The influence of deal hostility on the announcement effect of stock price returns for European listed targets

(1)

The influence of deal hostility on the announcement

effect of stock price returns for European listed targets.

Abstract

This thesis examines the effect of deal hostility on the cumulative abnormal returns for European listed target firms in the period between February 1998 and January 2014. For the event window [-40,0] the average cumulative abnormal returns for the target amount to 21.8% for hostile deals while friendly deals have average cumulative abnormal returns of 0.98%. From the results it can however not be concluded that there is a significant positive influence of deal hostility on the cumulative abnormal returns at the 10% significance level. The insignificance of the results may be caused by the substantially smaller size of hostile deals.

Author Bas Geenen Student Number: 6156533 Supervisor Mr. M.A. Dijkstra MSc University of Amsterdam

BSc Economics & Business Track Economics & Finance Amsterdam, 21 February 2014

(2)

2

1 Introduction

Martynova and Renneboog (2008) find that the European takeover market is becoming more hostile. Goergen and Renneboog (2004) study the effect of announcements of mergers and acquisitions on the stock price returns for European bidder and target firms. They find a significant positive influence of a merger or acquisition announcement on the targets stock price returns. Goergen and Renneboog (2004) also find evidence that the stock price returns of target firms increase more in the event of a hostile deal than in a non-hostile deal. Their research is based on a sample which consist of European bidders and targets. In their sample only large deals with a value of >90 million Euros are included. This is a confirmation of the results found by Servaes (1991) who also found evidence of a significant positive influence of deal hostility on the target firm’s stock price returns.

In this thesis the focus will be on the influence of hostile deals for the period from

February 1998 until January 2014. The minimum deal value is set on 0.5 million Euros this so study will also include the deal hostility influence for smaller deal values. Further variation from the existing literature consists of only taking completed acquisitions into account and not restricting the acquirer to be a European firm. The research question is the following:

RQ: Does bid hostility influence the target stock price returns after takeover announcement?

In the second chapter of this thesis the economic theories used are explained and the existing literature on deal hostility and the possible influence of other variables on the cumulative abnormal returns on announcement are discussed. Secondly in the third chapter the research method, the expected effects and the gathering of the data are described. In the fourth chapter the results are presented, interpreted and compared to the existing literature. The fifth chapter discusses possible improvements and suggestions for further research. The sixth and final chapter concludes.

(4)

4

2 Literature Review

2.1 The announcement takeover effect

The announcement effect of takeovers is the effect that the target firm’s share price increases after the announcement that the target firm will be taken over (Goergen and

Renneboog, 2004). The announcement signals investors that the target firm will be taken over and investors expect that the target firm will benefit from synergies (e.g. economies of scale, cost benefits due to larger size). This expected future benefit will therefore increase the targets current value and this value increase will be positively translated into the stock price at the announcement date (Healy et al., 1997).

The announcement effect is measured in the cumulative abnormal returns (CAR). This percentage is calculated by cumulating the difference from the actual returns minus the normal returns for a certain amount of days, the amount of days used is called the event window.

Schwert (1996) shows that the stock price can increase before the actual

announcement, because some investors obtain information about the announcement before it is announced. These investors will be able to buy shares before the price increase occurs. Therefore the effect of rumors and information leakages may already be translated into the stock price before the announcement date.

Fama and Malkiel (1970) explain that that there are three forms of market efficiency, which define the manner in which the new information is being translated into the stock price. The weak-form efficiency is only based on historical stock prices and implies that no

expected excess returns can be obtained by only interpreting historical stock prices. The semi-strong form adds to the weak-form efficiency that no expected excess returns can be made based on public information. Non-public information however, known only by insiders such as the management or information which has leaked can be used to make profit. An

implication of the semi-strong efficiency hypotheses is the fact that the announcement effect starts before the actual announcement. This means that the information of the announcement has leaked or insiders act on their information before it becomes public. In the strong form efficiency hypotheses this possibility of making profits is also excluded since in the strong form efficiency hypothesis all information will already be translated into the stock prices.

(5)

5 Goergen and Renneboog (2004) find significant cumulative abnormal returns in

European countries of 9% for target firms for a two-day event window [-1,0]. If the event window is enlarged by including the price run-up period, which was two months in their study, the cumulative abnormal returns significantly increases to 23%.

2.2 Bid hostility

Mörk et al., (1988) consider a bid to be hostile when it is initially rejected by the target’s board or when the management actively resists the bid. Mörk et al., (1998) consider two possible ways for the management to resist. Firstly the management could invite a ‘white knight’, which is a friendly third party (Jaggia and Thosar, 1993), to enter the bidding. And secondly Mörk et al., (1998) consider the scenario of a management buyout in which the sitting management will entirely or partially buy the firm. Mörk et al., (1998) consider non-hostile bids to be friendly.

