The announcement effect of acquisitions on shareholders’
wealth
The case of Western European investment in developed and developing countries
Author:
Jennifer Elisabeth Beeuwkes
University of Groningen
Faculty of Economics and Business
Business Administration, MSc Finance
Supervisor:
A.G. Schertler
2
The announcement effect of acquisitions on shareholders’
wealth
The case of Western European investment in developed and developing countries
Abstract
This paper studies whether the announcement effect of domestic and cross border acquisitions undertaken by Euronext firms has an effect on shareholder wealth of the acquiring and target firms. The sample consists of 82 transactions between the period of 2000 and 2010. The bidding shareholders
lost 0.496% and the target shareholders gained 6.852% on the announcement date. The findings suggest that the primary sources of the wealth gains for acquirers are whether they are in a related
industry, the size of the acquirer, the law of the target firms’ country, the level of economic development of the target firms’ country and the cultural distance between the target and acquiring
firms’ countries . Weak evidence is found for the method of payment and the size of the target determining wealth effects. Finally, shareholder protection is not found to be a source of wealth
creation.
JEL codes: G14, G34, G38
Key words: wealth effects, European, event study, acquisitions
3
Preface
This is my final assignment at the University of Groningen. My research started in December 2010 and was finalized in September 2011. During this period I got the opportunity to do an internship at Nestlé, where I discovered what I wanted to do in my further career. Even though it took me a long time to accomplish my thesis, I am thankful for this enriching experience.
I would like to express my gratitude to Ms. Schertler. I am thankful for her patience during this long path. She provided me with helpful insights and guidance during my research process. Futhermore, I would like to thank my family, my friends and boyfriend for their continuous support.
4
TABLE OF CONTENTS
1. INTRODUCTION ...5
2. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT ...8
2.1 Acquisitions and shareholder value ...8
2.2 Factors determining the impact on shareholder value ... 14
2.2.1. Deal, acquiring and target firm characteristics ... 14
2.2.2. Country characteristics ... 17
3. DATA ... 22
3.1 The sample ... 22
3.2 Dependent and independent variables ... 23
4. ABNORMAL RETURNS ... 26
4.1 Event study methodology ... 26
4.2 Descriptive statistics ... 28
4.3 Tests on significance ... 30
4.4 Event study results and discussion ... 31
5. THE DETERMINANTS OF THE ACQUISITION RETURNS ... 36
5.1 Ordinary Least Squares regression methodology ... 36
5.2 OLS regression results and discussion... 36
5.3 Robustness checks ... 40
6. CONCLUSIONS... 42
7. REFERENCES ... 46
5
1. INTRODUCTION
It has been broadly documented that acquisitions occur in waves (Hardford, 2005; Dong et al., 2006; Mitchell and Mulherin, 1996). Five waves have been identified in the literature, where the fifth and last merger wave was from 1993 to 2000. From $342 billion of deals in 1992, the worldwide volume of mergers increased steadily to $3.3 trillion worldwide in 2000 and nine of the ten largest deals in history took place during this wave (Lipton, 2005). The aim of this paper is to investigate the gains from acquisitions after this wave. This is especially interesting when including the period of the crisis from 2007 to 2010. Particularly when the findings of Mitchell and Mulherin (1996), that the
acquisition waves depend on the economic health and the emotional state of mind of the corporate world, are taken into consideration. In total European acquisitions accounted for $6.3 trillion in the period from 2000 to 2009. On average they accounted for 25% of the value of all the deals worldwide. It peaked in 2007 covering a value of $1.2 trillion and was 29.0% of the total worldwide value as depicted in figure I 1. The value decreased in 2008 and 2009, but the percentage of European deals compared to the total rose to 31.2% in 2008.
Figure I. Total deal values of completed mergers and acquisitions in the period from 2000 to 2009 (Financial Times)
Martynova and Renneboog (2008) draw attention to the geographical dispersion of the fifth wave, which was characterized by cross-border mergers. According to them the discovery that Continental European firms were participating on a similar level as their US and UK counterparts combined with
1
Financial times; http://www.ft.com/intl/cms/s/2/6cba359a-31ce-11df-9ef5-00144feabdc0.html#axzz1WLvPlQCl 0 1000000 2000000 3000000 4000000 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Overall Mergers & Acquisitions
6
the introduction of the Euro, the globalization process, technological innovation, deregulation and privatization, as well as the financial markets boom, drove European companies to take part in M&As during the 1990s.
Hernando and Campa (2004) argue that resistance against cross-border acquisitions within the Euro zone could be an indication that there are still a large number of legal, economic and cultural restraints to this activity. However, to sustain their competitive position in this highly competitive world, multinationals from Western European countries frequently turn to mergers and acquisitions. Developing countries are increasingly becoming an attractive target for Western European investors because of greater economic incentives, lower costs, and greater growth opportunities in developing countries. As a consequence of the worldwide integration of capital and product markets, and the emergence of new markets, globalization has become a crucial strategic issue for companies (Moeller and Schlingemann, 2005). Consequently, previously locally focused corporations often engage in cross-border M&A in order to cope with the tough global competition (Martynova and Renneboog, 2006).
The primary intention of this paper is to investigate whether the announcement effect of domestic and cross border acquisitions undertaken by Western European firms with public targets has an effect on shareholder wealth of the acquiring and target firms. This study is restricted to the separate outcomes of target and acquiring firm shareholders’ wealth and does not study the joint wealth effects. The study is carried out in three parts. First, the existing literature and theory on motives for acquisitions in general will be analyzed in order to formulate predictions about their sources of wealth creation. The second part calculates the wealth gains to bidding and target shareholders following the announcement of acquisitions, using the event study methodology explained by Brown and Warner (1985). The value effects are calculated based on share price performance. Finally, deal, firm and country characteristics which may explain wealth gains to bidding firm shareholders are identified.
The announcement effect will be related to different characteristics of target and bidding firms and countries and of the bid itself:
The two main questions that will be answered are:
- What overall impact do acquisitions with European bidders and public targets have on shareholders wealth during the period from 2000 to 2010?
- What were the sources of wealth gains for bidding firms’ shareholders from acquisitions with European bidders during the period from 2000 to 2010?
7
contributions are made to previous studies. This thesis will differentiate between firm, deal and country characteristics. Until now, the main focus in previous literature was on firm and deal
8
2. LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
This section will discuss the theoretical background and develop hypotheses concerning the
announcement effect of acquisitions of companies and different target companies on their shareholder value. Firstly, prior studies on the bidder and target firm returns following from an acquisition will be presented followed by hypotheses on this matter. Secondly, characteristics of the acquiring and target firm, the acquiring and target country and the acquisition deal that might determine the impact on shareholder value will be identified, where after the remainder of the hypotheses will be derived.
