Characteristics of value creating acquirers
Evidence from the EU
-MSc Business Administration: Finance Profile: Corporate Financial Management
August 2006 University of Groningen
Author Leo Ensing
Westkreek 7 8032 KM Zwolle 0384527191
StudentID. 1495348
ABSTRACT
In this paper I analyze the value creation of 200 acquirers situated in the EU. The ultimate goal is to distinguish firm and deal characteristics that positively influence the value creating possibilities of an acquisition.
Whether the acquisition created value for the acquirer has been measured with the short term event study methodology. This methodology focuses on the (abnormal) share price reactions around the acquisition announcement. This method is still the standard among researchers when examining announcement returns. I use the standardized cross sectional test of Boehmer et al. (1990) to examine if the share price reactions are significant. This test takes into account the increased variance of returns around the acquisition announcement.
TABLE OF CONTENTS
1.
INTRODUCTION... 4
2.
LITERATURE REVIEW... 7
2.1
Empirical results on value creation by acquirers ... 7
2.2
Explanatory firm and deal characteristics ... 9
3.
DATA AND METHODOLOGY ... 12
3.1
Sample selection and descriptive statistics... 12
3.2
Short term event study... 14
3.3
Statistical significance of the observed CAR ... 16
4.
RESULTS AND INTERPRETATION... 18
4.1
Negative CAR for whole sample ... 18
4.2
Correlations between firm and deal characteristics and the CAR... 20
4.3
Value creation for several subsamples ... 22
4.4
The effect of blockholders... 25
4.4
Rumour effect... 26
4.6
Cross sectional regression ... 28
1.
INTRODUCTION
The importance of Mergers and Acquisitions (M&As) has grown steadily over the last thirty years. The money involved, the jobs at stake, the reputation of the firm and its managers as well as the potential social impact all increased over the last decades. M&As often occur in waves mostly because firms react on the actions of their main competitors (Sudarsanam, 2003). The United States (US) already experienced five of such M&A waves and each wave was accompanied with higher deal values and an increased number of transactions (Sudarsanam, 2003). In Europe there have been only two waves so far. One between 1987 and 1992 and the last one during the late 1990s. The value of the M&A transactions in the first wave rose to $148bn in 1991 (Sudarsanam, 2003). The second wave followed the development in the US where the share prices of Internet and technology stocks skyrocketed. Investors believed that with the Internet the world economy had moved to another dimension with enormous growth potential and less risk. A lot of firms wanted to take part in this Internet hype and started to buy promising Internet companies. This resulted in increased M&As activities during this period. It peaked in 1999-2000 when the total value of acquisitions in the EU was respectively $1,129bn and $899bn and the number of acquisitions totalled 9301 in 1999 and 10405 in 2000 (Sudarsanam, 2003).
With hindsight it is easy to conclude that investors where overoptimistic about the possibilities of the Internet. This is best illustrated by a study of Cooper et al. (2001). The authors found that firms which only changed their company names into a name with .com in it, experienced abnormal increases in their share prices of 53% regardless of the actual involvement of the firm with the Internet. After the Internet bubble collapsed and the world economy slowed down the number of acquisitions in Europe and the US declined. Nowadays, there is a renewed interest in M&As. The worldwide number of acquisitions in 2004 was 21,620 with a total value of about $1,500bn (Cools and van de Laar, 2006). There are many reasons why a firm may decide to undertake an acquisition. Gaining market share, improving profitability and brand reputation, realizing economies of scale, gaining access to new technologies and exploring new markets are just some goals a firm may want to achieve with an acquisition. However, in the end every acquisition has to create shareholder value for the acquiring firm. Whether acquisitions create shareholder value for the acquiring firms is an important question without a direct answer.
On average the acquiring shareholders earn no abnormal returns. Why do so many firms enter into the acquisition process when the empirical results indicate that the acquiring shareholders earn on average zero abnormal returns?
The literature has come up with several explanations for this phenomenon. First of all, the average abnormal return (AR) is zero, but there are large variations in the returns for the acquiring firms indicating that there are firms that do create shareholder value. Many firms expect to beat the average. This may be warranted when the firm has undertaken some successful acquisitions in the past and has thus gained knowledge on how to create value from an acquisition. The learning effect for acquisitive firms is confirmed in empirical research where the total return for shareholders is equal or higher for frequent bidders (Fuller et al., 2002, Cools and van de Laar, 2006). A second reason that might explain the eagerness of firms to undertake acquisitions is related to the agency theory of Jensen (1986). Empire building management with a lot of free cash flow at hand has the incentive and the opportunity to undertake acquisitions even when these investment opportunities have a negative net present value. Moeller et al. (2004) and Harford (1999) found evidence that firms with large amounts of excess cash have poorer announcement returns. Finally, the cyclicality of M&As is also one of the reasons of the large amount of acquisitions undertaken. For example, a passive firm with the objective to keep its market share and competitive position at the same level has to react on acquisitions undertaken by its competitors even when this may result in value destruction in the short run.
2.
LITERATURE REVIEW
This part of the paper provides an overview of prior research done on the value creation of acquirers. Section 2.1 presents the empirical results on the value creation by acquirers. Why some firm and deal characteristics may explain part of the observed value creation will be explained in section 2.2.