Chen and Cornu (2002) explain why hostile bids trigger a higher bid premium by the

inefficient target management hypothesis. This hypothesis is based on the assumption that the target’s current management is inefficient and that replacement by a higher skilled

management will lead to improved company’s results. Therefore the acquirer’s management will be willing to pay a higher hostile bid premium to gain effective control of the firm.

Schwert (2000) finds that management turnover significantly increases after a hostile takeover.

Goergen and Renneboog (2004) study the influence of deal hostility on the cumulative abnormal returns. They included 187 M&A’s that have taken place in 18 European countries in the period 1993-2000 and find that 70 out of 136 deals with listed targets occurred in England and Ireland. La Porta et al., (1997) found that Anglo-Saxon firms have significantly more shareholder protection. Rossi and Volpin (2004) find that good shareholder protection, which they measure on a scale from 0-6, triggers hostile bids. A one-point increase in shareholder protection significantly increases the probability of hostile bids by 0.8%. Shareholders with more shareholder protection will have more power to accept a hostile bid against the will of the board if it will benefit the shareholders.

Moschieri and Campa (2009) find in their research 97% of the completed deals in Europe between 2001 and 2007 are not hostile. With regard to the completion of announced deals they find 64% of friendly deals are completed while only 41% of hostile deals are completed. They also find that higher concentration of ownership triggers friendly deals. The concentration of ownership is measured by the average stake belonging to the three largest

(6)

6 shareholders. When a few large shareholders have effective control over the target they will be able to reject a hostile deal by which they would have lost their effective control. Therefore in Anglo-Saxon countries, with a higher degree of shareholder protection and less

concentrated ownership, hostile deals are more likely than in non-Anglo-Saxon countries. Goergen and Renneboog (2004) find significant cumulative abnormal returns of 13% for hostile acquisitions in the event window [-1,0] while friendly acquisitions only amount to 6%. For the interval with a price run-up period of 40 working days [-40,0] they find

significant positive cumulative abnormal returns of 29% versus 20% cumulative abnormal returns for friendly acquisitions.

Servaes (1991), who investigates the hostile deal effect for 704 deals in the period 1972–1987, includes a run-up period of eleven working days. He finds 32% hostile deal cumulative abnormal returns versus 22% cumulative abnormal returns for friendly bids.

Based on the literature by Georgen and Renneboog (2004) and Servaes (1991) it can be concluded that bid hostility has a positive effect on the cumulative abnormal returns of the target firm.

2.3 Economic downturn

Martynova and Renneboog (2008) conclude that mergers and acquisitions appear in a wave pattern. They distinguish five waves: the 1900’s, 1920’s, 1960’s, 1980’s and the 1990’s. The first three waves mainly occurred in the United States. The European takeover waves started to grow from the 1980’s. From the 1990’s and onwards the European takeover market started reach the same total deal value traded as the American takeover market (Martynova and Renneboog 2008).

Martynova and Renneboog (2008) conclude that all five merger waves began in periods of economic recovery and ended during the collapse of stock markets.

Bhagat et al., (2005) find that the announcement effect of takeovers for the target firm is significantly higher in periods of economic recovery than in periods of economic downturn.

Shleifer and Vishny (1991) argue that takeover hostility is related to the merger waves. They find a significant increase of hostile takeovers in the economic recovery of the 1980’s which they consider to be a response to the creation of the inefficient conglomerates from the 1960’s. Conglomerates were known to be active in several completely different businesses. Shleifer and Vishny (1991) argue that the increase in hostile deals was created because the conglomerates did not adapt to the changing market situations. By the means of hostile deals these inefficient companies were taken over and later restructured.

(7)

7 Martynova and Renneboog (2008) contradict these findings for the United States in the 1990’s, and find a significant decrease in hostile deal activity while an increase would be expected given the takeover wave of the 1980’s. This decrease may however be caused by the changed takeover regulations in the United States in the 1980’s. From the 1990’s and onwards the hostile deals started to increase in European countries. Martynova and Renneboog (2008) argue that the increasing amount of hostile deals may be caused by the decrease in ownership concentration.

Based on the literature by Shleifer and Vishny (1991) and Goergen and Renneboog (2004) it can be concluded that economic downturn negatively influences the amount of hostile deals in European markets.

2.4 Cash payment

Travlos (1987) distinguishes three different methods of payment: cash, equity and a combination of both.