2.1 Acquisitions and shareholder value
Acquiring firm
Within studies on foreign direct investment, much of the research is done on the reasons for acquisitions. The results of previous research indicate that the main motive for firms entering into mergers and acquisitions are synergies. The literature points out two other recurrent reasons for mergers and acquisitions, namely maximizing shareholders’ wealth and management failure (Georgen and Renneboog, 2006; Ravenscraft and Scherer, 1989; Jensen, 1986). Acquisitions motivated by synergies are expected to create positive wealth effects. Operating synergies and informational synergies are the two different types. Operating synergies are accomplished if there is economies of scope and/or scale in the industry or if gains can be created from knowledge transfer, while
informational synergies are accomplished if the value of the merged firms is higher than the stand-alone values (Georgen and Renneboog, 2004). In addition to the two types of synergies mentioned by Georgen and Renneboog (2004) I would like to add the financial synergy following a reduction in costs of capital. Financial synergies may include improved cash flow stability, lower bankruptcy probability (Higgins and Schall, 1975) and cheaper access to capital (Martynova and Renneboog, 2004). Synergies are gained through three different types of acquisitions, namely horizontal, vertical or conglomerate. According to Berger and Ofek (1995) the type of acquisition has an impact on the level of abnormal returns generated by the acquisition. Horizontal mergers are undertaken to expand the current market. Vertical mergers are used to take control over a larger part of the production process, in order to decrease supplier or buyer power or to gain access to specific expertise. A vertical merger is more a supply chain integration. A conglomerate merger is a merger between two companies that operate in different industries, also called a diversifying merger. In theory, if acquisitions are motivated by synergies, the value of the newly merged firm will be higher than the stand-alone values. The added value will be shared between the target and the acquirer shareholders. Thus, when
9
and the target. According to several studies (Doukas and Travlos, 2002; Martynova and Renneboog, 2006; Hernando and Campa, 2004) diversifying mergers create most synergies.
However, negative abnormal returns could be expected from management failure which suggests that acquisitions are motivated by managers’ mistakes in evaluating targets. As a result managers engage in takeovers even when there are no synergies. According to Roll (1986), this is called the hubris motive, which explains why bids are made even when a valuation above the current market price represents a positive valuation error. Therefore, bidding firms pay too much for their targets. When there are no synergies, shareholders wealth shifts from the bidder to the target. Thus, acquisitions based on management failure as a consequence of the hubris motive, lead to wealth losses for the bidding firms’ shareholders and wealth gains for the target firms’ shareholders.
Managerial failure can also result in negative abnormal returns due to the agency conflict, with bidding company management aiming to maximize their own utility (Danbolt, 2004). By entering into acquisitions, managers create growth increasing manager’s power by enhancing the resources under their control (Jensen; 1986) and also managers compensation which is proved to be positively related to growth (Ciscell and Carroll; 1980). This theory suggests that self-interest of bidding firms
management is a motive for M&A’s. Agency costs are caused by a separation of ownership and control. Management (the agent) can use corporate assets to pursue investment strategies that yield them personal benefits of control, such as growth or diversification, without benefiting other
shareholders (the principal) (Jensen, 1986). Acquisitions motivated by management failure as a result of agency conflicts primarily take place because they increase bidding management’s wealth at the expense of acquiring shareholders.
Do the theories above match the data? The majority of the research on the bidding and target firms shareholders wealth is focused on the US and the UK in the 1980’s and 1990’s (Moeller and
10
Table I
Summary of results of previous event studies on the
announcement effect of acquisitions: returns to acquiring firms' shareholders Authors Region Research
period
Primary event window (days)
Positive/negative returns
Country coverage Number of observations
Additional results/Notes
Asquith, Bruner and Mullins
(1983)
US 1955-1979 (t=0) AR (2.8%) Domestic 214 Bidders' abnormal returns are positively related to the relative size of the merger partners.
Jensen and Ruback
(1983) US Till 1983 N/A Bidders do not lose
Domestic and
Cross-border N/A Overview of previous studies
Doukas and Travlos
(1988) US 1975-1983 (t=0)
AR (0.08%),
insignificant Cross-border 301
Bidding firms’ shareholders benefit most when their firm's expansion is taking place in less developed countries.
Conn and Connell
(1990) US-UK 1971-1980 (-30,0) CAR (3.39%) Cross-border 73
CARs for UK acquired firms are about half the CARs for US acquired firms at the month of merger announcement.
Morck, Shleifer and Vishny
(1990)
US 1975-1987 (-1,1) Bidder return
(0.70%) Domestic 326
Negative announcement period returns to bidding firms are due to managerial objectives.
Kaplan and Weisbach (1993)
US 1971-1982 (-5,5) CAR (-1.49%) Domestic 271
Acquirer returns at the acquisition announcement are significantly lower for unsuccessful divestitures than for successful divestitures and acquisitions not divested.
Corhay and Rad
(2000) The Netherlands 1990-1996 (-5,5) CAR (1.44%) Cross-border 84
Weak evidence is found that cross-border acquisitions are generally wealth-creating corporate activities.
Doukas, Holmén and Travlos
(2002)
Sweden 1980-1995 (-5,5) CAR (-2.53%) Domestic 101 Only diversifying acquisitions
Goergen and Renneboog (2004)
Intra-Europe 1993-2000 (-1,0) CAR (0.7%) Domestic and
Cross-border 187
Domestic acquisitions lead to higher wealth effects than cross-border acquisitions.
Hernando and Campa
(2004) European Union 1998-2000 (-1,1) & (-30,30)
CAR (0.44% and 0.56%) insignificant
Domestic and
11
Table I (continued)
Summary of results of previous event studies on the
announcement effect of acquisitions: returns to acquiring firms' shareholders Authors Region Research
period
Primary event window (days)
Positive/negative returns
Country coverage Number of observations
Additional results/Notes
Lowinski, Schierech and Thomas (2004)
Switzerland 1990-2001 (-1,1) CAR (1.07%) Domestic and
Cross-border 114
No significant difference in wealth creation between domestic and international merger activity is measured.
Conn, Cosh Guest and Hughes (2005)
UK 1984-1998 (-1,1) CAR (-0.82) Domestic and
Cross-border 4344 Moeller and Schlingemann (2005) US 1985-1995 (-1,1) CAR (1.173% and 0.307%) Domestic and Cross-border 4430
US firms who acquire cross-border targets experience lower abnormal returns than when acquiring domestic targets.
Martynova and Renneboog (2006)
Europe 1993-2001 (-60,60) CAR (-2.83%) Domestic and
Cross-border 2419
Alexandridis, Petmezas and Travlos (2010)
Global 1990-2007 (-2,2) CAR (-0.91) Domestic 4577
12
Conn and Connell (1990) provide empirical evidence that bidding firms from the US and the UK experienced positive returns in the month prior to the announcement of the merger. The wealth gains are be attributed to synergies. In addition, they find that the cumulative abnormal returns for UK acquired firms are about half the cumulative abnormal returns for US acquired firms at the month of the merger announcement. Asquith. et al. (1983) find that mergers generate positive net present value for US bidding firms’ shareholders and that merger activities are consistent with value-maximizing behavior by management. Similar to the US and the UK, positive announcement returns are found for European shareholders. Georgen and Renneboog (2004) examine the short-term wealth effects of large intra-European takeover bids in the period from 1993 to 2000. They find a positive statistically
significant announcement effect for the bidders. Martynova and Renneboog (2006) come to the same conclusion investigating the short-term wealth effects of European mergers and acquisitions for the same period.
Many studies find negative abnormal returns for bidding firms around the announcement date. Kaplan and Weisbach (1992) study a sample of large acquisitions and find negative returns to bidding firms shareholders (-1.49%) in the short-run. This is in line with the results from Morck, Shleifer and Vishny (1990) who studied acquisitions from the US between 1975 and 1987. A recent study by Alexandridis, Petmezas and Travlos, using a worldwide sample covering 39 countries from all continents, find negative abnormal returns for acquiring firms around the deal announcement. Regarding the theory and the literature, the following hypotheses are derived:
Hypothesis 1a: Announcements of acquisitions lead to wealth gains for bidding firms’ shareholders.