2.1 Empirical results on value creation by acquirers
Table 1
Empirical results on the value creation by the acquiring firm
Study
Sample
region Sample period
Sample size
Estimation
period (days) Benchmark 1
Event window
(days) 2 CAR (%)
Georgen, M. and EU 1993 - 2000 142 270 MM (All-share index) -1, 0 0,7***
Renneboog, L. (2004) -2, +2 1,2***
-60, +60 -0,5
Campa, J.M. and EU 1998 - 2000 262 150 MM (domestic stock market) -90, -1 2,6
Hernando, I. (2004) -30, +1 1,4
-1, +1 0,4
-30, +30 0,6
Mulherin, J.H. and US 1990 - 1999 281 - Constant return model -1, +1 -0,4
Boone, A.L. (2000)
Sudarsanam, S. and UK 1983 - 1995 519 250 Mean adjusted model -1, +1 -1,4***
Mahate, A.A. (2003) Market adjusted model
Size adjusted model
M-to-B value adjusted model
Sudarsanam, S. UK 1980 - 1990 429 250 MM -20, -1 0,6
Holl, P. and 0 -1,3***
Salami, A. (1996) -20, +40 -4***
Martynova, M. and EU 1993 - 2001 2419 250 MM (MSCI Europe) -1, +1 0,7***
Renneboog, L. (2006) -5,+5 0,8***
-60, +60 -2,8** Schwert, G.W. (1999) US 1975 - 1996 1286 253 MM (CRSP value weighted portfolio) -63, 126 -1 Jakobsen, J.B. and Denmark 1993 - 1997 138 250 MM (value-weighted market return) -1, +1 0,7***
Voetmann, T. (2003) -15, +15 0,4
*** Significantly different from zero at the 0.01 level ** Significantly different from zero at the 0.05 level The CARs (%) are rounded at one decimal
1
MM stands for market model
2
2.2 Explanatory firm and deal characteristics
The previous table showed that some studies found (significant) positive CARs for their sample of acquirers while other studies found (significant) negative CARs. In most studies the authors tried to distinguish firm and deal characteristics that may explain why some firms created value where others destroyed it. In this section I will discuss the most important characteristics which are relevant for my research. The characteristics which are relevant in a simplified world where management and ownership are in the same hands is discussed first. After that the additional firm and deal characteristics that become important when taking into account the separation of management and ownership and the resulting information asymmetry will be presented. Finally, I will discuss how the existence of blockholders in the acquiring firm can influence the value creation.
A profit seeking firm in which management and ownership are in the same hands will undertake projects with the highest net present value. In this case the acquisition of a firm in the same industry may on average have better value creating opportunities compared to acquisitions in unrelated industries. There are for example more opportunities to increase the revenues (i.e. higher market share, network externalities) and reduce the costs (i.e. economies of scale) in related acquisitions. Moreover, there is empirical evidence that firms following a diversifying strategy trade at a discount (Lang and Stulz, 1994, Berger and Ofek, 1995, Lins and Servaes, 1999). This has become known as the conglomerate discount. To investigate whether related acquisitions generate higher abnormal returns for acquiring firms the Standardized Industrial Classification (SIC) code is used. Firms with the same four digit SIC code are considered to be related. This proxy for relatedness has been used by several authors (Fuller et al. 2002).
Apart from the relatedness of the potential target it seems easier to create shareholder value with the acquisition of non-listed targets compared to the acquisition of listed targets for three reasons. First of all, listed firms have the obligation to publish their balance sheets and profit and loss accounts at least once a year. Their value is therefore much easier to assess and it is thus more difficult to take over the target at a bargain. Secondly, it is more likely that there will be more interested parties to take over a listed target. Multiple bidders drive up the price and hence limit the returns (Moeller et al. 2004). Thirdly, Fuller et al. (2002) indicate that private owners are often more willing to cash out resulting in a better negotiation position for the bidder when acquiring a non-listed company.
Renneboog (2004). Thus it is unclear whether domestic or cross border acquisitions generate higher returns. The EU market is still rather fragmented especially compared to the US and it will be interesting to analyze which argument dominates.
Summarizing, it has been argued that in a simplified world where management and ownership are in the same hands the related acquisitions in the domestic market of private firms with a single bidder offer the best opportunities for the acquirer.
Contrary to the simplified case presented above the real world is characterized by separation of management and ownership. This results in an information advantage for the management. They have private information about the true value of their firm and will try to exploit this advantage by choosing among different financing possibilities.
Table 2
Method of payment and financing of the acquisition
Cash 1) Excess cash
2) Issue new debt 3) Issue new equity Shares of the bidder Issue new equity
Target stockholders receive (method of payment)
Financing of the acquisition by the acquirer
As can be seen in the above table the method of payment for the bidder is either cash or shares of their own company. A cash offer by the bidder can be financed in three different ways. First of all, the bidder can use its excess cash balances. Secondly, they can issue new debt and pay out the proceeds as cash to the target. Thirdly, the bidder can issue new equity and pay out the proceeds as cash to the target. When the bidder and target arranged that the method of payment is equity the bidder can only issue new equity to finance the deal. The management of the bidding firm can use its information advantage and issue new equity when their firm is overvalued in the stock market, while it is more likely that they will use their excess cash balances when the firm is undervalued. The shareholders of the bidding firms are ex ante aware of this information asymmetry and will ceteris paribus react negatively on an acquisition that is financed with an equity issue compared to deals that are financed with excess cash (Georgen and Renneboog, 2004, Chemmanur and Paeglis, 2003, Martin, 1996). This has become known as the signalling function of equity.
leverage on the announcement returns of bidder firms. Their findings confirm that debt result in better managerial decisions because bidding firms with higher leverage earned higher abnormal returns. The size of the acquirer also becomes relevant in explaining the CARs observed in the event window. It turns out that large firms make large acquisitions with low or negative abnormal returns while small firms make many successful acquisitions but with rather small deal values. Moeller et al. (2004) found strong evidence for this ‘size effect’. They provide several explanations which are more likely to be present for large firms because for large firms the interests between the management and the shareholders are less aligned and they also have more resources to use. The first reason proposed by Moeller et al. (2004) is that managers of large firms are too eager to buy a particular firm (hubris hypothesis). Second, according to Jensen (1986) empire building management is more likely to pursue negative net present value acquisitions than to increase the dividends (free cash flow hypothesis). They try to increase the amount of resources under their control even if the firm grows beyond the optimal size. Murphy (1999) adds that firm size and managers’ remuneration are also positively correlated giving even larger incentives for the management to expand the firm via acquisitions.