Goergen and Renneboog (2004) argue that the choice of payment method depends on the perceived worth of the target firm estimated by the acquirer’s management compared to the targets market price. Goergen and Renneboog argue that if the management knows the target is undervalued it will prefer to pay in cash so only the acquirer will benefit from the perceived stock price increase. If the acquirer is unsure of the targets (potential) value the management will prefer to pay in equity. This can be seen as a method to share the risk of this uncertain value with the former owner (Travlos, 1987).

The method of payment plays a role for the target firm since the cumulative abnormal returns depend on the means of payment (Goergen and Renneboog, 2004). While equity bids have cumulative abnormal returns of 12.2% for the event window [-2,2] an all cash bid triggers 13.6% cumulative abnormal returns. The difference is significant at the 1% significance level.

If a run-up period of 40 working days is used the difference becomes even larger, the all cash cumulative abnormal return amount to a significant 27,5% and the equity bids amount to cumulative abnormal returns of 12.3%.

Based on the literature by Goergen and Renneboog (2004) it can be concluded that cash payment has a significant positive influence on the target cumulative abnormal returns.

(8)

8

3 Methodology and Data

3.1 Methodology

An event study analysis is performed to determine the effect of a particular event on the firm’s stock price. This can be investigated by comparing the actual stock prices of the firm to the expected stock prices that would have occurred without the influence of the event. The difference between the expected returns without the event and the actual returns are called the abnormal returns. The formula for the abnormal returns is the following:

(1) ARi,t = Ri,t – ERi,t

Where:

ARi,t are the abnormal returns of firm i on day t, Ri,t are the actual returns of share i on day t and ERi,t, are the normal or expected returns of share i on day t

De Jong (2007) explains an event study should be performed in three steps: first of all the event of interest has to be identified, secondly a benchmark model will have to be specified by which the normal expected returns can be calculated, thirdly the abnormal returns that result from the first two steps, the difference between the expected and the actual returns can be calculated and interpreted.

The event is the announcement of an acquisition (t=0) The event windows which will be investigated are [-2,2] and [-40,0], two days before announcement up to and including two days after announcement day, and forty days before announcement up to and including the announcement day. The event windows are based on the event windows used in the article by Goergen and Renneboog (2004).

The market model is used to compute the daily expected returns. The researches by Moschieri and Campa (2009) and Goergen and Renneboog (2004) both used the market model. To be able to compare the results from this study the market model will be used. An OLS regression estimates each firm’s alfa and beta for the market model. This regression is based on the daily target stock price data and the EUROSTOXX 50 stock price data for the period from 260 working days (one year) till 40 working days before announcement of the acquisition [-260, -40]. This period excludes the announcement date and 40 working days previous to it so the announcement effect and possible effects from insider trading are not included (De Jong, 2009).

(9)

9 The values found for each firm’s alfa and beta will be used to calculate the daily expected returns of the target firms using the EUROSTOXX 50 index as benchmark. The market model formula looks as follows:

(2) ER i,t = α + βiRmt

Where:

ER i,t are the expected returns of firm i on day t based on the firm’s α and β, α is an estimated constant for firm i, βi is firm i’s estimated coefficient of the responsiveness of the stock price to the market, Rmt are the market returns on day t.

Daily abnormal returns will be determined for each firm by subtracting the expected from the actual returns (eq. 1).

When the daily abnormal returns are computed the cumulative abnormal returns can be computed. This is done by cumulating the abnormal returns for the required number of days in the event window. The formula looks as follows:

(3) CARi,t = ∑−𝑛𝑛 ARi,t

Where: CAR i,t stands for the cumulative abnormal returns of target firm i for period t and ARi,t stands for the sum of the abnormal returns for firm i for the event window [-n,n].

(10)

10

3.2 Regression

The model which will be tested is the following:

car

i,t

= β

0

+ β

1

hos + β

2

crs + β

3

cr*hs + β

4

cash + β

5

ln(asset) + β

6

ln(deal) + β

7

gb +

β

8

fr + β

9

be + β

10

it + β

11

se + β

12

pl + β

13

dk + β

14

fi + β

15

de + ε

i,t Table 1: Regression variables

Variable Type Definition

car continuous cumulative abnormal returns(%)

hos dummy 1 = hostile, 0 ≠ hostile

crs dummy 1 = if after 15-09-2008, 0 if before 15-09-2008 cr*hs dummy 1 = both crs and hos are 1, 0 ≠ both crs,hos

cash dummy 1 if cash, 0 if equity

ln(asset) continuous natural logarithm of target’s total assets ln(deal) continuous natural logarithm of deal value

gb dummy 1 = United Kingdom, 0 ≠ United Kingdom

fr dummy 1 = France 0 ≠ France

be dummy 1 = Belgium, 0 ≠ Belgium

it dummy 1 = Italy, 0 ≠ Italy

se dummy 1 = Sweden, 0 ≠ Sweden

pl dummy 1 = Polen, 0 ≠ Polen

dk dummy 1 = Denmark, 0 ≠ Denmark

fi dummy 1 = Finland 0 ≠ Finland

de dummy 1 = Germany, 0 ≠ Germany

The Netherlands, nl, are taken as reference point so no nl dummy is included.