Hypothesis 1b: Announcements of acquisitions lead to wealth losses for bidding firms’ shareholders.
Target firm
13
Table II
Summary of results of previous event studies on the
announcement effect of acquisitions: returns to acquiring firms' shareholders Authors Region Research
Period
Primary Event window
Positive/negative returns
Country coverage Number of Observations
Additional results/Notes
Jensen and Ruback
(1983) US Till 1983 N/A Target firms benefit
Domestic and
Cross-border N/A Overview of previous studies
Harris and Ravenscraft (1991)
US 1970-1987 (-20, 4) CAR (26.33) Domestic and
Cross-border 1273
Target wealth gains are higher in cross-border takeovers than in domestic acquisitions.
Kaplan and Weisbach (1993)
US 1971-1982 (-5,5) CAR (26.9%) Domestic 271
Acquirer returns at the acquisition announcement are significantly lower for unsuccessful divestitures than for successful divestitures and acquisitions not divested. Danbolt (2004) UK 1986-1991 (-30,30) CAR (18.76%) and (19.60%) Domestic and Cross-border 630
Abnormal returns to target shareholders are higher in cross-border acquisitions acquired by companies based outside Europe or the USA
Goergen and Renneboog (2004)
Intra-Europe 1993-2000 (-1,0) CAR (9%) Domestic and
Cross-border 187
Domestic acquisitions lead to higher wealth effects than cross-border acquisitions.
Hernando and Campa
(2004) European Union 1998-2000 (-1,1) & (-30,30) CAR (4% and 9%)
Domestic and
Cross-border 262
Mergers in regulated industries show lower cumulative abnormal returns.
Martynova and Renneboog (2006)
Europe 1993-2001 (-60,60) CAR (26.70%) Domestic and
Cross-border 2419
Alexandridis, Petmezas and Travlos (2010)
Global 1990-2007 (-2,2) CAR (17.60%) Domestic 4577
14
Danbolt (2004) analyses the abnormal returns to target shareholders from acquisitions of UK companies. His results show that during the month of the bid announcement, large and highly
significant abnormal returns accumulate to target shareholders. The level of abnormal returns to target shareholders in acquisitions is found to be a little higher in bids by companies outside Europe or the USA. Georgen and Renneboog (2004) did not only find positive statistically significant announcement effects for the shareholders of the bidding firms, but they find even higher abnormal return for target shareholders. Harris and Ravenscraft (1991) study shareholder wealth gains for acquired US firms and their findings indicate that target wealth gains are significantly positive. Martynova and Renneboog (2006) also find a positive announcement effect for the target firms (9%). Finally, besides studying the shareholders wealth of bidding firm shareholders Hernando and Campa (2004) also investigated target firm shareholders wealth. They come to the conclusion that target firm shareholders receive a higher cumulative return (9%) than acquirer’s cumulative abnormal return, which is zero. Following the literature and the theory above:
Hypothesis 2: Announcements of acquisitions lead to wealth gains for target firm’s shareholders.
This overview of the shareholder values for bidding and target firms illustrates that there are no uniform empirical results confirming the wealth creation for the shareholders of bidding firms. However, previous studies are animus about the fact that target shareholders earn large positive abnormal returns. The literature finds that target firm shareholders gain a more positive abnormal return than bidding firm shareholders. For bidding firms there is little consensus in the literature about the sign of the price reaction to the announcement of the acquisition.
2.2 Factors determining the impact on shareholder value
Several factors explaining variations in wealth effects of French and Dutch acquisitions in various countries will be discussed. These factors may have a significant impact on the performance of acquiring firms in the days around the acquisition announcement. Deal, target and bidder firm related variables that have been identified in previous literature as affecting bidder returns will be presented. In addition, influences of country characteristics and differences will be studied. Numerous theories are presented to get a better understanding of the potential sources of value gains along with forming expectations about these factors. From this hypotheses will be derived.
2.2.1. Deal, acquiring and target firm characteristics
Relatedness
15
create most value with related acquisitions (Nayyar, 1993; Comment and Jarrell, 1995). The reduction of production and distribution costs (Martynova and Renneboog, 2006) lead to more positive value in related acquisitions because skills and resources can be used in related markets (Rumult, 1974). Therefore, most (operating) synergies are created from related acquisitions (Healy et al., 1997).
Hypothesis 3: Acquisitions with targets from unrelated industries lead to lower abnormal returns for bidding firms shareholders than acquisitions with targets from related industries.
Much attention has been paid by previous research to the relatedness of the target and the acquiring firms. These results are mixed. According to Berger and Ofek (1995) diversifying mergers reduce shareholder wealth. They compare the sum of the stand-alone values for individual business segments to the firm’s actual value. Their findings indicate a 13% to 15% value loss from diversifying mergers over the period from 1986 to 1991. Moeller and Schlingemann (2005) support this theory. However, Corhay and Rad (2000) study the wealth effects on international acquisitions by Dutch firms during the period of 1990 to 1996 and find that the diversification effects of entering into an unrelated business dominates the synergy effects of remaining within the same industry. Doukas and Travlos (2002) and Martynova and Renneboog (2006) and Hernando and Campa (2004) also surmise that conglomerate mergers create more value than the others. Lins and Servaes (1999) examine the valuation effect of diversification for large samples of firms in Germany, Japan and the United Kingdom. They find dissimilar significance concerning the diversification discount for the bidding firms from these countries. Their findings suggest that international differences in corporate governance affect the impact of diversification on shareholder wealth.
Method of payment
According to Martynova and Renneboog (2009) bidding firms use the method of payment as a tool to reduce the risks associated with the takeover deal. In a perfect market with symmetric information and the absence of taxes, the bidding firm’s shareholders would be indifferent between a payment in cash, debt or equity (Modigliani and Miller, 1958). The preference for payment in cash or in equity differs when we assume asymmetric information between the bidding management and bidding firm shareholders. The bidding firm’s management prefers a cash payment if they know that their shares are worth more than their current market value (Fishman, 1989). Conversely, management of the bidder will prefer an equity offer if they believe their shares are overvalued (Myers and Majluf, 1984). In this case, the method of payment will act as an information signal. Payment in cash usually
16
compensate for the taxation. The depreciation will reduce future profits and thus reduce tax burdens. Based on the theories above:
Hypotheses 4: Acquisitions paid in all-cash lead to higher abnormal returns for the bidding firms shareholders than using other methods of payment to pay for acquisitions.
It has been documented that the method of payment plays a major role in determining acquirer
performance. The hypothesis is in line with Martynova and Renneboog (2009). They find that all-cash bids generate higher bidder returns than all-equity acquisitions. Servaes (1991) analyzes the relation between takeover gains and the q ratios of targets and bidders over the period 1972 to 1987. According to him cash takeovers increase bidder abnormal returns by 11%. Sudarsanam and Mahate (2003) study the effect of different acquirer types based on financial status. Their results show that cash acquirers generate higher returns than equity acquirers. This theory is also supported by Rau and Vermaelen (1998) and Alexandridis, Petmezas and Travlos (2010) and Harris, Franks and Mayer (1987).