Finally, I hypothesize that the existence of large blockholders in the acquiring firm has a positive effect on the value creating possibilities. Most of the work done on the effect of blockholders is focused on the relationship between the premiums for the target firms and the presence of large shareholders in these targets (Moeller, 2005). Other research focused on the relationships between the value of a particular firm and the presence of blockholders (Thomsen et al. 2006, Maury and Pajuste, 2005). However, the relationship between blockholders in the acquiring firm and the value creating possibilities of the acquisition has not yet been well examined. Shleifer and Vishny (1997) discuss the effects of large shareholders in a general way. They argue that blockholders can influence the policies of the management and that they can block (value destroying) investment projects such as acquisitions. Furthermore, Shleifer and Vishny (1997) argue that the presence of large shareholders result in higher management turnover and less discretionary spending indicating that these blockholders improve the effectiveness of the firm. They also argue that at some point the costs of large investors (i.e. pursuing personal goals and expropriation of minority shareholders) may begin to outweigh the benefits indicating the existence of some kind of turning point.
The way the deal is financed, the amount of leverage, the size of the acquirer and the existence of
large blockholders all become relevant when adding the separation of management and ownership and
3.
DATA AND METHODOLOGY
In this part of the paper the data and the event study methodology is presented. In section 3.1 the sample selection criteria as well as some descriptive statistics of the sample are presented. How the event study is conducted is outlined in section 3.2. That section also presents the standardized cross sectional test of Boehmer et al. (1990) which will be used to determine whether the observed CARs are significantly different from zero.
3.1 Sample selection and descriptive statistics
The database Zephyr has been used to construct a sample of acquisitions. To be included in the sample the acquisition announcements had to satisfy the requirements in table 3.
Table 3 Sample criteria
Category Specification
Time period 1/1/2000 - 31/12/2003
Geography Acquirer and target situated in EU
Deal value > € 1 million
Quoted companies Acquirer is listed
Deal type Acquisition
Current deal status Announced, completed
Acquiring a controlling stake 50 - 100%
Table 4
Descriptive statistics about the acquisition announcements included in the sample
Number Number
Total sample 200 Blockholders (<5%) 61
Blockholders (5% - 10%) 31
Mean market value acquirer (in mln euros) 6.737 Blockholders (10% - 15%) 35
Median market value acquirer (in mln euros) 614 Blockholders (15% - 20%) 22
Blockholders (20% - 50%) 40
Listed targets 102 Blockholders (>50%) 11
Non-listed targets 98
Austria 6
Related acquisition * 94 Belgium 6
Unrelated acquisition 106 Denmark 3
Finland 8
Domestic acquisitions 136 France 21
Cross border acquisitions 64 Germany 11
Greece 8
Year 2000 43 Italy 16
Year 2001 41 Netherlands 8
Year 2002 55 Portugal 3
Year 2003 61 Republic of Ireland 3
Spain 6
All cash offer 136 Sweden 9
All shares offer 55 United Kingdom 92
Acquisitions with prior rumours 50
Acquisitions without prior rumours 150
Source: database Zephyr
* An acquisition is considered to be related when the target and acquirer have the same 4 digit SIC code
3.2 Short term event study
The event study methodology is the standard among researchers since Fama et al. (1969) and has been used for this research. The acquisition created value is measured by analyzing the share price movements during a number of days surrounding the acquisition announcement. The number of days is called the event window. The focus is on the observed abnormal returns in this period where the abnormal return is defined as the difference between the actual return on the shares of the bidder and the normal return. The normal return is estimated on the period before the start of the event window. This period is called the estimation period. The basics of the event study are shown in figure 1.
Figure 1
Structure of the event study
t= -280 t= +30 estimation period post event period event window t= -30 t= 0
The rest of this section covers the technical and statistical aspects of the performed event study which are based on the work done by Fama et al. (1969), Brown and Warner (1985) and Mackinlay (1997). Furthermore, I will elaborate on the assumptions and choices made.
There are several important decisions to be made when performing an event study. First of all, the length of the event window has to be defined. Initially I will use an event window of 61 days, starting 30 days prior (t = -30) to the official announcement day and ending 30 days after (t = +30) the announcement. This event window is large enough to capture pre-announcement share price changes triggered by information leakage as well as post-announcement effects. Furthermore, it is in line with important previous studies mentioned in the literature review section (table 1). Depending on the obtained results I will analyze the significance of the CARs over shorter event windows.
Finally, the length of the estimation period is set at 250 trading days prior to the start of the event window (t = -280, t = -30). Such a period is long enough for reliable estimates of the market model parameters and it is also used by most authors as mentioned in the literature review section (table 1). To increase the confidence in the results obtained via the procedure above I will perform several sensitivities, ranging from adjustments to the beta to changing the length of the estimation period.