In the next section the expected signs for the estimated coefficients will be predicted. The main explanatory variable used is the deal hostility. The group of hostile deals consists of deals that ZEPHYR classified as hostile initially. This group includes deals that later became recommended. At the time of the announcement the shareholders of the target firm do not know the deal will become recommended and it will therefore not influence the announcement effect.

As described by Goergen and Renneboog (2004) the cumulative abnormal returns of hostile deals are expected to be higher than the cumulative abnormal returns of non-hostile

(11)

11 deals since the cumulative abnormal returns found were 13% on the announcement day versus 6% for non-hostile deals. The coefficient of deal hostility is therefore expected to be positive.

To correct for possible effects of the crisis on the cumulative abnormal returns a dummy is included for deals that have taken place after the start of the crisis and an interaction term for crisis with deal hostility is included for a possible influence of the crisis on the hostile deal effect. The fall of the Lehman Brothers is taken as start of the financial crisis so deals with an announcement date after 15-09-2008 will include the dummy variable crisis. Based on the findings by Bhagat et al., (2005) it is expected that the crisis coefficient will be negative because in times of economic downturn the cumulative abnormal returns for the target will be lower.

Shleifer and Vishny (1991) and Martynova and Renneboog (2008) find that economic downturn is expected to negatively influence the amount of hostile deals for European targets. They do not consider the profitability of hostile deals during crisis. However it could be argued that managers will be more careful using hostile deals in periods of financial crisis since Moschieri and Campa (2009) find that hostile deals are less likely to be completed (41% hostile deals versus 65% of friendly deals is completed). Consequently managers will only acquire highly profitable targets, this would result in higher cumulative abnormal returns for hostile deals in periods of economic downturn.

The second control variable which is included is a dummy that controls for cash payments, from existing literature it appears that cash payments positively influence the cumulative abnormal returns (Goergen and Renneboog, 2004). They find that paying in cash results in 10% cumulative abnormal returns while paying in equity only triggers 6.7% cumulative abnormal returns. The coefficient for cash payment is therefore expected to positively influence the cumulative abnormal returns of the target firm.

Moeller et al., (2004) find that in the case of larger firms the short term cumulative abnormal returns are expected to be lower since it takes longer to implement the new management structure in larger firms, which are known to be harder to manage efficiently

(Chen and Cornu, 2002). Therefore it is expected that the target firm size will negatively

influence the (short term) cumulative abnormal returns.

There is no indication that deal value will influence the cumulative abnormal returns from existing literature but since there is a large variation in deal value (600 million Euros as biggest value and 1.2 million Euros as smallest) this control variable is added to control for possible effects of deal value on the cumulative abnormal returns.

(12)

12 From the article by Goergen and Renneboog (2004) it appears that if targets are listed in Great-Britain the cumulative abnormal returns significantly increase. The cumulative abnormal returns for Great-Britain based targets is as much as 12.3% on announcement day while in continental Europe the cumulative abnormal returns are only 6%. Because many different European deals are investigated for each country a dummy variable will be added to the model. The Great-Britain dummy’s coefficient is expected to be positive since the

cumulative abnormal returns in Great-Britain are expected to be higher.

3.3 Hypothesis

The following hypothesis will be tested:

Bid hostility has no influence on the target’s cumulative abnormal stock price returns after takeover announcement.

This implies: H0: β1 = 0

Bid hostility does influence the target’s cumulative abnormal stock price returns after takeover announcement.

(13)

13

3.4 Data

This study investigates European acquisitions for the period from the 28th of February 1998 until the 31th of December 2013. This period is chosen because the EURO STOXX 50 index, which is used as benchmark, has become available since that date.

While selecting the sample in ZEPHYR the following search criteria are used:

the announcement date is between 28-02-1998 and 31-12-2013, the target is listed in the EU, the acquisitions are completed, the deal value is at least 0.5 million Euros, ZEPHYR classifies the deal as hostile or initially hostile.

Using these search criteria 35 deals are found.

To be able to check for the effect of deal hostility non-hostile deals are included in the sample. When the hostility criterion is dropped from the search criteria 59036 deals remain in the sample. The 35 hostile deals are subtracted and from the remaining 59001 non-hostile deals 35 non-hostile deals are randomly selected using Excel.