Target and acquiring firm size
Although many studies measure relative size of the target, it is important to look at the size of the target and the size of the bidder separately because the sizes could be out of proportion due to large or small target or bidding firms. I focus on large acquisitions (minimum of 1 million) and therefore the relative size might be homogeneous.
Target firm size could have a negative effect on abnormal returns due to the agency problem, where managers may acquire targets for their best interest. Through empire building, the managers’ goal is to increase firm size and it will choose a target which is relatively large. Ciscell and Carroll (1980) and Conyon and Murphy (2008) show a positive relation between firm size and the managers pay levels and power. Therefore, top managers might have the incentive to increase firm’s size and invest in acquisitions that might generate negative returns. These acquisitions could be unprofitable, but may allow them to increase their private benefits over the other shareholders benefits.
Hypothesis 5a: Acquisitions with larger target firms lead to lower abnormal returns for bidding firm shareholders than acquisitions with smaller target firms.
17
Recalling the agency theory, Moeller et all. (2004) find that large acquiring firms pay higher premiums than small acquiring firms and they enter into acquisitions with negative synergy gains. They provide evidence that managers of large firms pay more for acquisitions and that the premium paid increases with firm’s size.
Large bidding companies are more likely to make bad decisions since their management is hard to lay off. Managers in big firms are freer to repeatedly make bad acquisitions without being replaced (Martin, Combs and Moran, 2009). The difficulty in monitoring the activities of managers increases with the size of the firm. Therefore,
Hypothesis 5b: Acquisitions with larger acquiring firm’s lead to lower abnormal returns for bidding firm shareholders than acquisitions with smaller acquiring firms.
Offenberg (2009) finds that managers of large firms enter into acquisitions that generate lower results than smaller firms. This is in line with the study of Conn, Cosh, Guest and Hughes (2005).
The studies mentioned above consider the size of the target and bidder separately.
2.2.2. Country characteristics
According to Lowinski (2004) no diversification gains can be generated from international acquisition activity if international capital markets are perfectly integrated, information is cheaply available and agents behave rationally. In a perfect capital market all securities are correctly priced and no excess returns occur. However, Campbell et al. (1993) argue that perfect efficiency is an unrealistic
benchmark that is unlikely to hold in practice. Cross-border acquisitions can offer additional sources of value gains or losses to acquirers. Theories (Harris and Ravenscraft, 1991; Caves, 1971; Hymer, 1990; Errunza and Senbet, 1981) have pointed out that the principal motives for cross border takeovers are: 1. imperfections and costs in product or factor markets, 2. imperfections and information
asymmetries in capital markets, and 3. differences in government and regulatory policies.
The first motive suggests that imperfections in factor and product markets are based on the assumption that domestic assets are worth more to a foreign buyer than to a domestic buyer (Harris and
Ravenscraft; 1991). Therefore, additional synergies are created when firms can internalize actions that are costly to execute throughout the market mechanism (Teece, 1985). By expanding into foreign markets they can capture rents that are not competitively priced (Georgen and Renneboog; 2004). This motive is expected to create positive wealth effects.
18
abnormal returns can be earned by the ‘lemons problem’, which argues that information asymmetry between capital markets prevents bidding firms from obtaining the true value of the target in less developed capital markets (Conn, Cosh, Guest and Hughes, 2005).
The third motive indicates that differences across borders in taxation (Scholes and Wolfson, 1990 and Servaes and Zenner, 1994), accounting standards (Bris, Brisley, Christos, 2008), tariff and trade policy (Harris and Ravenscraft, 1991, Hemphill, 2010) are expected to have a positive or negative influence on the value of targets and bidders. Government and regulatory policies are related to the legal environment and the corporate governance structure of countries. Positive wealth gains from cross-border acquisitions can be generated due to improvements in the governance of the bidding and target firms as a result of spillovers of corporate governance standards between two firms (Martynova and Renneboog, 2008). An important part of corporate governance is shareholder protection. Shareholder protection reflects the legal system of a country through the enforcement of regulations and laws. The shareholder protection of a country is based on the government and regulatory policies (Kuiper et al., 2009, and La Porta, Lopez-de-Silanes, Shleifer and Vishny, 2000). It is important because bidders face the risk that target managers and target shareholders will use the profits of the firm to benefit
themselves, instead of returning the money to the bidding firm shareholders and managers (La Porta, Lopez-de-Silanes, Shleifer and Vishny, 2000). Furthermore, cultural aspects and distance between the home and target countries are important determinants for abnormal returns in cross-border acquisitions (Elango; 2006, Brock; 2005). Cultural differences make integration difficult, time consuming and expensive due to the complexity in managing the merger process (Conn, Cosh, Guest and Hughes; 2005). Therefore, they are expected to have negative wealth effects on bidding firms’ shareholders. Numerous studies (Grubel, 1968; Bailey and Stulz, 1990; Meric and Meric, 1989; Dunning 1993) have identified international diversification as a source of wealth gains. However, several studies (Eckbo and Thorburn; 2000; Finnely and Conn, 2003) find negative abnormal returns for cross-border acquisitions.
I will put forward a number of sources of potential wealth gains or losses related to country differences in domestic and cross-border acquisitions.
Shareholder protection
Shareholder protection has been identified by literature as one of the drivers for acquisitions performance. The controlling shareholders of a company can implement policies that benefit
themselves at the expense of minority shareholders (La Porta, Lopez-De-Silanes, Shleifer and Vishny; 2000). With better shareholder protection, exploitation of company profits by controlling shareholders or management is restricted. Better shareholder protection is related to higher valuation of corporate assets (La Porta, Lopez-De-Silanes, Shleifer and Vishny, 2002). Moreover, when targets are
19
bidder returns are expected to be higher (La Porta, Lopez-De-Silanes, Shleifer and Vishny, 2000). Therefore,
Hypothesis 6: Target countries with higher shareholder protection lead to lower abnormal returns than countries with a low shareholder protection.
This is in line with Georgen et al. (2005) who argue that strong protection of minority shareholders makes acquisitions costly and therefore more wealth is created when companies acquire targets in countries with weaker shareholder protection. Moeller and Schlingemann (2005) agree that acquirer gains are lower for transactions involving targets in countries with a more restrictive institutional environment and thus have more shareholder protection. However, Yen and Andre (2007) find that acquiring firms with greater investor protection have a positive impact on the operating performance from acquisitions. Guerin (2006) studies the role of geography in foreign direct investment, trade and portfolio investment flows. The familiarity effect causes investors to favor close-by countries with similar characteristics and legal systems over more distant and institutionally different countries.
Law
Rossi and Volpin (2004) show that legal environment and takeover regulations are important
determinants of takeover gains. They find that the premiums paid by bidding firms are lower for firms based in countries with lower shareholder protection. Shareholders’ rights are protected in different ways depending on the legal tradition that provides the foundation for corporate law. The most important legal institution is corporate law, which specifies the rights and obligations of owners and managers (Guler and Guillén, 2010 ).
La Porta, Lopez-De-Silanes, Shleifer and Vishny (2000) indicate that one of the key remedies to the agency problems is law. Laws give shareholders certain powers to protect their investment against expropriation by managers. These laws give the minority shareholders more rights, for example retaining dividends and voting for directors. La Porta, Lopez-De-Silanes, Shleifer and Vishny (1999) show that common law countries appear to have the best legal protection. Whereas, civil law countries have the weakest protection leading to higher abnormal returns.