The market model is defined as follows.
it MSCIit i i it R R (1)
Where: Rit total return on stock i on day t
ai constant
ßi beta of stock i
Rmscit total return on MSCI country i on day t eit error term with expected value of zero
The total daily return on the individual stocks and the MSCI index is calculated by taking the natural logarithm of the stock price on a particular day divided by the stock price of the prior day. I used the stock prices corrected for dividend payments as provided by Datastream.
Ordinary least squares (OLS) regressions are used to estimate the a and ß over the 250 day estimation period for each security. The mean a and ß where 0,00 and 0,53 respectively. The sum of all the error terms (eit) in the estimation period was 0,000 which confirms that the OLS regressions where done correctly.
The AR on a particular day in the event window is the difference between the actual return and the normal return estimated with the market model. It is thus the error term of the market model.
MSCIit i i it it R R AR (2)
To make inferences about the value creation over the entire event window the AR of each day in the event window is aggregated to calculate the cumulative abnormal return.
The CAR for the whole sample or subsample is the mean of the CAR for the individual securities in the data sample.
N i it CAR N CAR 1 1 (4)
3.3 Statistical significance of the observed CAR
The previous section discussed the technical aspects of the event study and how the CARs are calculated. However, the ultimate goal is to determine whether the observed CAR is statistically significantly different from zero. The literature on event studies has developed multiple tests to determine the significance of the observed cumulative abnormal return. In essence all these tests relate the observed abnormal returns in the event window to the standard deviation of the returns to be able to conclude whether or not the CARs are significantly different from zero. However, the tests range from simple and easy to apply traditional tests to complex tests that take into account many statistically important characteristics. I will discuss three main issues that distinguish the different tests and are thus important for choosing between them. First, in the design of the test you have to make an assumption about the cross sectional (in)dependency of the residuals. I assumed that the security residuals are independent across the securities because the 200 securities in the data sample are situated in 14 different countries, are listed on different stock exchanges, operate in different markets and are not selected to have common characteristics other than being an acquisition, as can be seen in table 3. The traditional test of Brown and Warner (1980) is an example of a simple, easy to implement test which assumes cross sectional independency.
A second issue is whether the residuals are standardized. Brown and Warner (1985) and Boehmer et al. (1990) argue that standardization of the residuals has two main advantages. By standardizing the residuals you correct for the higher standard deviation that is applicable for the event window because you perform an out of sample prediction. Furthermore, it prevents the securities with large variances from dominating the sample. Brown and Warner (1985) use this test which is called the standardized residual method. This test also assumes cross sectional independency among the security residuals, thus the only difference with the traditional test is the standardization.
estimation period and the event window, thus this method takes into account the increased variance observed in the event window. His test is called the standardized cross sectional test and will be used for this research.
4.
RESULTS AND INTERPRETATION
In this section the results of the event study will be discussed. Section 4.1 presents information about the value creation by the acquirers. That section also provides a graphical presentation of the abnormal return on each day in the event window along with the corresponding significance levels. Section 4.2 provides correlations between firm and deal characteristics and the CAR to get a first impression of the relationships. Section 4.3 provides information on the value creation for several subsamples. The effect of blockholders is analyzed in section 4.4. In section 4.5 the rumour effect is discussed. A cross sectional regression analysis is performed in section 4.6 to provide further evidence on the proposed firm and deal characteristics. Several sensitivity analyses are presented in section 4.7.
4.1 Negative CAR for whole sample
The average (median) cumulative abnormal return for the whole sample was an insignificant -0,66% (-0,34%). Some descriptive statistics about the observed CARs are presented in table 5 with the t-test statistic between brackets.
Table 5
Descriptive statistics of the CAR during the 61 day event window
N 200
Mean
Median -0,34%
Maximum 77%
Minimum -139%
Source: own calculations
-0,66%
(-0,48)
following days there are mostly negative ARs which result in a negative but insignificant CAR of 0,66% over the entire event window.
Graph 1
The daily average abnormal return during the event window
-0,008 -0,006 -0,004 -0,002 0 0,002 0,004 0,006 0,008 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Days A b n o rm a l re tu rn
Source: own calculations
The CAR over the 61 day event window is presented in graph 2. That graph is constructed by adding all the abnormal returns on each day in the event window. The graph shows a pre-announcement price run up. The first 32 days (t = -30, t = +1) result in a significant (at the 0.10 level) CAR of 1,8%. Also the event window from t = -2 till t = +2 is highly significant with a cumulative abnormal return of 1,4%. Starting two days after the official announcement day the CAR gradually declines until an insignificant -0,66% at t = +30.
explain this finding. Whether firm and deal specific characteristics are to blame for the negative post-announcement drift in abnormal returns will be investigated in the rest of this paper.
Graph 2
The CAR for the whole sample during the 61 day event window
-0,01 -0,005 0 0,005 0,01 0,015 0,02 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 0 Days C A R
Source: own calculations
4.2 Correlations between firm and deal characteristics and the CAR
Table 7
Most important correlations Pearson
Correlations
cash financed acquisition
-0,136
(0,054*)
-0,149
(0,035**)
size of the acquirer (b)
-0,171
(0,016**)
Leverage (c)
0,159
(0,025**)
(b) The market capitalization at the day before the start of the event window is proxy for size (c) Leverage is the ratio of debt over debt plus equity
Source: own calculations
** Significant at the 0.05 level (2-tailed)
* Significant at the 0.10 level (2-tailed)
no block-holders
(<5%) non-listed domestic
4.3 Value creation for several subsamples
Although the whole sample generated on average a negative CAR, there are 97 acquisition announcements with a positive CAR. This section will explore whether there are firm and deal characteristics that contribute to a positive CAR. In the table below the sample is divided in several subsamples according to the reasoning presented in the literature review section. The t-test statistic is presented between brackets.