The data obtained about these deals is the following: payment method, deal value (million Euros), target total assets (million Euros), target ISIN number, and target country code.

The daily stock price returns of the target firms are retrieved from DataStream. After retrieving the stock price data only 27 hostile and 27 non-hostile firms remain, for the other 16 listed firms DataStream could not retrieve daily stock price data because an error occurred or only weekly stock prices were available.

The daily stock price returns for the benchmark, the EURO STOXX 50 index will be retrieved from DataStream as well. The EURO STOXX 50 is chosen as benchmark because all the firms in the sample are listed in Europe and the EURO STOXX 50 is based on the 50 most important stocks in Europe.

(14)

14

Table 2. Hostile and Non-Hostile Target Characteristics

Avg. St.dev. Max. Min.

Hostile CAR (%) [-2,2] 6.10 10.04 33.04 -2.97 CAR (%) [-40,0] 21.85 31.27 139.49 -14.08 Deal value (M) 475.70 1425.35 6949.00 1.22 Target assets (M) 2973.86 9147.28 45602.00 4.57 Non-Hostile CAR (%) [-2,2] 1.03 2.43 5.66 -3.67 CAR (%) [-40,0] 0.98 14.03 59.80 -20.30 Deal value (M) 58.31 180.54 911.15 1.25 Target assets (M) 31603.64 62075.74 227516.27 43.45

A large difference in average deal value and average target assets can be identified. Where deal value is substantially larger for hostile deals the target’s assets are substantially larger for non-hostile deals.

In graph 1 below an illustration of the daily cumulative abnormal returns for the event window [-40,2] is given. Graph 2, included in the appendix, provides an illustration of daily average abnormal returns for the event window [-40,2].

Graph 1. Overview of the daily Cumulative Average Abnormal Returns for [-40,2]

This graph shows that the daily cumulative average abnormal return is substantially larger for hostile deals than non-hostile deals. Especially on the day before announcement a larger increase is seen in for hostile deals compared to non-hostile deals.

(15)

15

Table 3. Descriptive statistics of dummy variables

Variable name N Value Percentage

Hostile 27 0= Non-Hostile 50%

27 1= Hostile(22) or initially Hostile(5) 50%

Crisis 27 0= Before 15-09-2008 50%

27 1= After 15-09-2008 50%

Crisis*Hostile 19 0= Not both Crisis and Hostile 35.2%

. 35 1= Crisis and Hostile 64.8%

Cash Payment 11 0= Equity 20.4%

(16)

16

4 Results

This section discusses the results for the cumulative abnormal returns of the 54 European target firms. First of all the results found for the influence of bid hostility on the target’s cumulative abnormal returns for the event window from two days before announcement up to and including two days after announcement are discussed. Secondly the results for the

influence of bid hostility for the event window from forty days before announcement up to and including the announcement date will be discussed.

4.1 Event window [-2,2]

Table 4: regressions on cumulative abnormal returns for the event window [-2,2]

CAR[-2,2] CAR[-2,2] CAR[-2,2] CAR[-2,2] CAR[-2,2]

1 2 3 4 5 hos 5.064** 4.633** 5.581** 0.193 -0.612 (1.988) (2.193) (2.333) (3.002) (4.447) crs 1.056 0.962 0.860 0.723 (2.193) (2.250) (2.124) (3.371) cs*hs 0.686 (4.772) ln(deal) -0.677 0.233 0.497 (0.516) (0.596) (0.740) ln(assets) -1.374** -1.185** (0.520) (0.570) cash -0.865 0.519 0.179 (2.577) (2.489) (2.818) gb 5.012 (7.734) Constant 1.032 0.719 7.644 19.303*** 9.303 (1.405) (1.558) (5.482) (6.804) (10.087)

Other country dum. No No No No Yes

R-squared 0.1109 0.1150 0.1464 0.2546 0.4157

Adjusted R 0.0939 0.0803 0.0767 0.1769 0.1851

Number of obs. 54 54 54 54 54

Significance levels at 1%, 5% and 10% presented by ***, **, *, respectively, Standard errors are between brackets

(17)

17 Model 1 in table 4 shows that when only the deal hostility and a constant are included in the regression model there seems to be a significant positive influence of deal hostility on the target cumulative abnormal returns. When the deal is hostile the average cumulative abnormal returns will be 5.1% higher than in non-hostile cases, significant at the 5% significance level. On the other hand it is hard draw conclusions based on the regression model with only one dummy variable and a constant. When the adjusted R-squared is taken into account, which gives an idea of the amount of variation which is explained by the model adjusted for the amount of variables, only 9.4% is explained.