Martynova and Renneboog (2008) state that takeover premiums in cross-border acquisitions increase with the differences in shareholder protection. Based on the theory above and recognizing that the bidding firms are from civil law countries, an acquisition with a target from a common law country will lead to lower abnormal returns.
20
Contrary to this hypothesis, Moeller and Schlingerman (2005) indicate that acquirer gains are lower for acquisitions involving countries with more restrictive capital markets and acquisitions involving countries with French civil law
Level of economic development
Acquisitions in emerging markets present opportunities that are different from transactions in developed markets, as these markets are characterized by imperfections in their financial, product and labor markets as well as in their corporate governance practices (Hoskisson et. al., 2000)
Developing countries have less developed capital markets, characterized by a high degree of
information asymmetry and therefore are less efficient. This high degree of information asymmetry is associated with the undervaluation of targets. The bidder pays less for the target because it cannot be valuated accurately. Therefore, bidding firms may realize significant positive returns from taking over targets in countries with less-developed capital markets. Furthermore, since the differences with emerging markets are larger, the benefits of this are greater from entering into acquisitions with less integrated economies (Doukas and Travlos, 1988). This level of integration depends on the level of economic development of the two counties. More benefits are created with targets firms from emerging countries, since the bidding firm in this study is from a developed country. This is in line with Francis, Hasan and Sun (2008) who find positive wealth effects for bidding firms entering into acquisitions with targets from segmented markets.
Therefore,
Hypothesis 8: Acquisitions with targets from less-developed countries will lead to higher abnormal returns for bidding firms shareholders than acquisitions from developed countries.
To the best of my knowledge, not many studies have examined the role of developed and developing countries in relation to shareholders’ wealth. However, the hypothesis is in line with Doukas and Travlos (1988) who find that shareholders from multinationals benefit the most when their firms’ expansion takes place in less developed countries. This article dates from 1988 and might not
represent the current market. Chari et al. (2004) find that cross-border M&A transactions in emerging markets lead to a creation of surplus value for bidding and target firm’s shareholders.
Cultural distance
21
that firms from the US target countries that are culturally closer to their home countries. This is in line with the cultural familiarity theory proposed by Lee, Shenkar and Li (2008). The theory states that firms are less likely to invest in countries which have a larger cultural distance. There is evidence of poorer performance regarding these investments. The cultural distance hypothesis by Hofstede (1983) suggests that costs and risks associated with cross-cultural contact increases with growing cultural differences between two individuals, group, or organization. Following the theories above, the subsequent hypothesis is derived:
Hypotheses 9: Acquisitions with firms with large cultural differences between the acquiring and target firm lead to lower abnormal returns for bidding firms shareholders.
22
3. DATA
In this section, I introduce the data set used in this study. Firstly, the data sources and the sample construction of the data are described. Secondly, the characteristics of the final sample will be illustrated.
3.1 The sample
This study is based on acquisitions which are undertaken by French and Dutch companies listed on the Euronext stock exchange market announced in the period from the 1st of January 2000 up to the 31st of December 2010. Firms from the Netherlands and France dominate the Euronext markets accounting for the largest amount of deals. This study succeeded in constructing a final dataset that includes 82 cross-border and domestic acquisition deals announced by Euronext firms and listed target firms from 29 developed and emerging countries2. The data has been obtained from the Zephyr database of Bureau Van Dijk to find the announcement dates. Zephyr is up-to-date and the most comprehensive database of deal information with integrated, detailed company information. The database is linked with other databases within Bureau Van Dijk and has a detailed search criterion that makes it possible to create a unique sample. I adopt the definition of an acquisition from the Zephyr database. For the target companies the classification of nationality used by Zephyr is assumed in this paper. An initial sample of 199,595 acquisition announcements of completed transactions is identified by Zephyr during this period:
The transactions that did not fulfil the following conditions were eliminated:
1. Acquiring firms needed to be listed on the Euronext stock exchange and target firms needed to be listed at a stock exchange in order to retrieve daily share price data for the estimation period.
2. The transaction involved the purchase of a majority stake of the target firm, more than 50%, in order to insure a definitive shift in hierarchical power from the target to the acquirer. This creates an influence on shareholders’ wealth.
3. The deal value had to be at least 1 million EUR.
4. The bids made by the same bidder if these bids occur within less than 300 trading days since the previous announcement of a bid, are excluded.
5. The bids made by the same bidder were excluded if these bids occurred within less than 300 trading days since the previous announcement of a bid. Here only the initial event is included.
2
23
The available stock data was collected from the Thomson Datastream after the total sample was identified. The Datastream adjusted total return index for the acquiring and target firm had to be available for the full event window and at least 30 days of the estimation period. This resulted in 82 deals for determining the acquiring firm’s cumulative abnormal returns and 59 deals for determining the target firm’s cumulative abnormal returns.
3.2 Dependent and independent variables
Following Georgen and Renneboog (2004) and Campa and Hernando (2004) I use two dependent variables for determining the shareholders value. I use the cumulative abnormal returns (CAR) of the acquirer and the cumulative abnormal returns of the target. The descriptive statistics of the dependent variables will be presented in the next section.
The independent variables in this paper include the size of the acquirer, size of the target, the relatedness of the industries, the method of payment of the deal, the cultural distance between the bidding and target firms’ country, the level of economic development of the target country, the law of the target country and the shareholder protection of the target country.
Table III shows the independent variables. When the data is not found in Zephyr, the missing data is obtained from Thomson One Banker. The actual size of the target and the acquirer separately is used to as a proxy for size. Following Doukas and Travlos (1988), Hernando and Campa (2004) and Yen and André (2007) the US 2 digit industry SIC code is used as a proxy for industry relatedness. I use a dummy variable that equals one when the acquirer and target operate in the same industry, zero otherwise. In my sample most acquirers select target companies from the same industry (57%), while the remainders are diversifying acquisitions. Moreover, cash is used as a proxy to measure the method of payment. This proxy is used in various other studies (Georgen and Renneboog, 2004; Harris and Ravenscraft, 1991). Almost half of the transactions were paid with cash only (55%). As documented in earlier studies (Lee et al., 2008; Slangen, 2006) Kogut and Singh’s index is used to measure the cultural distance between the targets and acquiring countries. They developed an index applying the 5 dimensions of Geert Hofstede (1980).
The index from the World Bank country classification index by the Organization for Economic Co-operation and Development (OECD) was used to capture the level of economic development. The international economic organizations’ aim is to stimulate economic progress and world trade. Their index classifies countries by their gross national income per capita into developing and developed countries3. The majority of the acquisitions were entered into with targets from developed countries (69%). Acquisitions were more prominently entered into with companies from civil law countries
24
(79%). The Netherlands and France are measured to have the same type of corporate governance (3.0) according to the self-dealing index. Therefore, differences in corporate governance may not be of influence for the bidding firms in this study. Because there are so many target countries involved in the data of this paper it is important to measure the extent of legal protection to the shareholders of these countries. La Porta, Lopez-De-Silanes, Shleifer and Vishny (1998) argue that the extent of legal protection of external investors differs immensely between countries. The legal system is often used as a proxy for the degree of corporate governance in a country. Taking this into consideration and
following the study of Moeller and Schlingerman (2005) law is used as a proxy for the degree of corporate governance in the target country. Following Yen and André (2007), a proxy for shareholder protection is used in the form of the revised anti-director rights index created by Djankov, La Porta, Lopes-de-Silanes and Shleifer (2006).4 The revised anti-self dealing index is a measure of legal protection of minority shareholder gains expropriation by corporate insiders.