Table 8
Cumulative abnormal return during 61 day event window for different subsamples
N 136 55 102 98 94 106 64 136
Mean
Median 0,024 -0,025 0,025 -0,029 -0,003 -0,003 -0,002 -0,008
Maximum 0,717 0,020 0,772 0,717 0,555 0,772 0,536 0,772
Minimum -1,386 -0,049 -1,386 -0,697 -1,387 -0,697 -0,872 -1,386
** Significantly different from zero at the 0.05 level
* Significantly different from zero at the 0.10 level
Source: own calculations
Unrelated acquisition Cross border -0,027 (-1,42*) 0,009 (0,46) -0,008 (-0,37) -0,005 (-0,30) Domestic acquisition -0,023 (-1,31*) -0,004 (-0,17) 0,004 (0,26) -0,008 (-0,46) Cash financed Equity financed Listed target Non-listed target Related acquisition
Graph 3
CAR divided in equity and cash financed acquisitions
-0,04 -0,03 -0,02 -0,01 0 0,01 0,02 0,03 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 0 Days C u m u la ti v e a b n o rm a l re tu rn Whole sample Equity financed Cash financed
Source: own calculations
Graph 4
CAR divided in listed and non-listed targets
-0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 0 Days C u m u la ti v e a b n o rm a l re tu rn Whole sample Listed Non-listed
Source: own calculations
As can be seen in table 8 the subsamples that distinguished whether the acquisition is related / unrelated or cross border / domestic did not yield significant CARs. Moreover, the differences in mean and median CAR between related and unrelated acquisitions as well as between cross border and domestic acquisitions are small. The results for the related and unrelated acquisitions do not confirm my ex ante reasoning in the literature review section in which I argued that related acquisitions have on average better value creating opportunities compared to acquisitions in unfamiliar industries. A possible explanation is that although the related acquisitions have more opportunities to increase revenues and reduce costs as argued in section 2.2, the bidder has to pay a higher price for these low risk acquisitions. When examining the deal values this reasoning is confirmed because the average deal values for the related acquisitions are about 20% higher compared to the deal values for unrelated acquisitions. Thus, the opportunities may be higher for related acquisitions but when this is offset by a higher offer price the value creating opportunities are actually lower.
In section 2.2 I already mentioned the unclear effect of cross border and domestic acquisitions. Unfortunately the results in table 8 do not provide evidence in favour of either the cross border or the domestic acquisition.
4.4 The effect of blockholders
I also hypothesized that the existence of blockholders in the acquiring firm may positively influence the observed abnormal returns. The data on the dispersion of ownership is obtained via the database Amadeaus. The mean, median, maximum and minimum CARs for acquirers with (=5%) and without (<5%) significant blockholders is presented below.
Table 9
CAR for acquiring firms with (=5%) and without (<5%) significant blockholders
<5% 5% -10% 10% - 15% 15% - 20% 20% - 50% >50% Mean Median 0,006 -0,003 -0,04 0,039 -0,002 0,06 Maximum 0,772 0,404 0,717 0,341 0,536 0,352 Minimum -1,386 -0,46 -0,7 -0,46 -0,87 -0,264
*** Significantly different from zero at the 0.01 level
** Significantly different from zero at the 0.05 level
Source: database Amadeus
0,017 (1,68**) -0,042 (-0,41) 0,033 (3,61***) Blockholders -0,002 (-0,17) 0,001 (0,12) -0,008 (-0,99)
Graph 8
Abnormal returns on official announcement day for several blockholders categories
-0,01 -0,005 0 0,005 0,01 0,015 0,02 0,025 0,03 0 Announcement day (t= 0) A b n o rm a l re tu rn Whole sample no blockholders >5% - 10% >10% - 15% >15% - 20% >20% - 50% >50%
Source: own calculations
4.4 Rumour effect
Table 10 Rumour effect Mean Median 0,025 -0,023 Maximum 0,772 0,717 Minimum -1,386 -0,872
** Significant at the 0.05 level
Source: database Zephyr
Acquisition with rumour -0,017 (-1,1) 0,025 (1,73**) Acquisition without rumour
The surprising finding of above table is that the mean and median CAR for acquisitions with prior rumours is actually higher. As can be seen in the graph below this is a result of the post-announcement period in which the acquisitions with prior rumours have an upward trend while the acquisitions without these prior rumours show a negative drift.
Graph 9
CAR divided in acquisitions with and without prior rumours.
-0,02 -0,015 -0,01 -0,005 0 0,005 0,01 0,015 0,02 0,025 0,03 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 0 Days C u m u la ti v e a b n o rm a l re tu rn ru mours n o ru mou rs whole s ample
The acquisitions with prior rumours have, as hypothesized, lower CARs in the 30 days before the official announcement day. However, the ARs for the acquisitions with rumours is much higher after the official announcement resulting in a CAR of 2,5% over a 61 day event window. Together with the CAR of 1,18% at the time of the rumours this results in a CAR of 3,68% compared to a negative CAR of 1,7% for the acquisitions without rumours. Thus, acquisitions with rumours create more value not only in the days surrounding the official announcement but also in the days following the rumour.
