2007, How are Firms Sold?, Journal of Finance 62, 847-875. Boone, A.L. and J.H. Mulherin, 2008, Do Private Equity Consortiums Impede Takeover Competition?, Working paper, University of Kansas.
Bradley, M., A. Desai and E. Kim, 1988, Synergistic Gains From Corporate Acquisitions and Their Division Between the Stockholders of Target and Acquiring Firms, Journal of Financial Economics 21, 3-40.
Brown, S.J. and J.B. Warner, 1980, Measuring Security Price Performance, Journal of Financial Economics 8, 205-258.
Brown, S.J. and J.B. Warner, 1985, Using Daily Stock Returns, Journal of Financial Economics 14, 3-31.
Cataldo, A.J., 2003, Information Asymmetry: A Unifying Concept for Financial and Managerial Accounting Theories, (Emerald Group Publishing, Amsterdam).
Cecchetti, S., 2009, Crisis and Responses: The Federal Reserve in the Early Stages of the Financial Crisis, Journal of Economic Perspection 23, 51-75.
Chen, X., J. Harford and K. Li, 2007, Monitoring: Which Institutions Matter?, Journal of Financial Economics 86, 279-305.
Dasgupta, S. and K. Tsui, 2004, Auctions with Cross-shareholdings, Economic Theory 24, 163-194.
C.S. van den Bos University of Groningen 37 Dittmar, A., D. Li and A. Nain, 2008, The Bright Side of Bidder Competition, Working Paper, University of Michigan.
Dodd, P., 1980, Merger Proposals, Management Discretion and Stockholder Wealth, Journal of Financial Economics 8, 105-137,
Edmans, A., 2009, Blockholder Trading, Market Efficiency, and Managerial Myopia, The Journal of Finance 64, 2481-2513.
Fama, E., 1980, Agency Problems and the Theory of the Firm, Journal of Political Economy 88, 288-308.
Fleck, R., M. Lombers, L. Robertson and I. Brown, 2005, How to Structure Bids, 24 International Financial Law Review 37, 37-41.
Fishman, M.J., 1988, A Theory of Preemptive Takeover Bidding, The RAND Journal of Economics 19, 88-101.
Gondhalekar, V., R. Sant, and S. Ferris, 2004, The Price of Corporate Acquisition: Determinants of Takeover Premia, Applied Economics Letters 11, 735-739.
Hayward, M. and D. Hambrick, 1997, Explaining the Premiums Paid for Large Acquisitions: Evidence of CEO Hubris, Administrative Science Quarterly 42, 103-127.
Heflin, F. and K.W. Shaw, 2000, Blockholder Ownership and Market Liquidity, Journal of Financial and Quantitative Analysis 35, 621- 633.
Hendricks, K. and R.H. Porter, 1988, An Empirical Study of an Auction with Asymmetric Information, The American Economic Review 78, 865-883.
Jarrel, G.A. and A.B. Poulsen, 1989, The Returns to Acquiring Firms in Tender Offers: Evidence from Three Decades, Financial Management 18, 12-19.
Kang, J.K. and J.M. Kim, 2008, The Geography of Block Acquisitions, The Journal of Finance 63, 2817-2858.
Keown, A. and J. Pinkerton, 1981, Merger Announcements and Insider Trading Activity: An Emperical Investigation, The Journal of Finance 36, 855-869.
Lee. C.I., S. Rosenstein, N. Rangan and W.N. Davidson III, 1992, Board Composition and Shareholder Wealth: The Case of Management Buyouts, Financial Management 21, 58-72. Lenz, R., 2011, Get Rid of Banks and Build Up a Modern Financial World!, Working paper, University of applied sciences Bielefeld.
March, J.G. and Shapira, Z., 1987, Managerial Perspectives on Risk and Risk Taking, Management Science 33, 1404-1418.
Martin, K.J., 1996, The Method of Payment in Corporate Acquisitions, Investment Opportunities, and Management Ownership, The Journal of Finance 51, 1227-1246.
C.S. van den Bos University of Groningen 38 Moeller, S.B., F.P. Schlingemann and R.M. Stulz, 2007, How Do Diversity of Opinion and Information Asymmetry Affect Acquirer Returns?, The Review of Financial Studies 20, 2047-2078.
Officer, M.S., O. Ozbas and B.A. Sensoy, 2009, Club Deals in Leveraged Buyouts, Working Paper, University of Southern California.
Peng, M. and P. Heath, 1996, The Growth of the Firm in Planned Economies in Transition: Institutions, Organizations, and Strategic Choice, Academy Of Management Review 21, 429-528.
Pringle, C.D., M.J. Kroll, 1997, Why Trafalgar was Won before it was Fought: Lessons from Resource-Based Theory, Acadamy of Management Executive 11, 73-89.
Prôa Bressane, B. and M. Verdini Maia, 2010, CEO Overconfidence and the Impact on M&A Activity, Working Paper, University of Rio de Janeiro.