These variables are measured without considering other factors such as share repurchases, dividends and capital restructuring. To control for external effects, following Danbolt (2004) and Hernando and Campa (2004) this paper will control for relative target firm size. Size is defined as the target firms total assets year end prior to the announcement divided by the bidding firms total assets year end prior to the announcement. The control variable GDP is measured by the log of the GDP of the target country to incorporate the differences between the minimum and maximum values. To check for outliers, histograms were made for all variables.
4
The index is based on the following six areas: (1) vote by mail; (2) obstacles to the actual exercise of the right to vote (i.e., the requirement that shares be deposited before the shareholders’ meeting); (3) minority
25
Table III Key variable characteristics
Variable Description Source Mean St. Dev.
Acquirer characteristics: Size Log of total assets of acquirer Zephyr 7.06 1.30
Target characteristics: Size Total assets of target Zephyr 5.46 1.24
Deal characteristics: Relatedness Acquirer and target SIC codes Zephyr 0.43 0.50 Method of payment Solely cash vs. Other Zephyr 0.55 0.50
Country characteristics Cultural distance Cultural distance index Kogut and Singh (1988) 0.76 1.05 Law Common law vs. Civil law www.cia.gov 0.21 0.40 Shareholder protection Shareholder protection index target country Revised Anti-Director
rights Djankov et. al (2006)
3.05 0.78
Level of economic
development Emerging vs. Developed country www.oecd.org 0.31 0.46
Size of acquirer (target) is the log of the total assets of the acquirer (target) the last fiscal year prior to the announcement. Relatedness is based on the first 2 US SIC digits between the acquiring and target firm. It is a dummy variable that equals 1
when the acquirer and the target operate in different industries, and 0 otherwise. Method of payment is a dummy variable equal to 1 when paid solely with cash, otherwise 0. Cultural distance is the differences between the acquiring and the target firm based on Kogut and Singh’s (1988) index. This index is based on the deviation along the four cultural dimensions (Power distance, uncertainty avoidance, masculinity/femininity and individualism) identified by Hofstede. Level of economic
development is a dummy variable equal to 0 for developed target countries and equal to 1 for emerging target countries based
on the World Bank Country Classification. Law is a dummy variable equal to 0 when the target country is a civil law country, 1 if target country is a common law country based on the Central Intelligence Agency’s World Factbook. Shareholder
protection is degree of target countries’ shareholder protection based on the Revised Anti-Director rights index of Djankov et
26
4. ABNORMAL RETURNS
4.1 Event study methodology
This paper measures the short-term wealth effects of bidding and target firms’ shareholders by calculating the cumulative abnormal returns in an event study (CAR). The standard event study method is used to measure the abnormal returns earned by firms announcing domestic and international acquisitions (Corhay and Rad, 2000; Martynova and Renneboog, 2006; Conn et al., 2005). The date of the acquisition announcement is the event date and is defined as day 0. The methodological approach that is used in this research follows the work of MacKinlay (1997) and Brown and Warner (1985). The assumption that capital markets should be efficient is essential indicating that security prices fully reflect all available information and that prices instantly change to reflect new public information. Therefore, information and trading costs (the costs of getting prices to reflect information), are zero (Fama, 1991).
The event window in this study ranges from 15 days prior to 15 days after the announcement of the acquisition. Interval (-15,15) is about the average used in event studies (Jensen and Ruback, 1983; Hernando and Campa, 2004). The abnormal returns were calculated for this interval to catch early stock price reactions stimulated by leakage of information and to identify potentially slow information processing after the event. An estimation window will be set to test for abnormal returns prior to the event window. According to MacKinay (1997) an estimation window of over 200 days prior to the event would be sufficient. Therefore, in this study the model is estimated over a 220 day period, starting 236 days prior to the event date and ending 16 days prior to the event date (-236,-16). The abnormal returns of each company are computed as the difference between the actual return and the normal return, which is the expected return in case there would not have been an acquisition announcement. The abnormal return is defined as:
it itit
R
E
R
AR
(4.1)
Where ARit, Rit and E(Rit) respectively indicate the abnormal, actual and normal returns for firm i on
day t. And the actual returns are is calculated as
1 1
it it it itR
R
R
R
.27
because by removing the portion of the return that is related to variation in the market’s return, the variance of the abnormal return is reduced:
it mt i i it
a
R
R
(
)
(4.2)Where
R
it andR
mt are, respectively, the period-t
returns on firm i and the market portfolio. The zero mean disturbance term is presented as
it. Finally,
i anda
i are the parameters of the market model. Coefficienta
i is the intercept term for the regression equation related to the bidder’s security. Coefficient
i captures the systematic risk of the bidder’s security. The market portfolio is arepresentative broad based stock index which reflects the overall market fluctuations. In this paper the Euro Stoxx 50 is used as a proxy for the market. The Euro Stoxx 50 index is Europe’s leading index for the Eurozone and covers 50 stocks from 12 Eurozone countries.5 Applying the market model, the daily abnormal return is calculated:
mt i i it it
R
R
AR
ˆ
ˆ
(4.3)Here
ˆ
i and
ˆ
i are the ordinary-least-squares estimates obtained from regressions of firm i’s daily returns on the market return over the estimation period from t=-236 to t=-16.Subsequently, individual securities’ abnormal returns are aggregated. Given N events, the abnormal returns for period t are:
N i it tAR
N
AR
11
(4.4)N is the number of events. In order to draw overall conclusions about the impact of the event on the value on shareholder’s wealth, the abnormal returns need to be aggregated. This forms the cumulative average abnormal returns (CAR), from which
1 to
2 is the sum of the included abnormal returns:
2 1 2 1,
)
(
AR
tCAR
(4.5)Where
1, is the beginning of the event and
2 is the end of the event. To determine the statistical significance of the results, t values for each CAR is calculated. The CARs are used to examine the effect the announcement has on the value of the firm. Assuming market efficiency, average CAR should not significantly differ from zero. Positive CARs imply that the market expects the acquisition to create value and negative CARs imply value losses.5
28
4.2 Descriptive statistics
The first variable used is the cumulative abnormal return of bidding firms from Euronext listed companies for the event window based on estimated alphas and betas. Firms from Amsterdam and Paris make up for the substantial amount (82) of transactions. Whereas firms from Brussels are only responsible for 9 transactions. Therefore, transactions from Amsterdam and Paris are used to represent the Euronext acquiring companies. This dependent variable is used to investigate whether
announcements of acquisitions create wealth for bidding firms’ shareholders (hypotheses 1). The second dependent variable is the cumulative abnormal return of listed target firms for the event window based on estimated alphas and betas. This dependent variable is used to investigate whether announcements of acquisitions create wealth for target firms’ shareholders (hypothesis 2). The same deals as for the first dependent variables are used. However, only 59 observations are found because several listed targets are not traded on the stock market. I have not found an explanation for this. The majority of the acquisitions occurred in France (54), while the average transaction value and the total transactions values were larger for The Netherlands, respectively EUR 799,000 and EUR 2,2376,000, accounting for 80.0% and 67.9% of the sample.