4.6 Cross sectional regression
I also performed a cross sectional regression on the observed CAR with the firm and deal characteristics discussed previously. The results which are presented in table 11 are rather weak, with a low overall fit of the model (F statistic is 0,336) and no significant coefficient parameters. However, the sign of the parameters are consistent with the previous analyses. As before, the blockholders with share holdings between 15% and 20% as well as those above 50% have a positive effect on the CAR. The acquisitions with prior rumours as well as the cash financed acquisitions have a positive effect on the CAR and are thus consistent with the previous analyses. Furthermore, the analysis also confirms that the acquisition of non-listed targets does create more value (positive coefficient). The initially observed negative CAR is thus indeed a result of the fact that those deals are often equity financed.
Table 11
Results of cross sectional regression analysis
Parameter Significance
Intercept -0,061 0,371
Dummy = 1 for acquisitions with prior rumours 0,040 0,392
Dummy = 1 for cash financed acquisition 0,032 0,433
Leverage (debt / [debt + equity]) 0,045 0,495
Dummy = 1 for acquisition of non-listed target 0,020 0,619
Size of the acquirer (market capitalization) 0,000 0,651
Dummy = 1 for acquisition of related target -0,002 0,967
Dummy = 1 for a domestic acquisition -0,002 0,955
Dummy = 1 for blockholders 5% - 10% 0,006 0,923
Dummy = 1 for blockholders 10% - 15% 0,006 0,914
Dummy = 1 for blockholders 15% - 20% 0,018 0,781
Dummy = 1 for blockholders 20% - 50% -0,030 0,571
Dummy = 1 for blockholders >50% 0,040 0,635
N 200
F statistic 0,336
Source: own calculations
4.7 Sensitivity analyses
I performed several sensitivity analyses to check the robustness of the obtained results. I altered the assumptions and choices made in section 3.2. First of all, I used Blume’s (1971) technique to adjust the beta from the market model. Blume (1971) found that the betas tend to regress towards the mean of one. Thus, the estimates for the beta improve when the raw historical beta is adjusted for this mean reversion. Secondly, the length of the estimation period was extended to 375 days. The advantage is that more data is used to estimate the parameters. However, the downside is that recent changes in the firm’s characteristics are now given less weight. Finally, the market model is estimated on a 250 day estimation period ending 125 days prior to the start of the event window. Table 12 reports the CAR for the whole sample as well as for the subsamples under the altered assumptions. The differences between the analyses are small. Table 13 in Appendix III shows the significance of the different analyses with the significance level between brackets. None of the analyses showed significantly different results. Table 12 Sensitivity analyses Analyses A -0,007 0,004 -0,027 0,009 -0,023 -0,008 -0,005 -0,004 -0,008 B -0,001 0,010 -0,026 0,010 -0,014 -0,007 0,003 -0,004 0,000 C -0,008 -0,004 -0,010 0,008 -0,025 -0,007 -0,009 -0,004 -0,010 D -0,010 0,005 -0,046 0,001 -0,022 -0,012 -0,009 -0,006 -0,012 Analyses A -0,002 0,001 -0,008 0,017 -0,042 0,033 B 0,000 0,006 0,001 0,022 -0,032 0,026 C -0,002 -0,007 -0,025 -0,003 -0,023 0,051 D -0,005 0,002 0,011 0,032 -0,064 -0,029
A is the initial analyses
B uses the adjustment of Blume (1971)
C uses a 375 day estimation period ending at the start of the event window
D uses a 250 day estimation period ending 125 days before the start of the event window Source: own calculations
5.
CONCLUSION
The observation that many firms undertake acquisitions while the acquirers earn on average zero abnormal returns triggered this research. I examined the acquisition announcements during the period 2000-2003. The data sample was reduced to 200 acquisition announcements after the collection of the relevant firm and deal characteristics.
I used the short term event study methodology which is still well appreciated among researchers. By taking into account the event induced variance the research method should produce reliable results. The acquirers in the data sample experienced a negative CAR in the 61 day event window of 0,66%. Thus, on average the acquirers destroyed value although the results where insignificant. The CAR in the first 31 days in the event window showed an upward trend with a significant AR of 0,57% on the day of the announcement. However, the 30 days following the official announcement showed a negative drift of the CAR. This is consistent with the finding of Georgen and Renneboog (2004), Sudarsanam et al. (1996) and Martynova and Renneboog (2006) who all find a (significant) negative post-announcement drift while the pre-announcement period yielded (significant) positive CARs. In the rest of the paper I tried to distinguish firm and deal characteristics that positively influence the observed CAR. Some of the already well-known characteristics are confirmed by my research. The cash financed acquisition for example generates a significant positive CAR while the equity financed acquisition shows a significantly negative CAR. This confirms the equity signalling hypothesis and is consistent with the research of Georgen and Renneboog (2004), Chemmanur and Paeglis (2003) and Martin (1996). The non-listed targets seemed unexpectedly to under perform compared to the acquisition of listed targets. However, the observed negative AR in the post-announcement period turned out to be a result of the fact that these acquisitions are often financed with equity. Apart from examining the well-known firm and deal characteristics I focused on two relatively new concepts. First of all, I discovered that the existence of large blockholders in the acquiring firms have a positive influence on the observed CAR. Especially the firms with blockholders that have a direct ownership between 15% and 20% or with more than 50% benefit. These blockholders have the ability to perform tight monitoring actions and influence the firm’s policy. Secondly, the acquisitions that had rumours more than one month prior to the official announcement day experienced a significant positive cumulative abnormal return around the rumour day as well as around the official announcement day. This was unexpected because I hypothesized that the value creation around the official announcement day would be lower because some of the news of the acquisition was already incorporated in the share prices at the start of the rumours.