Ravid, S.A. and M. Spiegel, 1999, Toehold Strategies, Takeover Laws and Rival Bidders, Journal of Banking & Finance 23, 1219-1242.
Renneboog, L. and T. Simons, 2005, Public-to-Private Transactions: LBOs, MBOs, MBIs and IBOs, Discussion paper, Tilburg University.
Roll, R., 1986, The Hubris Hypothesis of Corporate Takeovers, Journal of Business 59, 197-216.
Rolvink, F., 2011, Limiet Renteaftrek Kost Bedrijven ‘Miljarden’, Financieel Dagblad, June 10.
Roosenboom, P., J. Fidrmuc and T. Teunissen, 2009, Under the Hammer: Do Private Equity Investors Really Acquire Target Firms for Less?, Working paper, Rotterdam School of Management.
Schnitzer, M., 1996, Hostile versus Friendly Takeovers, Economica 63, 37-55.
Schwert, G.W., 1996, Markup Pricing in Merger and Acquisition, The Journal of Financial Economics 41, 153-192.
Shleifer, A. and R. Vishny, 1992, Liquidation Values and Debt Capacity: A Market Equilibrium Approach, The Journal of Finance 47, 1343-1366.
Slusky, A. and R. Caves, 1991, Synergy, Agency, and the Determinants of Premia Paid in Mergers, Journal of Industrial Economics 39, 277-296.
Varaiya, N.P., 1986, An Empirical Investigation of the Bidding Firm’s Gains from Corporate Takeovers, Research in Finance 6, 149-178.
Varaiya, N.P. and K. Ferris, 1987, Overpaying in Corporate Acquisitions: The Winner’s Curse, Financial Analysts Journal 43, 64-69.
Walkling, R. and R. Edminster, 1985, Determinants of Tender Offer Premiums, Financial Analysts Journal 27, 30-37.
C.S. van den Bos University of Groningen 39 Wansley, J., W. Lane and H. Yang, 1983, Abnormal Returns to Acquired Firms by Type of Acquisition and Method of Payment, Financial Management 12, 16-22.
Wilson, R., 1967, Competitive Bidding with Asymmetric Information, Management Science 13, 816-820.
C.S. van den Bos University of Groningen 40 DATABASES
Bloomberg
MergerMarket (The Financial Times Group) OneSource (Infogroup)
C.S. van den Bos University of Groningen 41 LIST OF TABLES
Table 1 : Overview previous event studies Page: 9
Table 2 : Overview previous research on acquisition premiums Page: 10 Table 3 : Descriptive statistics of the return in the estimation window Page: 20
Table 4 : Acquisition premium Page: 21
Table 5 : Significance of the AARs Page: 22
Table 6 : Significance of the CAARs Page: 23
Table 7 : Sensitivity analysis Page: 26
Table 8 : Regression analysis, weekly premium Page: 30
Table 9 : Country sensitivity Page: 31
Table 10 : Bidder type sensitivity Page: 32
Table 11 : Regression variables Page: 42
Table 12 : Correlation matrix Page: 43
Graph 1 : AARs in the event window Page: 23
C.S. van den Bos University of Groningen 42 APPENDIX A – Regression variables
Table 11 Regression variables
The table defines the different regression variables including the control variables.
Variable Description
OwnedBefore The percentage of target shares that the acquirer already owns before the transaction.
PrivateEquity Dummy variable that is 1 if the acquirer is a private equity firm or a venture capitalist, 0 otherwise.
Consortium Dummy variable that is 1 if the acquiring party consists of multiple firms, 0 otherwise.
IndustryInsider Dummy variable that is 1 if the acquirer and the target are active in the same sector, 0 otherwise.
CrossBorder Dummy variable that is 1 if the acquirer and the target are not based in the same country, 0 otherwise.
HostileBid Dummy variable that is 1 if of a hostile bid, 0 otherwise.
US Dummy variable that is 1 if the acquirer is a US based company, 0 otherwise.
EU Dummy variable that is 1 if the acquirer is a EU based company, 0 otherwise.
Financial Dummy variable that is 1 if the acquirer is active in the financial sector, 0 otherwise.
Consumer Dummy variable that is 1 if the acquirer is active in the consumer (both cyclical and non-cyclical) sector, 0 otherwise.
AnnouncedValue The total amount in euros the acquirer is willing to pay on the announcement date.
CompleteOwnership Dummy variable that is 1 if the acquirer owns 100% of the target shares after completing the transaction, 0 otherwise.
Crisis Dummy variable that is 1 if the transaction occurs before the September 15th, 2008. Transactions that are both announced and completed after September 15th, 2008 have the value 0. Deals that are announced before this specific date, but were completed afterwards are considered as missing data.
InterestRate The interest rate set by the central bank of the acquirer's domestic country. PeriodLength The number of days between the transaction announcement and the deal
C.S. van den Bos University of Groningen 43
APPENDIX B – Correlation matrix
Table 12 Correlation matrix The table presents the correlation between the different factors of the regression analysis.