In model 2 crisis is added to the model and the coefficient for deal hostility decreases to 4.6%, but remains significantly positive at the 5% level. The crisis coefficient is positive, which does not stroke with the expectation, and is not significant. The coefficient crisis has a negative influence on the adjusted R-squared since it decreases by 1.3%. By including the crisis coefficient the amount of explained variation decreases after correcting for the extra coefficient so it is still questionable whether we can draw conclusions based on the effect of two dummies.

In model 4 all control variables except for the country effects and the interaction term between crisis and hostile are added. In this model the hostile deal effect is no longer

significant. The only difference of model 4 compared to model 3 is that the target asset coefficient is included into the model. So adding the target assets coefficient leads to insignificance of the positive deal hostility coefficient.

When looking at the results from model 5, in which all control variables are added, the adjusted R-Squared amounts to 0.185 which indicates that 18.5% of the variation can be explained by the model. No significant influence of bid hostility on the target cumulative abnormal returns can be identified. Moreover a negative influence of bid hostility is suggested by the coefficient of -0.612. This is not in line with the expectation of a positive relation between bid hostility and the target’s cumulative abnormal returns, and not in line with the results found in model 1-4. The insignificant Great-Britain dummy has a positive coefficient which suggests that the cumulative abnormal returns will be 5% higher for targets listed in Great-Britain compared to Dutch targets. This strokes with the expectation of higher cumulative returns for Great-Britain based targets.

(18)

18

4.2 Event window [-40,0]

Table 5: regressions on cumulative abnormal returns for the event window [-40,0]

CAR[-40,0] CAR[-40,0] CAR[-40,0] CAR[-40,0] CAR[-40,0]

1 2 3 4 5 hos 20.834*** 15.350** 18.117** 7.233 5.117 (6.598) (7.048) (7.355) (9.864) (13.547) crs 13.461* 14.515** 14.309** 9.807 (7.058) (7.092) (6.978) (10.271) cs*hs 13.421 (15.541) ln(deal) -3.031* -1.194 -1.957 (1.626) (1.959) (2.256) ln(assets) -2.775 -0.976 (1.709) (1.173) cash 4.007 6.804 2.549 (8.125) (8.176) (0.286) gb -4.593 (23.564) Constant 1.032 0.719 21.770 45.318** 33.300 (1.405) (1.558) (17.285) (22.350) (30.734)

Other country dum. No No No No Yes

R-squared 0.1109 0.1150 0.2733 0.3111 0.5354

Adjusted R 0.0939 0.0803 0.2140 0.2393 0.3521

Number of obs. 54 54 54 54 54

Significance levels at 1%, 5% and 10% presented by ***, **, *, respectively Standard errors are between brackets

For the event window [-40,0] the same five models are tested to be able to compare the results from the different event windows.

In the first regression the positive coefficient for hostility has increased from 5.064 till 20.834 and is significant at the 1% significance level. This increase is similar to what was expected.

In the second regression the bid hostility has decreased to 15.350, this is similar to the findings in the [-2,2] event window and is caused by adding the crisis dummy to the model. Contrary to the first model tested the crisis dummy is significant at the 10% significance level. The coefficient of crisis is again positive which does not stroke with the expectations. In

(19)

19 the second model the adjusted R-squared decreases by the same amount which means the amount of explained variation again decreases after correcting for the extra coefficient.

In the fourth model the now insignificant coefficient of deal hostility decreases, remains positive and again becomes insignificant. The interaction variable between crisis and hostile increases strongly, the crisis first positively influenced the cumulative abnormal returns of hostile deals by 0.1230% for the event window [-2,2] now the crisis positively influences the cumulative abnormal returns by 10.954% for the event window [-40,0]. This result is not significant and it does not stroke with the negative influence of crisis on the cumulative abnormal returns which was expected. The coefficient of the target total assets is no longer significant and remains negative in the second event window, this strokes with the expectations. When the third model is compared to the fourth the addition of the target total assets coefficient again causes the deal hostility coefficient to become insignificant.

In the fifth model hostile deals to create a 5.117% increase in the cumulative abnormal returns, this contradicts the findings in the [-2,2] event window and strokes with the expected positive influence of deal hostility on the cumulative abnormal returns which was also

identified in the first four models of both event windows. This coefficient is not significant. In the fifth model the coefficient for Great-Britain negatively influences the cumulative

abnormal returns by 4.593%, this result is not significant and does not stroke with the findings in the first event window and the expected higher cumulative abnormal returns for Great-Britain based firms. The adjusted R-squared has increased to 0.3521 in the fifth model for the [-40,0] event window, in the [-2,2] event window it was only 0.1851, this means an increase of more than 16% in the amount of variation explained.