REFERENCES
Beaver, W.H. 1968. The Information Content of Annual Earnings Announcements. Journal of
Accounting Research, 6: 67-92.
Berger, P.G. and Ofek, E. 1995. Diversification’s effect on firm value. Journal of Financial
Economics, 37: 39-65.
Boehmer, E., Musumeci, J., and Poulsen, A.B. 1990. Event-study methodology under conditions of event-induced variance. Journal of Financial Economics, 30: 253-272.
Blume, M.E. 1971. On the assessment of risk. Journal of Finance, 26: 1-10.
Brown, S.J. and Warner, J.B. 1980. Measuring security price performance. Journal of Financial Economics, 8: 205-258.
Brown, S.J. and Warner, J.B. 1985. Using Daily Stock Returns: The Case of Event Studies. Journal of
Financial Economics, 14: 3-31.
Bruner, R.F. 2002. Does M&A Pay? A Survey of Evidence for the Decision-Maker. Journal of
Applied Finance, 12: 48-69.
Campa, J.M. and Hernando, I. 2004. Shareholder Value Creation in European M&As. European
Financial Management, 10: 47-81.
Chemmanur, T.J. and Paeglis, I. 2003. The Choice of the Medium of Exchange in Acquisitions: A Direct Test of the Double-Sided Asymmetric Information Hypothesis. Working paper.
Cools, K. and van de Laar, M.M. 2006. The Performance of Acquisitive Companies. Working paper.
Cooper, M.J., Dimitrov, O., and Rau, P.R. 2001. A Rose.com by Any Other Name. Journal of
Finance, 56: 2371-2388.
Faccio, M., McConnell, J.J. and Stolin, D. 2006. Returns to Acquirers of Listed and Unlisted Targets.
Journal of Financial and Quantitative analysis, 41: 197-220.
Fuller, K., Netter, J., and Stegemoller, M. 2002. “What Do Returns to Acquiring Firms Tell Us? Evidence from Firms That Make Many Acquisitions”. Journal of Finance, 57: 1763-1793.
Georgen, M. and Renneboog, L. 2004. Shareholder Wealth Effects of European Domestic and Cross-border Takeover Bids. European Financial Management, 10: 9-45.
Harford, J. 1999. Corporate Cash Reserves and Acquisitions. Journal of Finance, 54: 1969-1998.
Jakobsen, J.B. and Voetmann, T. 2003. Post-acquisition performance in the short and long run. Evidence from the Copenhagen Stock Exchange 1993-1997. The European Journal of Finance, 9: 323-342.
Jensen, M.C. 1986. Agency Costs of Free Cash Flow, Corporate Finance and Takeovers. American
Economic Review, 76: 323-330.
Kalay, A. and Loewenstein, U. 1985. Predictable events and excess returns; The Case of Dividend Announcements. Journal of Financial Economics, 14: 423-449.
Lang, L.H.P. and Stulz, R.M. 1994. Tobin’s q, Corporate Diversification and Firm Performance. The
Journal of Political Economy, 102: 1248-1280.
Lins, K. and Servaes, H. 1999. International Evidence on the Value of Corporate Diversification. The
Journal of Finance, 54: 2215-2239.
Mackinlay, A.C. 1997. Event Studies in Economics and Finance. Journal of Economic Literature, 35: 13-39.
Maloney, M.T., McCormick, R.E. and Mitchell, M.L. 1993. Managerial Decision Making and Capital Structure. Journal of Business, 66: 189-217.
Martin, K. 1996. The Method of Payment in Corporate Acquisitions, Investment Opportunities, and Management Ownership. Journal of Finance, 51: 1227-1246.
Maury, B and Pajuste, A. 2005. Multiple large shareholders and firm value. Journal of Banking &
Finance, 29: 1813-1834.
Moeller, T. 2005. Let’s make a deal! How shareholder control impacts merger payoffs. Journal of
Financial Economics, 76: 167-190.
Moeller, S.B., Schlingemann, F.P. and Stulz, R.M. 2004. Firm size and the gains from acquisitions.
Journal of Financial Economics, 73: 201-228.
Mulherin, J.H. and Boone, A.L. 2000. Comparing acquisitions and divestitures. Journal of Corporate
Finance, 6: 117-139.
Murphy, K. 1999. Executive compensation. Orley Ashenfelter and David Card (eds.), Handbook of Labor Economics 3, North Holland.
Schwert, G.W. 1999. Hostility in Takeovers: In the eyes of the beholder? Working paper.
Shleifer, A. and Vishny, R.W. 1997. A Survey of Corporate Governance. The Journal of Finance, 2: 737-783.
Sudarsanam, S. 2003. Creating Value from Mergers and Acquistions. Pearson Education Limited, Edinburgh.
Sudarsanam, S. and Mahate, A.A. 2003. Glamour Acquirers, Method of Payment and Post-acquisition Performance: The UK Evidence. Journal of Business Finance and Accounting, 30: 299-341.
Sudarsanam, S., Holl, P. and Salami, A. 1996. Shareholder wealth gains in mergers: Effect of synergy and ownership structure. Journal of Business Finance and Accounting, 23: 673-698.