29
Table IV
Distribution of the number of transactions over time, the minimum transaction value, the maximum transaction value, the average transaction value and the total transaction values
Year Number of Transactions Minimum Transaction value (EUR min) Maximum Transaction value (EUR mil) Average Transactions value (EUR mil) Total Transaction values (EUR mil)
2000 5 7 1,516 751 3,753 2001 4 4 308 106 425 2002 3 44 211 110 330 2003 2 36 40 38 76 2004 7 3 39 45 314 2005 7 8 2,746 628 4,399 2006 9 2 4,2 842 7,579 2007 16 2 1,955 204 3,076 2008 8 2 330 90 716 2009 8 9 334 83 661 2010 13 2 5,585 863 11,654 2000-2010 82 2 5,585 402 31,963
Table V provides an overview of the distribution of the transactions per industry. The industries are classified by US primary Standard Industry Classifications (SIC) codes assigned by the government. The primary SIC code for the company defines the company’s core business. Computers, Electronics & Software, Finance, Real Estate and Services make up for more than half (57.3%) of the sample. The other industries account for the minority of the deals. With five industries (Healthcare, Industrial Products, Insurance, Mining and Oil & Gas) representing two deals per industry.
Table V
Number transactions per industry
Target industry Number % of sample
Computers, Electronics & Software 14 17.1%
Finance 13 15.9%
Real Estate 11 13.4%
Services 9 11.0%
Food, Beverages and Tobacco 7 8.5% Telecommunications 7 8.5% Transportation 5 6.1% Chemicals 3 3.7% Construction 3 3.7% Healthcare 2 2.4% Industrial Products 2 2.4% Insurance 2 2.4% Mining 2 2.4%
Oil & Gas 2 2.4%
30
4.3 Tests on significance
The results generated from the abnormal return calculations are not sufficient to make conclusions about the wealth effect. In order to test whether acquisitions have wealth effects, the abnormal ARs and CAR’s must be statistically different from zero. Parametric and non-parametric tests are
conducted to examine the significance of the abnormal returns for the event window. Results should be normally distributed in order to improve the validity of the parametric test. Non-normality can decrease the validity of the test. Therefore, a Jarque-Bera test will be carried out first.
The Jarque-Bera tests the null hypothesis that the data are normally distributed, based on the sample kurtosis and skewness6. For normal distribution the skewness needs to have a value of zero and the kurtosis coefficient must be three. The normality results for the bidding firms show a kurtosis of 3.234, a skewness of 0.666 and the Jarque Bera is 16.638. The target firm’s results for normality show a kurtosis of 0.151, a skewness of 0.125 and the Jarque Bera is 75.29. Therefore, the results show that the average ARs in the estimation period is normally distributed for the acquiring firms, since the skewness and kurtosis are close to the value indicated for normality. This implies that the parametric tests will offer the most consistent results in the event study. However, the Jarque Bera test for the target firms show signs of non-normality and therefore a non-parametric test is conducted.
Parametric and non-parametric test
To test the significance of the CARs and the ARs the standard parametric test statistics are computed like in Brown and Warner (1985). To determine the statistical significance of the results, t values for each of the AR and CAR are calculated. A two sided test is applied with the null hypothesis that the average ARs and the average CARs for the event window is equal to zero. The critical values on which the t-test will be tested are on a 10%, 5% and 1% level and the degrees of freedom will be N-1.
In addition to the parametric t-test, a non parametric test will be conducted. Brown and Warner (1985) point out that solely analyzing data with a t-test can lead to a distorted picture since a t-test assumes normality of the data. This study will apply a non-parametric test for both dependent variables in order to enhance the robustness of the results. Even though, the parametric test conducted above did not find any evidence of non-normality for the bidding firm’s abnormal returns. Following Martynova and Renneboog (2009 and 2006) and MacKinlay (1997) the Corrado rank test (Corrado, 1989) is used as a non-parametric test in this paper. According to Corrado this rank test does not require symmetry in cross-sectional excess returns distributions for correct specification and is better specified under the
6
The Jarque-Bera test is defined as: JB =
4 ) 3 ( 6 2 2 K S N
31
null hypothesis. The abnormal returns of each bidding and target company will be ranked over the period that includes the estimation and the event window.
Finally, a Pearson correlation analysis for the independent variables was conducted to check for multicollinearity, illustrated in table 1 in the appendix. The correlation table shows that there is some evidence of multicollinearity between the independent variables. When explanatory variables are correlated, the coefficient estimates might change in response to small changes in the data. Therefore, various tests have been conducted to test the robustness of the results.
4.4 Event study results and discussion
Bidding firm returns
32
Table VI
Daily average abnormal returns and cumulative average abnormal returns for 82 acquiring firms
Day relative to announcement Daily average abnormal return (AR) in % t-value Corrado-test statistic Cummulative daily average abnormal return (CAR) in % t-value -15 -0.061 -0.220 1.022 -0.061 -0.0137 -14 0.059 0.211 0.448 -0.002 -0.001 -13 -0.064 -0.229 0.020 -0.066 -0.017 -12 -0.750 -2.701*** 1.460 -0.816 -0.226 -11 0.085 0.308 -0.564 -0.730 -0,219 -10 0.107 0.386 -1.673* -0.623 -0.204 -9 0.318 1.144 -0.584 -0.305 -0.110 -8 -0.152 -0.546 0.953 -0.457 -0.181 -7 -0.224 -0.805 1.693* -0.681 -0.307 -6 0.184 0.664 -1.177 -0.496 -0.255 -5 -0.562 -2.024** 0.457 -1.058 -0.635 -4 -0.312 -1.120 1.673* -1.369 -0.986 -3 0.142 0.511 -1.732* -1.227 -1.105 -2 -0.277 -0.100 -0.068 -1.504 -1.807* -1 -0.161 -0.579 1.051 -1.665 -3.000** 0 -0.470 1.693* -1.868* -2.135 -7.692*** 1 0.134 0.485 -10.411 -2.000 -3.604*** 2 -0.407 -1.466 2.063** -2.408 -2.891*** 3 -0.109 -0.393 0.885 -2.517 -2.267** 4 -0.232 -0.838 -0.37 -2.749 -1.981* 5 0.361 1.230 -1.382 -2.388 -1.434 6 0.374 1.347 -1.810* -2.014 -1.037 7 0.102 0.369 0.603 -1.912 -0.861 8 -0.045 -0.162 1.138 -1.957 -0.784 9 -0.089 -0.322 0.691 -2.047 -0.737 10 0.080 0.287 -0.409 -1.967 -0.644 11 0.320 1.151 -1.236 -1.647 -0.494 12 -0.149 -0.537 0.808 -1.796 -0.498 13 -0.053 -0.191 1.090 -1.849 -0.476 14 -0.219 -0.790 0.234 -2.068 -0.497 15 -0.364 -1.311 0.924 -2.432 -0.548
33
announcements cannot affect the abnormal returns because the overlapping event dates of the same bidder are excluded. The absence of significant abnormal returns after the announcement date (expect for +2 and +6 in Corrado) indicate that the wealth gains are entirely incorporated within only several days of the announcement. The results of the non-parametric Corrado test also show acquisitions lead to significant negative abnormal returns during the estimation period. The test is significant for several days during the estimation period, including the day of the announcement. Almost all significant abnormal returns are negative.