(20)

20

4.3 Discussion of the Results

In this section the results found will be compared to those found in the literature review. For both event windows there is positive influence of deal hostility on the cumulative

abnormal returns when no control variables or country dummies are added. These results are in line with the findings of Servaes (1991) and Goergen and Renneboog (2004) who find a significant positive influence of deal hostility on the cumulative abnormal returns. Goergen and Renneboog (2004) find significant cumulative abnormal returns for hostile deals of 6.62% higher than the non-hostile cumulative abnormal returns. For the [-40,0] event window

Goergen and Renneboog (2004) find cumulative abnormal returns for hostile deals of 9.89% higher than non-hostile cumulative abnormal returns.

The first two models for both event windows in this study suggest that the positive influence of deal hostility is between 5.1% and 4.6% for the [-2,2] event window and between 20.8% and 15.4% for the [-40,0] event window. These results are significant but explain only 8-9% the variation for both event windows and only two dummy variables are used to describe the data. Based on the third model a significant 18.1% increase in cumulative abnormal returns is found for the [-40,0] event window in case of a hostile takeover, the adjusted R-squared increases to 21.4% for this model. Based on these findings the positive effect of deal hostility on the cumulative abnormal returns would even be larger than in the study by Goergen and Renneboog (2004).

In model 4 the coefficient target asset value, significant for both models in the [-2,2] event window, is added to the regression, after addition of the target asset value the effect of deal hostility is reduced and becomes insignificant for both event windows. These findings are in line with the findings by Moeller et al., (2004) who argue that the short term cumulative abnormal returns for large firms are expected to be lower since it takes longer to implement the new management structure in larger firms. This strokes with the fact that the target assets coefficient is more significant in the [-2,2] event window than in the [-40,0] event window. The fact that the deal hostility coefficient becomes insignificant when target assets are added to the model could also be explained by the fact that hostile deals involve substantially smaller targets. Table 2 shows that the average target asset value for hostile deals is 2973.86 million Euros while the average value of target assets for non-hostile deals is 31603.64 million Euros. Therefore the hostile deal effect will be reduced by adding target assets to the model.

For the event window [-2,2] the influence of deal hostility on the cumulative abnormal returns even becomes negative, this contradicts the findings by Goergen and Renneboog

(21)

21 (2004) and the findings by Servaes (1991), who found cumulative abnormal returns which are 10% higher for hostile deals when adding control variables.

With respect to the crisis for all regressions in both event windows positive values are found. In the second event window the variable for crisis is significant at the 10% level. These findings contradict the expected negative influence of crisis on the cumulative abnormal returns which were expected for European countries. This may be caused by the fact that the reaction to economic downturn does not appear immediately or the possibility that this crisis is different from previous ones with respect to the cumulative abnormal returns earned by targets.

With respect to the interaction term of hostility and crisis the values found are positive and insignificant for both event windows. The positive effect indicates that hostile deals trigger higher cumulative abnormal returns in economic downturn. This could possibly be explained by managers being more careful selecting hostile deals during crisis. Especially since hostile deals are less likely to succeed (Moschieri and Campa, 2009).

The control variable cash payment shows an insignificant positive effect for models 3 and 4 in both event windows. This is in line with the expectations based on the article by Goergen and Renneboog (2004) who found significant 15% higher cumulative abnormal returns for cash deals when cash and equity deals are compared for the event window [-40,0]. In this study the insignificant effect for this event window lies between 4.1% and 2.5%.

For the control variable Great-Britain the results are insignificant and differ in the event windows tested, the [-2,2] event window suggests a positive influence of 5.5% on the cumulative abnormal returns, which was expected by Goergen and Renneboog (2004) who found significant 8.6% higher cumulative abnormal returns for the same event window. This could be based on less concentrated ownership and higher amount of shareholder protection (La Porta et al., 1997). For the [-40,0] event window 23.4% higher cumulative abnormal returns were found by Goergen and Renneboog (2004). In this study an insignificant negative 4.6% influence on the cumulative abnormal returns is found for the event window [-40,0]. This negative influence of Great-Britain on the cumulative abnormal returns is strange since the values for the second event window [-40,0] contradict the values found in the first event window [-2,2] while a longer price run-up period is included.

(22)

22

5 Discussion

Due to the insignificance of the results for the influence of bid hostility on the target cumulative returns a discussion is in order. In this study several differences have been adopted compared to the study by Chen and Cornu (2002), who also focus on the influence of hostility on the cumulative abnormal returns using an event study. These differences possibly

contribute to the insignificance of the results. The most important difference seems to be the fact that Chen and Cornu (2002) used a matched sample. This way deals are selected that are comparable apart from the researched hostility while in this study a random sample of friendly deals was taken. This possibly led to the high, and significant negative value for in the control variable target assets which indicates that the cumulative abnormal returns are negatively influenced for large targets. In the sample taken it appeared that on average targets from hostile deals are more than ten times smaller than non-hostile targets, this may have led to the insignificance of the hostile deal coefficient.