APPENDIX I
Table 6
Correlation table between the firm and deal characteristics and the observed CAR
Pearson Correlations C A R 1 0,064 (0,372) -0,005 (0,944) -0,007 (0,927) 0,061 (0,389) -0,028 (0,695) 0,012 (0,866) 0,013 (0,854) -0,003 (0,969) 0,033 (0,644) -0,071 (0,321) 0,038 (0,590) 0,073 (0,301) 0,062 (0,382) non-listed target 1 0,101 (0,153) -0,008 (0,914) -0,136 (0,054*) 0,101 (0,153) 0,085 (0,234) -0,050 (0,482) -0,049 (0,493) 0,057 (0,424) -0,060 (0,399) 0,017 (0,810) 0,404 (0,000***) 0,065 (0,360)
related acquisition (a) 1
-0,041 (0,562) -0,063 (0,378) 0,050 (0,478) -0,015 (0,838) 0,012 (0,867) -0,065 (0,363) 0,053 (0,455) 0,030 (0,673) -0,007 (0,916) 0,012 (0,871) -0,011 (0,882) domestic acquisition 1 -0,149 (0,035**) -0,171 (0,016**) -0,011 (0,875) -0,091 (0,199) 0,090 (0,204) 0,070 (0,325) -0,059 (0,407) 0,024 (0,731) 0,099 (0,163) -0,214 (0,002***)
cash financed acq. 1
-0,096 (0,177) -0,128 (0,072*) 0,175 (0,013**) -0,023 (0,751) 0,070 (0,325) -0,059 (0,407) 0,024 (0,731) -0,124 (0,081*) 0,088 (0,213)
size of the acquirer (b) 1
0,001 (0,984) -0,033 (0,646) 0,106 (0,134) -0,019 (0,795) -0,022 (0,761) -0,065 (0,364) 0,072 (0,311) 0,065 (0,362) no blockholders (<5% ) 1 -0,284 (0,000***) -0,305 (0,000***) -0,233 (0,001***) -0,331 (0,000***) -0,160 (0,024**) 0,119 (0,093*) 0,159 (0,025**) blockholders (>5% ) 1 -0,197 (0,005***) -0,151 (0,033**) -0,214 (0,002***) -0,103 (0,145) -0,088 (0,217) -0,006 (0,930) blockholders (>10% ) 1 -0,162 (0,022*) -0,230 (0,001**) -0,111 (0,117) -0,053 (0,455) -0,054 (0,446) blockholders (>15% ) 1 -0,176 (0,013*) -0,085 (0,232) 0,018 (0,795) -0,004 (0,959) blockholders (>20% ) 1 -0,121 (0,089*) 0,000 (1) -0,100 (0,161) blockholders (>50% ) 1 -0,038 (0,593) -0,040 (0,571) prior rumours 1 0,015 (0,834) Leverage (c) 1
(a) Target with the same 4 digit SIC code is considered related
(b) The market capitalization at the day before the start of the event window is proxy for size (c) Leverage is the ratio of debt over debt plus equity
Source: own calculations
** Significant at the 0.05 level (2-tailed)
* Significant at the 0.10 level (2-tailed)
non-listed C A R related block-holders (>15% ) block-holders (>10% ) block-holders (>5% )
*** Significant at the 0.01 level (2-tailed)
APPENDIX II
Graph 5
CAR divided in related and unrelated acquisitions
-0,03 -0,02 -0,01 0 0,01 0,02 0,03 0,04 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 0 D ays C u m u la ti v e a b n o rm a l re tu rn Whole samp le Related U nrelated
Source: own calculations
Graph 6
CAR divided in domestic and cross border acquisitions
-0,015 -0,01 -0,005 0 0,005 0,01 0,015 0,02 0,025 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 1 0 1 2 1 4 1 6 1 8 2 0 2 2 2 4 2 6 2 8 3 0 D ays C u m u la ti v e a b n o rm a l re tu rn Whole samp le D omestic Cross border
Graph 7
CAR divided in classes of blockholders
-0,06 -0,04 -0,02 0 0,02 0,04 0,06 0,08 -3 0 -2 8 -2 6 -2 4 -2 2 -2 0 -1 8 -1 6 -1 4 -1 2 -1 0 -8 -6 -4 -2 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Days C u m u la ti v e a b n o rm a l re tu rn Whole sample no blockholders >5% - 10% >10% - 15% >15% - 20% >20% - 50% >50%
APPENDIX III
Table 13
Significance of the sensitivity analyses
Analysis A-B A-C A-D B-C B-D C-D Analysis <5% >5% >10% >15% >20% >50% A-B -0,071 (0,944) -0,138 (0,891) -0,173 (0,864) -0,141 (0,889) -0,230 (0,820) 0,124 (0,903) A-C -0,014 (0,989) 0,215 (0,831) 0,340 (0,736) 0,519 (0,609) -0,469 (0,641) -0,334 (0,746) A-D 0,073 (0,942) -0,028 (0,977) -0,363 (0,719) -0,400 (0,693) 0,544 (0,590) 1,162 (0,272) B-C 0,057 (0,955) 0,355 (0,725) 0,528 (0,601) 0,687 (0,500) -0,236 (0,814) -0,418 (0,685) B-D 0,144 (0,886) 0,110 (0,913) -0,196 (0,846) -0,270 (0,790) 0,763 (0,450) 0,947 (0,366) C-D 0,086 (0,932) -0,217 (0,830) -0,744 (0,462) -0,822 (0,420) 1,115 (0,272) 1,695 (0,121) A is the initial analyses
B uses the adjustment of Blume
C uses a 375 day estimation period ending at the start of the event window
D uses a 250 day estimation period ending 125 days before the start of the event window Source: own calculations