Table VII provides a summary of the bidding firm CARs based on several event windows. I find negative CARs for all event windows. However, the primary event window (-15,15) is not significant. The event windows (-15,0), (-5,0), (-1,0) and (-1,1) are significantly different from zero when
executing a t-test. It appears that acquisitions from the Netherlands and France positively influence shareholders’ wealth on the first day after the announcement date. Yet, they have a negative effect over the period surrounding the announcement and also over a longer period. The CARs including days prior to the announcement (-15,0), (-5,0) and (-1,0) are highly significant. This could be due to the fact that the average returns on 12 and 5 days before are significant at 1% and 5%. This is in line with hypothesis 1b, and therefore this hypothesis is supported. Therefore, hypothesis 1a is not supported.
Table VII
Acquiring firm cumulative abnormal returns for different event periods
Target firm returns
Table VIII depicts the event study results of the 59 target firms. The table illustrates that acquisition announcements create positive abnormal returns. On the announcement date an average abnormal return of 4.921% is earned, which is significant at 1% according to the parametric and non-parametric Corrado tests.
Event window CAR % t-value
34
Table VIII
Daily average abnormal returns and cumulative average abnormal returns for 59 target firms
Daily average abnormal return (AR) in % t-value Corrado-test statistic Cummulative daily average abnormal return (CAR) in % t-value 0.061 0.135 -0.398 0.061 0.008 -0.857 -1.895* 0.894 -0.796 -0.117 1.205 2.664** -0.962 0.409 0.065 0.186 0.411 -0.214 0.595 0.101 0.216 0.477 -1.477 0.811 0.150 0.208 0.460 0.846 1.019 0.205 -0.184 -0.406 0.690 0.835 0.185 0.089 0.196 0.233 0.924 0.227 0.029 0.064 0.311 0.953 0.263 0.571 1.261 -1.322 1.524 0.481 -1.024 -2.264 1.331 0.500 0.184 -0.647 -1.430 1.040 -0.147 -0.065 0.478 1.056 0.068 0.331 0.183 0.731 1.615 2.264** 1.061 0.782 0.369 0.816 1.691* 1.430 1.581 4.921 10.876*** 4.369*** 6.352 14.037*** 1.562 3.451*** 3.136*** 7.913 8.744*** 3.742 8.271*** 3.535*** 11.656 8.586*** 0.443 0.980 0.408 12.099 6.685*** 0.540 1.194 0.301 12.640 5.587*** -0.888 -1.962 2.352** 11.752 4.329*** -0.660 -1.458 1.837* 11.092 3,502*** 1.228 2.715*** -0.293 12.320 3.403*** 0.097 0.215 -1.312 12.417 3.049*** -0.337 -0.745 -0.311 12.080 2.670*** 0.295 0.653 0.632 12.375 2.486** -0.107 -0.237 0.466 12.268 2.259** 0.112 0.248 -0.914 12.380 2.105** -0.519 -1.146 1.409 11.862 1.872* -0.206 -0.456 0.175 11.655 1.717* -0.626 -1.383 1.205 11.029 1.523
35
following the announcement date, with the exception of -13, -14 and +7, also. The Corrado test shows less significant following the announcement date. However, it is in line with the parametric test.
Table IX provides a summary of the target firm CARs based on several event windows. I find highly positive significant abnormal returns for all event windows except for event window (0,15). The primary event window (-15,15) is significant at a 10% level. It shows that acquisitions have a positive influence on target shareholders’ wealth and thus hypothesis 2 is supported.
Table IX
Target firm cumulative abnormal returns for different event periods
Summarizing the results, the wealth losses for the bidding firm shareholders were 0.470% on the event date and 2.432% for the event window. During this event window they are not significantly different from zero. Surrounding the announcement (-1,1) the abnormal returns are highly significant and 0.496% is lost. The short term window (-1,1) catches most of the announcement effect and therefore hypotheses 1b is supported. According to the theory described in section 2.1 negative wealth effects are generated from management failure. Therefore, the losses for bidding firms’ shareholders can be attributed to this aspect.
The target firm shareholder gained 4.921% on the event date and gained 11.029% for the whole investigation period. Around the announcement date (-1,1) 6.852% was gained. As a result, hypothesis 2 stating that acquisitions lead to higher abnormal returns for target firm shareholders, is supported. According to the theory described in section 2.1 positive wealth effects are generated from synergies. Therefore, the gains for target firms’ shareholders can be attributed to this aspect.
Event window CAR% t-value
36
5. THE DETERMINANTS OF THE ACQUISITION RETURNS
5.1 Ordinary Least Squares regression methodology
Following previous studies (Moeller and Slingemann, 2005; Conn, Cosh and Hughes, 2005; Lowinski et al., 2005) an ordinary least squares (OLS) regression analysis is conducted to test the relationship between the average CARs and the firm, deal and country characteristics described in section 2. The estimated regression refers to formula (5.1) and gives insight into hypotheses H3 to H9.
CAR
ti =
j+
1RELATi,j +
2LAWj +
3SHARPROTj +
4CASHi +
5SIZEACQi + (5.1) 6
SIZETAR j+
7CULDISi,j +
8ECDEVj +
10RELSIZEi,j +
11GDPTARj +e
iCARti The cumulative abnormal return for a firm i during event window t
RELATi,j Dummy variable 1 if firm i and j operate in different industries, zero otherwise
LAWj Dummy variable 1 if firm j is based in a common law country, zero otherwise
SHARPROTj Level of shareholder protection of the country where firm j is located in
CASHi Dummy variable 1 if the acquisition is paid for with cash, zero otherwise
SIZEACQi Log of the total assets of the acquirer
SIZETAR j Log of the total assets of the target
CULDISi,j Cultural distance between the countries where firm i and j are located in
ECDEVj Dummy variable 1 if firm j is located in an emerging country, zero otherwise
RELSIZEi,j Firm j’s total assets year end prior to the announcement divided by firm i’s
firms total assets year end prior to the announcement GDPTARj The log of the GDP of firm j
i
e
Error term of firm i
5.2
OLS regression results and discussion
37
Most of the previous studies only draw upon one or two event windows. It is interesting to analyse to what exact event window the wealth gains take please because emerging and developing markets are included in the sample. This exposes the possible cause of the delayed or premature wealth gains.
Table X illustrates the results of the OLS analysis modelling the relationships between firm, deal and country characteristics and firm performance (CAR) based on the four models. The regression tests the hypotheses H3 to H9. The overall explanatory powers of the regression are low. The adjusted R2 is 7.6% for model 1, 4.7% for model 2, 12.1% for model 3 and 4.8% for model 4. The relatively low adjusted R2 can be explained by the fact that there are many unobserved variables that are not incorporated in the regression. The low adjusted R2 are in line with previous studies (Danbolt, 2004 and Conn, Cosh, Guest and Hughes, 2005). My models are as successful as other studies in explaining the variance of the abnormal returns.
38
Table X
Results of Ordinary Least Square analysis modelling the relationships between firm, deal and country characteristics and acquiring firm performance (CAR)
Model 1 Dep.var = (-1,1) Model 2 Dep.var = (-10,10) Model 3 Dep. Var = (-10,0) Model 4 Dep. Var = (0,10) Variable Expected
sign Coef. T-value Coef. T-value Coef. T-value Coef. T-value