(23)

23

6 Conclusion

In this thesis the influence of bid hostility on stock price returns after announcement of an acquisition is studied. The stock price returns after announcement are measured by the

cumulative abnormal returns of the stock prices for the event windows [-2,2] and [-40,0]. When only the bid hostility and a constant are included in the regression model a significant positive effect is found for both event windows, for the [-2,2] event window the predicted positive effect of hostility on the cumulative abnormal returns is 5.1% and for the second event window the predicted positive effect amounts to 20.8%. Given the fact that no control variables are added to this regression no conclusions can be drawn based on this effect.

For the [-40,0] event window an 18.1% increase in the target’s cumulative abnormal returns is found when the acquisition is hostile. This effect is significant at the 5%

significance level when the deal value and cash payment control variables are added to the model.

For both event windows no significant effect for bid hostility can be identified for the regression model in which all control variables are added. This means there is no hard evidence that bid hostility positively influences the cumulative abnormal returns of target firms. Based on the findings for the [-40,0] event window, the results suggest that there is a positive influence of 5.1% for the model where all control variables are included but this effect is not significant at the 10% significance level. Therefore H0, that states that bid hostility does not influence the cumulative abnormal returns of the target firm after announcement, cannot be rejected.

The results suggests there is a positive effect of crisis on the cumulative returns. This effect is insignificant and was not expected but can possibly be explained by a delayed effect for the influence of crisis on the target cumulative abnormal returns or the possibility that this crisis differs from previous crises. The influence of crisis on the deal hostility is insignificant and positive and may be explained by the fact that managers are more careful selecting profitable deals during the crisis, generating higher cumulative abnormal returns.

(24)

24

References

Andrade, G., Mitchell, M., Stafford, E., (2001) New Evidence and Perspectives on Mergers, Harvard Business School Working Paper No. 01-070

Chen, C., & Cornu, P. (2002). Managerial performance, bid premiums, and the characteristics of takeover targets. Annuals of Economics and Finance, 3(1), 67-84.

Goergen, M., & Renneboog, L. (2001). Investment policy, internal financing and ownership concentration in the UK. Journal of Corporate Finance, 7(3), 257-284.

Goergen, M., & Renneboog, L. (2004). Shareholder Wealth Effects of European Domestic and Cross‐border Takeover Bids. European Financial Management,10(1), 9-45. Jaggia, S., & Thosar, S. (1993). Multiple bids as a consequence of target management

resistance: a count data approach. Review of Quantitative Finance and Accounting,

3(4), 447-457.

Malkiel, B. G., & Fama, E. F. (1970). Efficient Capital Markets: A Review Of Theory And Empirical Work*. The journal of Finance, 25(2), 383-417.

Martynova, M., & Renneboog, L. (2008). A century of corporate takeovers: What have we learned and where do we stand?. Journal of Banking & Finance,32(10), 2148-2177.

Moeller, S. B., Schlingemann, F. P., & Stulz, R. M. (2004). Firm size and the gains from acquisitions. Journal of Financial Economics, 73(2), 201-228.

Morck, R., Shleifer, A., & Vishny, R. W. (1988). Characteristics of targets of hostile and friendly takeovers. In Corporate takeovers: Causes and consequences (pp. 101-136). University of Chicago Press.

Rossi, S., & Volpin, P. F. (2004). Cross-country determinants of mergers and acquisitions. Journal of Financial Economics, 74(2), 277-304.

Schwert, G. W. (1996). Markup pricing in mergers and acquisitions. Journal of Financial

economics, 41(2), 153-192.

Schwert, G. W. (2000). Hostility in takeovers: in the eyes of the beholder?. The Journal of

Finance, 55(6), 2599-2640.

Shleifer, A., & Vishny, R. W. (1991). Takeovers in the'60s and the'80s: Evidence and Implications. Strategic Management Journal, 12(S2), 51-59.

Stigler, G. J. (1950). Monopoly and oligopoly by merger. The American Economic

(25)

25

Travlos, N. G. (1987). Corporate takeover bids, methods of payment, and bidding firms' stock returns. The Journal of Finance, 42(4), 943-963.

(26)

26

Appendix

Graph 2. Overview of the daily Average Abnormal Returns for [-40,2]

A large increase in average abnormal return can be identified for hostile deals on the day before announcement.

The influence of deal hostility on the announcement effect of stock price returns for European listed targets